## 317OFIRST EFFICACY RESULTS FROM THE TURANDOT PHASE III TRIAL COMPARING TWO BEVACIZUMAB (BEV)-CONTAINING REGIMENS AS FIRST-LINE THERAPY FOR HER2-NEGATIVE METASTATIC BREAST CANCER (MBC)

### Abstract

#### Background

TURANDOT is the first prospective trial to compare BEV combined with either paclitaxel (PAC) or capecitabine (CAP). We report the planned interim analysis (IA) of efficacy.

#### Methods

Patients with HER2-negative mBC who had received no prior chemotherapy for mBC were randomised to receive either BEV–PAC (BEV 10 mg/kg d1 & 15 + PAC 90 mg/m d1, 8 & 15 q4w) or BEV–CAP (BEV 15 mg/kg d1 + CAP 1000 mg/m bid d1–14 q3w) until disease progression or unacceptable toxicity. The primary objective is to demonstrate non-inferior overall survival (OS) with BEV–CAP vs BEV–PAC. Interim and final OS analyses were planned after 175 and 389 deaths, respectively, in the per-protocol (PP) population to reject the null hypothesis of inferiority (hazard ratio [HR] ≥1.33) with 80% power and overall α = 0.025. Secondary endpoints include response rate (RR), progression-free survival (PFS), safety and quality of life.

#### Results

Median follow-up was 19 months at data cut-off for this IA (1 Sep 2011). Baseline characteristics were generally similar in the 2 treatment arms.

BEV-PAC (n = 285) BEV–CAP (n = 279)
Median age, years 59 59
Visceral metastases, % 65 73
Prior (neo)adjuvant taxane, % 20 18
OSa
Events, % 33 35
1-year OS rate, %b 81 79
HR (97.5% RCIc)
for non-inferiority
1.04 (−∞ to 1.69)
p = 0.0593d
RR
Overall, % 44 27
CMH test (superiority) p < 0.0001
PFS
Events, % 62 77
Median, months 11.0 8.1
HR (95% CI) 1.36 (1.09 to 1.68)
Log-rank (superiority) p = 0.0052
BEV-PAC (n = 285) BEV–CAP (n = 279)
Median age, years 59 59
Visceral metastases, % 65 73
Prior (neo)adjuvant taxane, % 20 18
OSa
Events, % 33 35
1-year OS rate, %b 81 79
HR (97.5% RCIc)
for non-inferiority
1.04 (−∞ to 1.69)
p = 0.0593d
RR
Overall, % 44 27
CMH test (superiority) p < 0.0001
PFS
Events, % 62 77
Median, months 11.0 8.1
HR (95% CI) 1.36 (1.09 to 1.68)
Log-rank (superiority) p = 0.0052

RCI = repeated confidence interval

aPP population (n = 533)

bKaplan–Meier estimate

cUsing O'Brien–Fleming boundaries

dNon-inferiority not shown as p > 0.00105 (α at IA) AEs were consistent with the known safety profiles of BEV, PAC and CAP. The most common grade ≥3 AEs were neutropenia (18%), peripheral neuropathy (14%) and leucopenia (7%) with BEV–PAC and hand-foot syndrome (16%), hypertension (6%) and diarrhoea (5%) with BEV–CAP.

#### Conclusion

In this planned IA, the non-inferiority criterion has not been met but OS results do not indicate relevant differences. Final results are expected in 2014. PFS and RR were better with BEV–PAC and very similar to previous data for BEV–PAC (E2100) and BEV–CAP (RIBBON-1).

#### Disclosure

R. Greil: RG has received research support and honoraria from Roche.

S. Beslija: SB has received research support and honoraria from Roche.

D. Messinger: Employee of IST GmbH, CRO which is providing various services and consultancies for Hoffmann-La Roche and CECOG.

T. Brodowicz: TB has received honoraria from Roche.

All other authors have declared no conflicts of interest.

## 318OPH IB/II STUDY OF BKM120 PLUS TRASTUZUMAB (T) IN PATIENTS WITH T-RESISTANT HER2+ ADVANCED BREAST CANCER (BC)

### Abstract

#### Background

PI3K/AKT/mTOR pathway upregulation has been implicated in T resistance, and thus the impact of pathway inhibition on restoration of therapeutic sensitivity is being investigated. The RP2D of BKM120, an oral pan-class I PI3K inhibitor, plus T is 100 mg/d. Here, we present Ph II results of BKM120 + T in pts with T-resistant advanced HER2+ BC.

#### Methods

Pts with HER2+ locally adv/metastatic BC resistant to T (progression while on T, or within 4 wks [metastatic] or 12 mths [adjuvant] of last T dose) received daily BKM120 (100 mg) and the standard wkly dose of T. Ph II eligibility criteria: ≥1 measurable lesion, ≥1 but ≤4 prior anti-HER2 regimens (incl. trastuzumab [required], lapatinib, and/or T-DM1), and ≤3 lines of prior chemotherapy for metastatic disease. Ph IB pts treated at the RP2D who met Ph II eligibility criteria were included in the analysis.

#### Results

As of 23 March 2012, 53 pts were included in the Ph II analysis (safety set; incl. 8 pts from Ph IB). 49 pts were evaluable for response (full analysis set); median age 52 yrs (28–75); median no. prior antineoplastic regimens 4 (1–10); 5 pts had a baseline CNS lesion (3 measurable target; 2 non-target). At data cut-off, 9 pts were still on study. Most pts discontinued treatment due to disease progression (55%); 8 pts (16%) withdrew due to AEs. Mean duration of BKM120 exposure was 11 wks (0.1–41). Most common suspected study-drug related G3/4 AEs: ALT increased (5 pts); rash (5 pts); AST increased (4 pts); asthenia (3 pts); nausea, anxiety, skin photosensitivity, hyperglycemia (2 pts each). Partial responses (RECIST) were seen in 4 pts (8%), and stable disease (SD) was noted in 20 pts (41%); the disease control rate (CR, PR, or SD) was 49%. Preliminary results indicate that, of the 5 pts with baseline brain mets (BM), 2 pts had SD in the CNS without evidence of progression at study withdrawal; 2 pts had overall SD (1 for 90 days and 1 for 106 days) before progression in the CNS; 1 pt was not evaluated in the CNS after study entry.

#### Conclusion

BKM120 in combination with T has an acceptable safety profile, and has shown encouraging preliminary activity in heavily pretreated HER2+ metastatic BC pts with T resistance, including pts with BM.

#### Disclosure

Q. Ru: Employee of Novartis.

S. Ruquet: Employee of Novartis.

D.W. Sternberg: Employee of Novartis.

All other authors have declared no conflicts of interest.

## 319OSIGNIFICANT ANTITUMOR ACTIVITY OF E-3810, A NOVEL FGFR AND VEGFR INHIBITOR, IN PATIENTS WITH FGFR1 AMPLIFIED BREAST CANCER

### Abstract

#### Background

Amplification of the FGFR1 gene occurs in subsets of tumors, notably breast cancer (BC), where the altered FGF pathway may be clinically relevant.

#### Methods

E-3810 is a kinase inhibitor targeting FGFR1 and VEGFR1, 2, 3. Its safety and activity, at the daily oral dose of 20 or 15 mg on a continuous schedule, are being assessed in patients with solid tumors and FGFR1 amplification or potentially sensitive to antiangiogenic agents. The study is an open label non comparative extension of the first in man dose-escalation trial; the efficacy threshold, set for the FGFR1+ cohort only, requires 3/14 confirmed objective RECIST responses or non-progressive disease ≥ 6 cycles (one-stage Fleming design: H0 5%, H1 30%, power 80%).

#### Results

46 patients with various tumor types (including 13 FGFR1+) were recruited. 8 patients (4/13 and 4/33 treated respectively at 20 and 15 mg; none in the FGFR1+ cohort) were withdrawn for toxicity including: G3 proteinuria (5), headache and vomiting (2), pancreatic enzymes increase (1), recovered in all cases. Hypertension G2-3, proteinuria G2, GI intolerance, asthenia and weight loss led to dose reduction in 20 patients; frequent TSH increase required supplementation. Tolerability was better in the FGFR1+ cohort, in line with a limited prior exposure to antiangiogenic treatments. Antitumor activity is shown in the table:

Antiangiogenic Sensitive

FGF-amplified (1)

Evaluable PR SD PD Evaluable PR SD PD
Breast cancer   4**
Other 23 3*** 13  1*
Total 24 14 11
Antiangiogenic Sensitive

FGF-amplified (1)

Evaluable PR SD PD Evaluable PR SD PD
Breast cancer   4**
Other 23 3*** 13  1*
Total 24 14 11

(1) incl. the two BC patients with 11q amplification; * SD ≥ 24 weeks; ** 1 PET response (bone lesions); ***2 thyroid; 1 thymic carcinoma.

12 women with BC were recruited (9 HR + , 1 HER2 + /HR + , 2 TN); 8 were FGFR1 + , 2 more had 11q amplification (CGH). They were treated with a median of 5 prior chemotherapy lines; 9, 10 and 5 patients also received ≥ 1 endocrine, antiangiogenic and experimental therapies, respectively. There were 4 PRs sustained over 4-6 courses (3 FGFR1 and one FGF4 amplification), with 3 responders still on treatment.

#### Conclusions

E-3810 has shown significant activity in heavily pretreated BC, with durable responses in patients with altered FGF pathway. Further studies are planned in this population.

#### Disclosure

R. Cereda: Currently an employee and stock holder at EOS SpA.

M.G. Camboni: Currently an employee and stock holder at EOS SpA.

All other authors have declared no conflicts of interest.

## 320PDSIGNIFICANTLY SUPERIOR OUTCOME AFTER NEOADJUVANT CHEMOTHERAPY WITH DOCETAXEL, ANTHRACYCLINE AND CYCLOPHOSPHOMIDE VERSUS DOCETAXEL PLUS CYCLOPHOSPHOMIDE: RESULTS FROM THE NATT TRIAL IN TRIPLE NEGATIVE OR HER2 POSITIVE BREAST CANCER

### Abstract

#### Objective

To evaluate the activity and safety of docetaxel plus cyclophosphomide (TC) compared with docetaxel, anthracycline and cyclophosphomide (TAC) in neoadjuvant treatment of triple negative or HER2 positive breast cancer, which may help us to determine the role of anthracycline in breast cancer treatment.

#### Methods

Clinical stage IIB or III breast cancer patients were treated with six cycles of TC (docetaxle 75 mg/m2 and cyclophosphomide 600 mg/m2) or TAC (docetaxle 75 mg/m2, anthracycline, and cyclophosphomide 500 mg/m2). Either epirubicin 60mg/m2 or adimycin 50mg/m2 was allowed as an anthracycline regimen. The primary end point was pathological complete remission (pCR), defined as no residual invasive cancer in breast and axillary lymph node. Seconday end points included safety, clinical response rate, event free survival (EFS), disease free survival (DFS), and overall survival (OS).

#### Results

102 patients were randomized and 96 patients were available for analysis (TC: n = 45; TAC: n = 51). After surgery, pCR rate was 6.8% (3/45) and 17.6% (9/51) in TC and TAC group, respectively, P = 0.113. There was also no difference in clinical response rate. However, with mean follow up of 20 (3-36) months, TAC treatment, compared with TC, resulted in a significantly superior EFS (80.4% vs 60.0%, respectively; adjusted HR = 2.42 (95% CI: 1.11 to 5.30); P = 0.027), better DFS (84.3% vs 64.4%, respectively; adjusted HR = 2.85 (95% CI: 1.21 to 6.74); P = 0.017) and a trend towards less death (96.1% vs 86.7%, respectively; adjusted HR = 2.52 (95% CI: 0.41-15.38); P = 0.315). Patients treated with TAC had relative high rates of grade 3-4 neutropenia and leukopenia, and rates of other severe adverse events were similar and no patients died on treatment.

#### Conclusions

pCR rate was not significantly different between TAC and TC in neodajuvant treatment of triple negative or HER2 positive breast cancer. Adding anthracycline to TC can significantly improve patients' outcome, which deserves further investigation. (Sponsored by Sanofi; ClinicalTrials.gov number, NCT00912444.)

#### Disclosure

All authors have declared no conflicts of interest.

## 321PDIMPACT OF MULTIFOCAL OR MULTICENTRIC DISEASE ON SURGICAL, LOCOREGIONAL, AND DISTANT SURVIVAL AFTER NEOADJUVANT CHEMOTHERAPY IN 3562 BREAST CANCER PATIENTS

### Abstract

#### Purpose

To evaluate patients outcome with multifocal or multicentric breast cancer after neoadjuvant chemotherapy by type of surgery.

#### Patients and methods

Participants of the GeparTrio and GeparQuattro trials with operable or locally advanced tumors received anthracycline-taxane (+/- anti-HER2-based neoadjuvant chemotherapy and were classified as having unifocal (one lesion detected by physical examination, sonography, and mammography ± MRI), multifocal (≥2 lesions in one breast quadrant), or multicentric (≥ 1 lesion in ≥2 quadrants) disease. Breast conservation was allowed when tumor-free margins were achieved.

#### Results

Tumors of 3,562 participants were classified as unifocal (N = 2793; 78.4%), multifocal (N = 429; 12.0%), and multicentric (N = 340; 9.5%). Breast conservation was performed in 71.7%, 56.2%, and 35.1%, respectively (P < 0.0001). At surgery pathological complete response (pCR) rates were 18.7%, 14.1%, and 14.9%, respectively (P = 0.047). After median follow up of 46.3 months locoregional and distant-relapse-free survival were worse in patients with multicentric disease versus uni- or multifocal disease treated with mastectomy (P = 0.007 and 0.061, respectively), but not when treated by breast conservation (P = 0.634 and 0.650, respectively). Patients with pCR showed a low locoregional relapse rate irrespective of focality (P = 0.713) but a higher distant relapse rate in case of multicentric disease (P = 0.003). Prognostic factors for locoregional recurrence in multivariable analysis were tumor and nodal status at surgery, grading, hormone-receptor status, and type of surgery, but not focality of the tumor. Overall survival was not statistically different through all focality groups.

#### Conclusions

Breast conservation for multifocal or multicentric breast cancer after neoadjuvant chemotherapy is feasible and seems not to impair outcome if tumor-free margins were achieved.

#### Disclosure

All authors have declared no conflicts of interest.

## 322PDA TREATMENT-INTERACTION ANALYSIS BALANCING PATHOLOGICAL COMPLETE RESPONSES (PCR) AND CARDIOTOXICITY OF SINGLE-(S)/DUAL-(D) HER2 INHIBITION AND NEOADJUVANT CHEMOTHERAPY (CT) BACKBONE IN OPERABLE/LOCALLY ADVANCED BREAST CANCER (O/LABC) PATIENTS

### Abstract

#### Purpose

Given the toxicity drawbacks potentially related to the combination of anthracylines and a D-HER2 inhibition, a treatment interaction analysis of the available randomized trials was accomplished.

#### Methods

pCR (breast + axilla), breast conserving surgery (BCS), grade 3-4 neutropenia/cardiotoxicity, and febrile neutropenia (FN), events were extracted from papers/presentation and cumulated according to a random-effect model; 95% confidence intervals (CI) were derived. A sensitivity analysis according to S/D HER2 inhibition, hormonal receptors (HRs) and CT (anthracyclines-taxanes: anthra-TAX; TAX alone) was accomplished to test for interaction. Absolute differences (AD) with 95% CIs, and the number of pts needed to treat/harm (NNT/NNH) for 1 to benefit were calculated to derive the Likelihood of being Helped or Harmed (LHH).

#### Results

8 trials (2092 pts) with 1955 pts treated with anti-HER2 therapy (Trastuzumab, Lapatinib and Pertuzumab), were gathered; pCR rates according to HER2 inhibition follow: pCR rates were significantly higher in the HRs negative population, regardless of the HER2 inhibition and CT backbone (Neg. vs. Pos. Anhtra-TAX [S] AD 9.4%, p = 0.002; TAX [S] AD 15.3% p < 0.0001; TAX [D] AD 28%, p < 0.0001). With regard to CT, a significant interaction (p < 0.0001) in favour of adding Anthra to TAX was found in the context of S-HER2 inhibition subgroup with regard to pCR (AD 15.4%) and BCS (AD 10.8%) with no interaction in the D-subgroup. No significant differences in FN and cardiotoxicity were found according to HER2 inhibition and CT. In the Anhtra-TAX [S] population weighted for cardiotoxity, LHH was 77.

CT HER2-Inhibition Rates (95% CI) Interaction (p) NNT
Anthra-TAX 37.0 [34.0, 40.0] 0.22 2.7
44.3 [33.3, 55.2]  2.2
TAX 21.7 [18.1, 25.3] <0.0001 4.6
42.4 [36.4, 48.5]  2.3
CT HER2-Inhibition Rates (95% CI) Interaction (p) NNT
Anthra-TAX 37.0 [34.0, 40.0] 0.22 2.7
44.3 [33.3, 55.2]  2.2
TAX 21.7 [18.1, 25.3] <0.0001 4.6
42.4 [36.4, 48.5]  2.3

#### Conclusions

On the basis of the available data, anthracyclines should be considered for O/LABC patients receiving TAX-based CT plus HER2 inhibition, given the likelyhood 70-times greater to achieve pCR than to be harmed by clinically meaningful cardiotoxicity.

#### Disclosure

All authors have declared no conflicts of interest.

## 323PDPHASE III TRIAL EVALUATING THE ADDITION OF BEVACIZUMAB TO ENDOCRINE THERAPY AS FIRST-LINE TREATMENT FOR ADVANCED BREAST CANCER: THE GEICAM/GBG LEA STUDY. SAFETY ANALYSIS

### Abstract

#### Background

We designed the randomized phase III LEA study of first-line bevacizumab in combination with endocrine therapy, to address the hypothesis that anti-VEGF treatment can prevent resistance to endocrine therapy in patients (pts) with advanced breast cancer sensitive to such treatment.

#### Methods

A multicentre, binational, randomised, open label, phase III study investigated the addition of Bevacizumab (B) 15mg/kg every 3 weeks to an endocrine therapy (ET) with letrozole or fulvestrant (250mg/4 weeks) as first-line therapy in metastatic breast cancer. Postmenopausal pts with HER2-negative and hormone-receptor-positive breast cancer were eligible. The primary objective was to compare progression-free survival (PFS) in the treatment arms. Secondary objectives were overall survival, time to treatment failure, overall response rate, response duration, clinical benefit rate and safety. The recruitment was completed in September 2011. Efficacy analysis will be triggered after 270 events.

#### Results

From 11/2007 to 11/2011, 380 pts with ER/PgR+ and HER2-tumours were randomised to ET ± B. 38 patients received fulvestrant. Baseline characteristics were well balanced. Median age was 65 years (38-85). 27% had visceral metastases. 36% patients had a prior AI. 110 pts are still on treatment. So far safety data from 335/380 patients are available. Overall 58 serious adverse events in 46 patients were reported 44 in ET-B of which seven were fatal (thromboembolic event x2, hypertension, heart failure, sudden death, liver failure, cerebellum infarction) but only 2 drug related to bevacizumab by the investigator and 14 in ET arm. The main side effects any grade per patient ET-B vs ET were as follows anemia 76% vs 44%, p < 0.001; fatigue 50%vs 31%, p = 0.001; hemorrhage 19% vs 2%, p < 0.001; hypertension 55% vs 12%, p < 0.001; proteinuria 21%vs 3%, p < 0.001; thrombosis grade 3-4, 2.3%vs 0%, p = 0.057.

#### Conclusions

LEA is the first study to explore the use of an anti-angiogenic drug in combination with an endocrine treatment. The main side effects are of grade 1-2. Final safety data will be presented at the meeting.

#### Disclosure

All authors have declared no conflicts of interest.

## 324PDEFFICACY AND SAFETY OF EVEROLIMUS IN POSTMENOPAUSAL WOMEN WITH ADVANCED BREAST CANCER (BOLERO-2): EFFECT OF VISCERAL METASTASES AND PRIOR ENDOCRINE THERAPY

### Abstract

#### Introduction

Postmenopausal women with hormone receptor-positive (HR+) breast cancer (BC) who relapse or progress on a nonsteroidal aromatase inhibitor (NSAI) have limited treatment options. At a median follow-up of 18 months, BOLERO-2 demonstrated that everolimus (EVE), an oral mammalian target of rapamycin (mTOR) inhibitor, plus exemestane (EXE), a steroidal aromatase inhibitor, prolonged progression-free survival (PFS) compared with EXE alone in this setting (7.8 vs 3.2 mo, respectively; hazard ratio [HR] = 0.45 [95% confidence interval (CI) = 0.38, 0.54]; log-rank P < .0001).

#### Methods

BOLERO-2 is a phase 3, double-blind, randomized study that compared EVE (10 mg/d) + EXE (25 mg/d; n = 485) versus placebo (PBO) + EXE (n = 239) in postmenopausal women with advanced HR+ HER2 BC who had progression or recurrence after NSAI therapy. The primary endpoint was PFS by local investigator assessment. In view of the limited efficacy of endocrine therapies in patients with visceral involvement, EVE + EXE was evaluated in patient subgroups defined by the presence of visceral metastases (including lung, liver, spleen, pleural effusions, pericardial effusion, peritoneum, ascites, ovary and central nervous system).

#### Results

At a median follow-up of 18 months, adding EVE to EXE prolonged median PFS versus PBO + EXE in patients with visceral metastases (n = 406; 6.8 vs 2.8 mo, respectively; HR = 0.47 [95% CI = 0.37, 0.60]). Similarly, EVE + EXE extended PFS in patients without visceral metastases (n = 318; 9.9 vs 4.2 mo for PBO + EXE; HR = 0.41 [95% CI = 0.31, 0.55]). Patients with visceral involvement had shorter PFS compared with patients with bone-only disease regardless of treatment. Nonetheless, PFS improvements with EVE + EXE were similar in patients with visceral and bone-only disease (n = 151; 12.9 vs 5.3 mo for PBO + EXE; HR = 0.33 [95% CI = 0.21, 0.53]).

#### Conclusions

Adding EVE to EXE markedly extended PFS by ≥ 4 mo among patients with advanced HR+ HER2 BC regardless of the presence of visceral metastases.

#### Disclosure

M. Campone: M. Campone is a consultant to and has received honoraria from Novartis.

S. Noguchi: S Noguchi has received grant support and honoraria from AstraZeneca, Chugai, Pfizer, sanofi-aventis, GSK, Taiho, Novartis, and Takeda.

K. Pritchard: Consult sanaven AZE Roche PFE NVR ABR AMG GSK res funding NCICCT Grp contracted AZE BMS SanAven AMG PFE NVR GSK & OrthoBio honoraria & Spkr B SanAven AZE PFE Roche NVR & AMG paid expert test SanAven AZE & GSK AdCom SanAven AZE Roche PFE NVR GSK & AMG.

H. Rugo: H. Rugo has received grant support from Novartis, Pfizer, and Merck, and has received travel support from Novartis.

G.N. Hortobagyi: Member of the Scientific Advisory Board of Allergan, consultant to Allergan, Novartis, Genentech, and sanofi-aventis, received grant support from Novartis, and travel expense reimbursement from Novartis, Genentech, and sanofi-aventis.

J. Baselga: J Baselga is a consultant to Novartis, Roche, Merck, sanofi-aventis, Verastem, Bayer, Chugai, Exelixis, Onyx, Constellation.

A. Panneerselvam: Employee of Novartis with stock/stock options.

T. Taran: Employee of Novartis with stock/stock options.

T. Sahmoud: An employee of Novartis with stock/stock options.

M. Piccart: Board PharmaMar, consultant sanofi-aventis, Amgen, BMS, GSK, Boehringer, Roche, & Bayer, & grant Pfizer, Amgen, Bayer, Boehringer, BMS, GSK, Roche, & sanofi-aventis, honoraria Bayer, BMS, GSK, Boehringer, Roche, Amgen, sanofi-aventis, & AZE.

## 325PDPHASE I/II TRIAL OF ABIRATERONE ACETATE (AA) IN ESTROGEN RECEPTOR (ERα) OR ANDROGEN RECEPTOR (AR) POSITIVE METASTATIC BREAST CANCER (MBC)

### Abstract

#### Background

Abiraterone irreversibly inhibits 17-hydroxylase/c-17-20 lyase (CYP17), reducing androgen and estrogen levels and improves overall survival from castration resistant prostate cancer. We hypothesized that: A) Postmenopausal ERα+ MBC continue to be ERα + /AR driven; and, B) Postmenopausal ERα- AR+ MBC can be driven by AR.

#### Methods

This Phase I/II trial of AA with hydrocortisone evaluated tolerability, pharmacokinetic (PK)-pharmacodynamic (PD) profile and anti-tumor activity. Two parallel but non-randomized Phase II arms utilized a Gehan design (95% probability of detecting a 24wk clinical benefit rate (CBR, partial response [PR] + stable disease) of > 20%; 14 patients [pts] in the first stage; 11 in the second stage for each arm). Prior therapy with ≥ 2 lines of endocrine therapy (for ERα+ arm); ≥ 1 line of chemo (for AR + ERα- arm); and prior trastuzumab if HER2-positive was required. ERα and AR positivity was defined as immunoreactivity in ≥ 1% cells.

#### Results

In the phase I study, daily dosing of AA was well tolerated with variable PK at all dose levels. PD studies of CYP17 blockade demonstrated suppression of circulating estradiol and androgen levels below the limit of assay detection with 1000mg and 250-2000mg AA respectively; 1000mg was selected for Phase II evaluation. In the ERα+ arm, 6 pts (Phase I) and 25pts (Phase II) received 1000mg AA, of whom 4 were HER2-positive. The median age (range) was 60 (46-80), prior lines of hormonal and chemotherapy were 3 (2-4) and 2 (0-5) respectively. There was 1 partial response (PR) lasting 14m in a pt who had received 4 and 5 lines of hormonal and chemotherapy respectively. Median progression-free survival was 11wk. CBR at 24wk was 21%. In the AR + ERα- arm, recruitment is ongoing. Hypokalaemia easily managed by hydrocortisone administration, was the commonest drug related adverse event (AE).

#### Conclusion

AA was well tolerated and merits further evaluation in MBC. Cancer Research UK (Drug Development Office) Sponsored and funded the trial. Johnson & Johnson provided AA.

#### Disclosure

M. Dowsett: The Institute of Cancer Research has a commercial interest in Abiraterone. Mitch Dowsett is an employee of the Institute of Cancer Research, which has a ‘Rewards to Inventors’ scheme and is a recipient of the scheme.

J.S. de Bono: The Institute of Cancer Research has a commercial interest in Abiraterone. Prof de Bono is an employee of the Institute of Cancer Research, and recipient of a ‘Rewards to Inventors’ scheme. He is also a consultant for Johnson and Johnson.

All other authors have declared no conflicts of interest.

## 326PDPACLITAXEL-BASED VERSUS DOCETAXEL-BASED REGIMENS IN METASTATIC BREAST CANCER: A SYSTEMATIC REVIEW AND META-ANALYSIS OF RANDOMIZED CONTROLLED TRIALS

### Abstract

#### Background

Docetaxel and paclitaxel show significant clinical activity in metastatic breast cancer (MBC) and have been approved for MBC by U.S. Food and Drug Administration, but it is still unclear whether paclitaxel-based regimen improves outcome over docetaxel-based regimen in patients with MBC. We therefore performed a meta-analysis of randomized controlled trials that compared paclitaxel-based with docetaxel-based regimens in MBC.

#### Method

We systematically searched for randomized controlled trials that comparing paclitaxel-based with docetaxel-based regimens in patients with metastatic breast cancer in PUBMED, EMBASE and Cochrane databases. Abstracts presented at the conference were also searched. The primary endpoint was overall survival (OS). Secondary endpoints were progression free survival (PFS), time to progression (TTP), overall response rate (ORR), and grade 3 or 4 toxicity. Data were extracted from the studies by two independent reviewers. The meta-analysis was performed by Stata version 12.0 software (Stata Corporation, College Station, TX, USA).

#### Results

Seven eligible trials involving 1694 patients with MBC were selected. Our meta-analysis results showed that paclitaxel-based regimen was comparable to docetaxel-based regimen in terms of OS (HR: 0.87, 95%CI: 0.60-1.27, p = 0.476),PFS(HR:0.76,95%CI:0.58-1.00, p = 0.052), TTP(HR: 1.13, 95%CI: 0.81-1.58, p = 0.459), and ORR(RR:1.01, 95%CI: 0.88-1.15, p = 0.915), but less grade 3 or 4 adverse events including anemia(RR:0.64,95%CI:0.44-0.94,p = 0.023), neutropenia (RR:0.74,95%CI:0.58-0.93, p = 0.011), febrile neutropenia (RR:0.38, 95%CI:0.15-0.96, p = 0.041), thrombopenia (RR:0.62, 95%CI: 0.41-0.96,p = 0.033),mucositis (RR:0.082, 95%CI:0.025-0.27,p = 0.000),diarrhea(RR:0.194,95%CI:0.081-0.47,p = 0.000) and fatigue(RR:0.434,95%CI:0.20-0.96,p = 0.039) were observed in paclitaxel-based regimen.

#### Conclusion

The present systematic review and meta-analysis demonstrate that both taxanes-based regimens show comparable efficacy for patients with MBC, and paclitaxel-based regimen is associated with less toxicity and better tolerability, especially in older patients and when used in weekly regimens.

#### Disclosure

All authors have declared no conflicts of interest.

## 327PDBODY MASS INDEX AND PROGNOSIS OF WOMEN WITH METASTATIC BREAST CANCER

### Abstract

Obesity is a well known risk factor for the development of breast cancer and an adverse prognostic factor in those women who have already been diagnosed with breast cancer. However, the effect of obesity on the prognosis of MBC women has not been explored so far. The aim of this study was to evaluate the influence of Body Mass Index (BMI) on the prognosis of MBC patients receiving first line chemotherapy.

#### Methods

The relationship between BMI (kg/m2), and progression free (PFS) or overall survival (OS) was assessed in 698 MBC patients enrolled in 3 clinical trials of first line chemotherapy, conducted by the same collaborative group. Conventional WHO BMI definitions were applied: normal 18.5-24.9 kg/m2, overweight 25-30 kg/m2, obese >30 kg/m2. PFS and OS were calculated by Kaplan-Meier estimation; multivariate Cox analysis was performed adjusting for age, menopausal status, PS and study.

#### Results

Information on BMI at the time of study entry, was available on 489 women. All patients received first line chemotherapy regimens including anthracyclines and taxanes, either in sequence or combination. Median follow up was 18 months (range 0.4 to 88.3). Overall, 40.3% of the patients were normal, 38.2% were overweight and 21.5% were obese. Median age was 57 years (range 25 to 73); PS was 0 in 93% of the patients. A non significant trend toward improved PFS was found in overweight women: median PFS was 13.1 months (95% CI 11.0-15.2) in BMI 25-30 versus 10.8 months (8.9-12.6) in BMI < 25 and 12.2 (95% CI 10.4-14.0) in BMI >30, p = 0.2. Median OS was 32.0 months (95% CI 24.9-39.1) in BMI < 25 versus 32.9 (95% CI 27.7-38.1) in BMI 25-30 and 30.7 (95% CI 24.3-37.0) in BMI >30, p = 0.8. By multivariate analysis, none of the considered variables was significantly associated with differences in PFS and/or OS.

#### Conclusions

In this study BMI at baseline was not significantly associated with the outcome of MBC patients treated with first line chemotherapy; however, a non significant trend for an improved PFS was observed in overweight women as compared to normal weight and obese patients. These data suggest that overweight should not be regarded as an adverse prognostic factor patients with MBC.

#### Disclosure

All authors have declared no conflicts of interest.

## 328PDPROGRESSION-FREE SURVIVAL AS A SURROGATE FOR OVERALL SURVIVAL IN METASTATIC BREAST CANCER

### Abstract

#### Objectives

Progression-free survival (PFS) is frequently used to establish the clinical efficacy of anti-cancer drugs. However, the surrogacy of PFS for overall survival (OS) remains a matter of uncertainty in metastatic breast cancer (mBC). Therefore, the aim of this study was to assess the relationship between PFS and OS in mBC using a trial-based approach.

#### Methods

A systematic literature review was conducted using the PICO method: Population consisted of women with mBC; Interventions and Comparators were standard treatments for mBC or best supportive care; Outcomes of interest were median PFS or TTP and median OS. A correlation analysis between median PFS and OS was first performed and subgroup analyses were conducted to explore possible reasons for heterogeneity. Then, the relationship between the treatment effect on PFS and OS was assessed. The treatment effect on PFS and OS was quantified by the absolute difference of median values. Linear regression analysis was also conducted to predict the effects of a new anticancer drug on OS on the basis of its effects on PFS.

#### Results

A total of 5041 studies were identified and 144 fulfilled the eligibility criteria. Selected studies included 315 treatment arms, which represents 43,459 patients with mBC. There was a significant relationship between median PFS and median OS across included trials (r = 0.428; p < 0.01). The correlation between median PFS/TTP and median OS was higher for studies evaluating chemotherapy alone (r = 0.575; p ≤ 0.01) or in combination (r = 0.632; p ≤ 0.01) compared with those evaluating hormone therapy (non-significant r). The correlation coefficient for the treatment effect on PFS and OS was estimated at 0.427 (p < 0.01). The linear regression equation was: ΔOS = -0.088 (95% CI, -1.347 to 1.172) + 1.753 (95% CI, 1.307 to 2.198) * ΔPFS, with a proportion of variation explained (R2) of 0.86. Results of the regression analysis predict that a difference in median PFS/TTP of 5, 10, 15, and 20 months would translate into a difference in median OS of 8.7, 17.4, 26.2, and 35.0 months respectively.

#### Conclusion

The present findings point toward a statistically significant correlation between PFS and OS in the context of mBC and support the surrogacy of PFS for OS in this cancer setting.

#### Disclosure

All authors have declared no conflicts of interest.

## 329PPATIENT-REPORTED OUTCOMES (PROS) FROM EMILIA, A PHASE 3 STUDY OF TRASTUZUMAB EMTANSINE (T-DM1) VS CAPECITABINE AND LAPATINIB (XL) IN HER2-POSITIVE LOCALLY ADVANCED OR MBC

### Abstract

#### Background

The antibody-drug conjugate T-DM1 combines the antitumor activities of trastuzumab (T) with intracellular delivery of the cytotoxic agent DM1. In the EMILIA study, the efficacy and safety of T-DM1 was compared to XL. Here we report the PRO results.

#### Methods

Pts with centrally confirmed HER2-positive MBC and prior therapy with T and a taxane were randomized to T-DM1 or XL administered in 21-day cycles. Pts completed the FACT-B, a 37-item questionnaire composed of 5 subscales at baseline and on day 1 every 2 cycles until disease progression (PD), 6 wks after PD and every 3 mos thereafter. A 24-item subset of FACT-B, the Trial Outcome Index (TOI) provides a summary of physical and functional well-being and breast cancer-specific symptoms. Time to FACT-B TOI worsening (ie, ≥5 point decrease in TOI), the main PRO analysis, was assessed by Kaplan-Meier methods and a Cox model in female pts with baseline and ≥1 post-baseline TOI score. An exploratory PRO end point was the incidence of symptoms measured by the Diarrhea Assessment Scale (DAS), in which pts rated diarrhea symptoms in 4 domains (frequency, urgency, discomfort, stool consistency) on a 4-point scale at baseline and on day 1 of each cycle until PD.Results: Pts treated with T-DM1 had longer PFS (9.6 vs 6.4 months; HR=0.650; P<0.0001) and fewer grade 3/4 AEs (41% vs 57%) than those treated with XL. These gains in efficacy and safety were accompanied by improvements in patient-reported health status in the 450 T-DM1 and 445 XL PRO-evaluable pts. T-DM1 pts maintained stable FACT-B TOI scores longer than XL pts (time to TOI worsening 7.1 vs 4.6 months; HR=0.796). A comparison in FACT-B subscales between treatment arms will be presented. From cycle 2 through 8, more pts in the XL arm than in the T-DM1 arm reported diarrhea symptoms such as having >1 stool per day (57% vs 30%); loose or watery stools (70% vs 37%); somewhat to very urgent stools (54% vs 26%) and mild-moderate to very severe abdominal discomfort with bowel movements (45% vs 25%).

#### Conclusions

Improvements in efficacy and safety were accompanied by, clinically significant benefits in health-related quality of life in pts treated with T-DM1 vs XL.

#### Keywords

diarrhea, T-DM1, patient-reported outcomes, HER2-positive.

#### Disclosure

V. Diéras: *Compensated consultant/advisory relationship with Genentech/Roche, Novartis, Sanofi, Amgen, Clovis, Pfizer, GSK *Honoraria from Genentech/Roche, Novartis, Sanofi, Amgen, Clovis, Pfizer, GSK J. Sohn: Research funding from Roche, Amgen and GSKS. Hurvitz: *Support for study-related travel from Roche *Research funding for institution from Roche D. Lalla: Genentech employee, owns Roche stockL. Fang: Genentech employee, owns Roche stockE. Guardino: Genentech employee, owns Roche stock All other authors have declared no conflicts of interest.

## 331PIMPACT OF RESPONSE SHIFT ON TIME TO QUALITY OF LIFE SCORES DETERIORATION IN BREAST CANCER PATIENTS: IS IT TIME TO MOVE FOR QOL RECIST CRITERION?

### Abstract

#### Background

Time to quality of life (QoL) score deterioration (TD) is a method of longitudinal QoL data analysis that has been proposed for breast cancer (BC) patients (Hamidou et al Oncologist 2011). As for RECIST criteria, the optimal definitions dealing with reference should be explored. This study aims to study the impact of changes in internal standards (CIS) of response-shift (RS) and the influence of baseline QoL expectancies on TD.

#### Methods

A prospective multicenter study including all women hospitalized for a primary BC was conducted. The EORTC-QLQ-C30 and BR-23 questionnaires were used to assess the QoL at baseline, at the end of 1st hospitalization, and 3 and 6 months after. CIS was investigated by the then-test method. QoL expectancy was assessed at baseline using Likert scale. Deterioration was defined as a decrease in QoL scores reaching at least the mean difference identified as minimal clinically important difference (MCID) using Jaeschke's transition question. Sensitivity analyses were done using the then-test score as reference score, and considering 5 and 10 points as MCID. TD was estimated using Kaplan-Meier method. Cox regression analyses were used to identify factors influencing TD.

#### Results

From February 2006 to February 2008, 381 women were included. For role functioning dimension, the median TD increased from 3.2 months [95% CI: 3.1-3.36] to 4.76 months [3.3-6.2] when adjusting on CIS. For body image when adjusting on CIS, sentinel lymph node biopsy became significantly associated with longer TD (HR: 0.64[0.43-0.94]) as compared to axillary lymph node dissection, radiotherapy to a shorter TD (HR: 0.63[0.42-0.95] and the type of surgery had no effect on TD. For global health, cognitive and social functioning dimensions, patients expecting deterioration in their QoL had a significantly shorter TD. For fatigue and breast symptom scales, patients expecting no change had a significantly shorter TD, as compared to patients expecting an improvement. Sensitivity analyses using a MDCS of 5 or 10 confirmed these results.

#### Conclusions

Our results suggest that it would be more accurate to take into account CIS component of RS as well as QoL expectancies to estimate TD of QoL scores in patient with BC.

#### Disclosure

All authors have declared no conflicts of interest.

## 332PHEALTH STATE UTILITY DETERMINATION IN ADVANCED STAGE BREAST CANCER PATIENTS

### Abstract

#### Background

Quality-adjusted Time Without Symptoms of disease and Toxicity (Q-TWiST) survival analysis where time without disease progression is rewarded while time without disease progression but with toxicities is penalized by applying a utility weight is increasingly common in health technology assessments of cancer treatments. AIM: We assess women's preferences for health states specific to advanced stage breast cancer including a baseline diagnoses of ABC, treatment response, no treatment response but no disease progression, disease progression and hormonal therapy specific toxicities.

#### Method

Using FACT-B quality of life data from a randomized clinical trial in ABC univariate statistics were obtained for each item for each health state without regard to treatment. These item scores were paired to the actual narrative in the FACT-B to construct health state narratives consisting of physical, social, emotional, functional well-being and additional concerns content domains. The order of the content domains was varied to prevent order bias. One hundred and nine peri- or post-menopausal women were recruited and interviewed by a single woman in their age group using visual analogue and standard gamble techniques. Univariate and multivariate analyses were performed to control for age, marital status, menopausal status and whether the interviewee has had breast cancer or any other cancer.

#### Results

Of the 109 recruited 100 women completed the interview, mean age was 55.76 years, 64% were postmenopausal, 11% had breast cancer previously and 16% had another type of cancer previously. Multiple regressions results for the VAS scores yielded values of 51.8 (p < 0.01) for baseline ABC diagnosis, 82.5 (p < 0.01) for treatment response, 57.5 (p < 0.01) for no response no progression and 38.4 (p < 0.01) for disease progression. The SG results were 0.64 (p < 0.01), 0.76 (p < 0.01), 0.67 (p < 0.01), and 0.50 (p < 0.01), respectively. Women who previously had breast cancer related the health states consistently higher (p < 0.05) in the VAS and SG analyses. The trade-off between a chance of treatment response yet the possibility of toxicity yielded a utility score of 0.34 (p < 0.01).

#### Conclusion

These VAS and SG scores can be used to better assess women's preference for treatment options in ABC.

#### Disclosure

W.R. Simons: This research was conducted independent of Eisai Inc.

## 333PCOST EFFECTIVENESS OF ADJUVANT CYCLOPHOSPHAMIDE CONTAINING REGIMES TO AT IN THE TREATMENT OF BREAST CANCER IN GERMANY

### Abstract

#### Objectives

58,000 women in Germany are diagnosed with breast cancer each year. Anthracycline-containing regimes - in addition to other treatment options - have been standard adjuvant breast cancer regimens. The purpose of this analysis was to estimate the cost-effectiveness of AT (doxorubicin, docetaxel) compared with AC (doxorubicin, cyclophosphamide), CMF (cyclophosphamide, methotrexat, 5-fluoruracil) and FEC (5-fluoruracil, epirubicin, cyclophosphamide) administered as adjuvant therapy to women with node-positive breast cancer in Germany.

#### Methods

We developed a multi-country Cost-Utility-Model to simulate the long-term consequences from initiation of adjuvant chemotherapy over 10 years. Markov modelling techniques were used to estimate incidence of complications during chemotherapy (febrile neutropenia, chemotherapy induced nausea/vomiting, dose-reduction, dose-delay, other grade 3/4 adverse events) and long-term consequences like local or distant relapse, secondary acute myelogenous leukemia, chronic heart failure and death. Monte-Carlo-simulation accounted for uncertainty. Probabilities were derived from clinical and epidemiological studies; direct costs from published sources from the payer's perspective. QALYs, life years and costs were discounted at 3% p.a.

#### Results

Over a 10-year timeframe, costs and outcomes associated with AT amounts to 21,055.91 € and 6.3 QALYs (7 LYs). Costs associated with AC are 17,018.04 €, while outcomes are comparable to AT. The cost saving potential associated to AC vs. AT amounts to 4,037.87 € per patient. Costs and outcomes associated with CMF are 17,790.42 € and 6.3 QALYs (6.94 LYs), leading to a cost saving potential of € 3,265.49 with CMF vs AT. FEC associated total costs are 18,471.84 €. Quality-adjusted life expectancy increases to 6.5 years, which represents a QALY gain of 0.2 QALYs over 10 years vs. AT. The increase in life expectancy without quality adjustment amounts to 7.4 years and leads to 0.4 LYs gained with FEC versus AT. Accordingly, FEC dominates AT.

#### Conclusion

Cyclophosphamide-based regimens (FEC, AC and CMF) demonstrate a better performance from cost-effectiveness perspective vs. AT (doxorubicin, docetaxel).

#### Disclosure

All authors have declared no conflicts of interest.

## 334PHEALTH-RELATED QUALITY OF LIFE (QOL) IN METASTATIC BREAST CANCER PATIENTS TREATED WITH EVEROLIMUS AND EXEMESTANE VERSUS EXEMESTANE MONOTHERAPY

### Abstract

#### Background

The phase 3 BOLERO-2 study randomized 724 patients with hormone-receptor–positive (HR+) advanced breast cancer and recurrence or progression during/after nonsteroidal aromatase inhibitor therapy to everolimus (EVE) plus exemestane (EXE) or EXE and placebo (PBO). Interim analysis after 12 months' median follow up demonstrated that EVE + EXE significantly improved progression-free survival (PFS) vs EXE + PBO, with a higher rate of grade 3/4 adverse events but no deterioration in QOL. We report here on additional post hoc analyses of patient-reported QOL.

#### Methods

Using the EORTC QLQ-C30 questionnaire and QLQ-BR23 module, QOL was assessed at baseline and every 6 weeks thereafter until progression or discontinuation. The QLQ-C30 consists of 30 items combined into 15 subscales, including Global Health Status (GHS). The BR23 consists of 23 items specific to breast cancer combined into symptom and functioning subscales, including breast symptom (BS) and arm symptom (AS) scales. Average difference in change from baseline between treatment groups was evaluated using linear mixed models with several adjustment covariates. Sensitivity analysis was conducted using pattern-mixture models to examine the effect of study dropout on or before week 24. Treatments were compared using differences of least squares mean (LSM) changes from baseline at each timepoint and overall.

#### Results

Linear mixed models indicated no statistically significant overall difference between EVE + EXE and EXE + PBO for GHS (LSM difference = –2.0; 95% CI = –4.8, 0.9), breast symptoms (LSM difference = 0.3; 95% CI = –2.8, 2.4), or arm symptoms (LSM difference = –0.2; 95% CI = –2.8, 2.4). Pattern-mixture models indicated that patients who dropped out early had worsening QOL over time in both treatment arms; EVE + EXE patients who did not drop out early had stable QOL, whereas EXE + PBO was associated with worsening QOL over time.

#### Conclusions

These additional analyses from the BOLERO-2 study confirm that compared with EXE alone, EVE + EXE improved PFS without adversely impacting QOL in patients with HR+ advanced breast cancer progressing despite nonsteroidal aromatase inhibitors.

#### Disclosure

J.T. Beck: Has received grant support from Pfizer and Novartis.

H. Rugo: Has received grant support from Pfizer and Merck, and has received travel support from Novartis.

J. Baselga: Is a consultant to Novartis, Roche, Merck, sanofi-aventis, Verastem, Bayer, Chugai, Exelixis, Onyx, and Constellation.

T. Taran: Is an employee of Novartis with stock options.

L. Bennett: Is an employee of RTI Health Solutions, which contracted with Novartis for data analysis services.

J. Ricci: Is a consultant to Novartis.

T. Sahmoud: Is an employee of Novartis with stock options.

G.N. Hortobagyi: Member of the Sci Ad Board of Allergan; consultant to Allergan, Novartis, Genentech, and sanofi-aventis; has received grant support from Novartis; travel expense reimbursement from Novartis, Genentech, and sanofi-aventis.

All other authors have declared no conflicts of interest.

## 335PDYSPHONIA AS A PREVIOUSLY UNREPORTED SIDE EFFECT OF BEVACIZUMAB TREATMENT IN PATIENTS WITH METASTATIC BREAST CANCER

### Abstract

#### Introduction

Bevacizumab is a humanized monoclonal antibody directed against vascular endothelial growth factor A and has been approved for the treatment of several metastatic tumours. There is considerable heterogeneity in the response to treatment with bevacizumab, both in effectiveness and in toxicity. Here we describe a previously unreported adverse drug reaction (adr) in pts with MBC treated with bevacizumab.

#### Methods

In a teaching hospital in the Netherlands (from Sep 2009 to Jul 2011), 32 consecutive pts with MBC treated with chemotherapy and bevacizumab were registered in a retrospective database. TNM stage, comorbidities, concomitant medication, prior treatment for the primary tumour, date of metastatic disease, prior treatment for metastatic disease and toxicities were recorded. The WHO global individual case safety report database, VigiBase, contains summaries of suspected spontaneous case reports summated by health care professionals and pts to national pharmacovigilance centres . As of May 2010, VigiBase contained >5 million case reports. . We searched the VigiBase extraction of Dec 2011 for dysphonia. Reporting odds ratios (ROR) were calculated for the occurrence of dysphonia compared with other adr for bevacizumab and paclitaxel.

#### Results

In total, 9/32 pts (28%) reported dysphonia during treatment with bevacizumab and 5/9 pts underwent ENT examination. In several pts marked oedema of the vocal cords and/or chronic laryngitis were found. As of Dec 2011, 6,880,361 reports were available in VigiBase, of which16,239 were related to dysphonia. For bevacizumab there were 51 reports for dysphonia and 46,041 reports for other adr. Corresponding figures for all other drugs were 22,108 reports for dysphonia and 25,151,628 reports for other adverse effects: ROR of 1.26 (95% CI: 0.95-1.66). For paclitaxel there were 45 reports for dysphonia and 85,988 reports for other adr. Corresponding figures for all other drugs were 22,114 reports for dysphonia and 25,111,681 reports for other adverse effects: ROR of 0.59 (95% CI: 0.44-0.80)meaning that the risk of angiooedema is significantly higher in bevacizumab users compared to paclitaxel users.

#### Conclusion

Dysphonia is a previously unreported side-effect in pts with MBC treated with bevacizumab and paclitaxel.

#### Disclosure

All other authors have declared no conflicts of interest.

## 336PPATTERNS OF CLINICAL MANAGEMENT AND RESOURCE UTILIZATION FOR POSTMENOPAUSAL HORMONE-RECEPTOR–POSITIVE HER2-NEGATIVE (HR+ HER2–) ADVANCED BREAST CANCER (ABC) IN EUROPE

### Abstract

#### Objective

To understand treatment patterns and quantify resource utilization of HR+ HER2 aBC, with the overall aim of comparing costs and disease burden as patients progress from hormonal therapy (HT) to chemotherapy (CT).

#### Methods

We conducted a chart audit in France, Germany, The Netherlands, Belgium, and Sweden of 375 living and deceased postmenopausal female patients (75 per country) diagnosed with ER+ HER2 aBC in the past 4 years. Patients were required to have progressed on ≥ 1 line of prior HT either in the adjuvant or advanced setting and to have completed ≥ 1 line of CT treatment (minimum 2 full cycles) in the aBC setting. The chart audit was completed online using a standardized form developed with the assistance of European academic physicians, pharmacy directors, and hospital administrators. Participation was sought from 25 oncologists per country, except in Germany (15 oncologists and 10 gynecologists to reflect local clinical practice). Data collection was compliant with European and country market research regulations.

#### Results

Our report details the patient care pathway, CT side effects, and resource utilization in the inpatient and outpatient settings throughout the continuum of aBC care. Preliminary analyses indicate that 55% of HR+ HER2 aBC patients are first treated with HT and switch to CT after 1.5 lines of HT. This switch is primarily influenced by the extent (56%) and progression rate (36%) of metastases. The switch from HT to CT is associated with increased resource utilization and the associated costs of treating aBC. In addition to cost of drug therapies, the main drivers of cost are treatment for CT side effects (chiefly febrile neutropenia and diarrhea) and related hospitalization events. CT side effects that have the greatest impact on the overall disease burden of aBC include alopecia, nausea, vomiting, fatigue, and peripheral neuropathy.

#### Conclusions

Our results highlight the increased costs and disease burden for postmenopausal ER+ HER2 aBC patients treated with CT versus HT.

#### Disclosure

G. Jerusalem: Consultant, Novartis Pharmaceuticals Corporation.

N. Marinsek: Consultant, Novartis Pharmaceuticals Corporation.

J. Ricci: Advisor, consultant, Novartis Pharmaceuticals Corporation.

J. Etchberger: Consultant, Novartis Pharmaceuticals Corporation.

R. Degun: Consultant, Novartis Pharmaceuticals Corporation.

G. Benelli: Employees of Novartis with stock/stock options.

S. Saletan: Employee of Novartis with stock/stock options.

F. André: Financial support from sanofi-aventis, Novartis, Roche, AstraZeneca

## 337PEFFICIENCY FRONTIER ANALYSIS (EFA) OF METASTATIC BREAST CANCER (MBC) TREATMENTS: A UK PERSPECTIVE

### Abstract

#### Background

As newer therapies for mBC become available, understanding the efficiency of these therapies will be important for HTA recommendations and treatment decisions. EFA may be a useful method of assessing the efficiency of newer interventions. The EFA displays the outlay (cost) and gains with technologies, and displays the next most efficient option going forward. This study was designed to evaluate whether EFA could be useful in identifying the efficiency of mBC therapies adopted by the NHS, and to identify the efficiency frontier for newer technologies.

#### Methods

A literature search was performed to identify mBC treatments that underwent HTA in the UK. Reports were reviewed to identify treatment efficacy and HTA recommendations. Costs were determined for a course of treatment. The incremental costs per patient were plotted on the horizontal axis and incremental median overall survival (ΔOS) of each treatment was plotted on the vertical axis to construct the EFA line. Treatments below this line are considered inefficient options. Treatments above this line have better OS and may redefine the efficiency frontier. Treatments in the upper right quadrant beyond the frontier line are in an area where ceiling price has not been defined. Treatments in the lower right quadrant beyond the frontier line are inefficient due to higher cost for lower OS.

#### Results

Ten reports that evaluated efficacy in terms of median OS were included in the EFA. The therapies are paclitaxel albumin, gemcitabine, trastuzumab, bevacizumab, lapatinib, eribulin and fulvestrant. On the frontier line are paclitaxel albumin (ΔOS of 2.3 months at £2020), gemcitabine (ΔOS of 2.8 months at £6020), and trastuzumab (ΔOS of 4 months at £16939); all received positive recommendations. Lapatinib (ΔOS of 1.9 months at £10180), bevacizumab (ΔOS of 1.7 months at £36560), eribulin (ΔOS of 2.5 months at £4834) and fulvestrant (ΔOS of 2.3 months at £2481) are all below the frontier line and received negative recommendations.

#### Conclusion

EFA may be a useful method for assessing the efficiency of new mBC treatment options for clinical use. Further studies are needed to better understand value in terms of efficiency of treatments in other tumor types and disease areas.

#### Disclosure

All authors have declared no conflicts of interest.

## 338PA CONJOINT ANALYSIS OF WILLINGNESS TO PAY TO AVOID METASTATIC BREAST CANCER SIDE EFFECTS

### Abstract

#### Background

Patients with metastatic breast cancer (MBC) are treated with a variety of regimens with differing side effect profiles. In addition to efficacy, side effect profile can be an important consideration in therapy choice. Conjoint analysis is a research method used to evaluate how trade-offs are made between different attributes. This study assessed the willingness to pay (WTP) to avoid side effects related to MBC treatment using conjoint analysis. The WTP thus informs clinicians of patients' preferences and which side effects they are most desirous of avoiding.

#### Methods

An online, self-administered survey of MBC patients in the US was conducted. The survey was fielded with a sample of adult female patients with a diagnosis of MBC. Key variables (attributes) included in the analysis and with levels described in lay terms were: Alopecia, Diarrhea, Fatigue, Pain, Nausea, Neuropathy, Neutropenia and Out of pocket costs. 15 scenarios (choice-based conjoint questions) were presented where patients selected the most preferred therapy. Each therapy was described with three distinct variables. The choices of variables for each therapy included two side effects and an out of pocket price. The survey also collected information on prior treatment regimens, previous side effect history, and demographics.

#### Results

There were 298 responses. Most respondents were white (84%), married (57%) over 40 years of age (86%), and had private insurance (57%). MBC patients were willing to pay $3,894 to avoid severe diarrhea,$3,479 to avoid being hospitalized due to infection, $3,211 to avoid severe nausea,$2,764 to avoid severe tingling in hands and feet, $2,652 to avoid severe fatigue,$1,853 to avoid obvious hair loss and \$1,458 to avoid severe pain. The most important attributes when selecting a therapy for MBC in terms of average utility were neutropenia, diarrhea and nausea.

#### Conclusions

Patients most highly value the avoidance of diarrhea, neutropenia, and nausea with MBC treatment regimens. These are common side effects seen in many regimens used for treatment of MBC. This information can aid clinical decision making when selecting between MBC treatment options. Treatment regimens providing clinical efficacy while decreasing or eliminating key side effects would be an important choice to consider.

#### Disclosure

D. Lalla: Dr. Lalla is an employee of Genentech, which funded this analysis.

M. Brammer: Dr. Brammer is an employee of Genentech, which funded this analysis.

R. Carlton: Dr. Carlton is an employee of Xcenda, which received funding from Genentech to conduct this analysis.

T. Bramley: Dr. Bramley is an employee of Xcenda, which received funding from Genentech to conduct this analysis.

A. D'Souza: Dr. D'Souza is an employee of Xcenda, which received funding from Genentech to conduct this anaysis.

## 339PFULVESTRANT 500 MG AS FIRST-LINE TREATMENT IN HORMONE-POSITIVE (HR+) METASTATIC BREAST CANCER (MBC) PATIENTS (PTS): PROSPECTIVE EVALUATION OF ACTIVITY, SAFETY, QUALITY OF LIFE (QOL) AND TUMOUR MARKERS CHANGES

### Abstract

#### Background

Fulvestrant 500 mg has been shown to have a biologically greater effect and to produce a clinical meaningful benefit over Fulvestrant 250 mg. We carried out a prospective phase II trial to evaluate the activity and safety of such an option as 1st line treatment of HR+ MBC; an analysis of QoL, patient compliance, and prognostic value of tumour markers monitoring was also performed.

#### Patients and methods

Pts with HR+ MBC relapsing after Tamoxifen and/or Aromatase Inhibitors adjuvant therapy received Fulvestrant 500 mg i.m. on day 0, then 500 mg on days 14-28 and every 28 days thereafter, until disease progression or unacceptable toxicity or patient refusal. Changes in cancer antigen 15.3 (CA 15.3) and carcinoembryonic antigen (CEA) were monitored monthly over treatment; QoL and treatment tolerability were assessed every 2 cycles.

#### Results

Forty-eight consecutive pts were enrolled and treated. The overall clinical benefit rate was 68%, with 3 (6%) complete (CR) and 12 (25%) partial responses (PR), and 18 (48%) stable diseases (SD ) lasting >24 weeks. The median PFS was 11 months, median response duration was 9 months. Toxicity was manageable, also in pts given long-duration therapy (>18 months). Analysis of QoL by QLQ-BR23 showed no deterioration of the evaluated items over treatment, while the compliance assessment documented an improvement of tolerability of treatment compared with their previous adjuvant hormonotherapy. In pts achieving a CR or PR both CA 15.3 and CEA serial levels decreased significantly over the first 4 months of treatment, while in long-lasting SD a statistically significant difference from baseline was reached within a median of 8 months.

#### Conclusions

Our results confirm the known activity and tolerability of Fulvestrant 500 mg as 1st line therapy for HR+ MBC, with good patient compliance over long-term treatment and preservation of QoL. Of interest, changes in tumour markers resulted significantly predictive of treatment activity in responding pts, in contrast with previously reported data with Fulvestrant 250 mg, reflecting and additional difference between the two drug dosages.

#### Disclosure

All authors have declared no conflicts of interest.

## 340PFULVESTRANT AFTER FAILUR OF ADJUVANT HORMONAL THERAPY: MULTICENTER RETROSPECTIVE OBSERVATIONAL STUDY

### Abstract

#### Background

Aromatase inhibitors (AI) are superior to TAM in the treatment of hormonal receptor positive metastatic breast cancer (MBC). Fulvestrant (F), an estrogen receptor antagonist demonstrated activity when used after failure of previous TAM or AI. Objective: To evaluate the position of F in the hormonal strategy sequence in hormonal responsive MBC related to the evolution of adjuvant hormonal therapy (HT).

#### Material and methods

MBC patients with disease progression after adjuvant HT or after treatment (CT or HT) for metastatic disease, candidate to receive F at the dose of 250 mg/ month (registered in AIFA database) from May 1st 2006 to July 31st 2008. The study registered the HT sequence in clinical practice. Overall 201 patients were collected from 13 Italian Centers. Main patients' characteristics are summarized in the table.

Table: 340P

Characteristics N 201 (%) Characteristics N 201 (%)
T size Tis-T1-T2 T3-T4 T Unk 138(68,6) 40 (20) 23 (11,4) PgR Pos Neg Unk 142(70,6) 33 (16,4) 26 (13)
Nodal Status N0 N+ Nx 54 (26,9) 120 (59,7) 27 (13,4) Adj CT Yes No Unk 105 (51,7) 94 (46,8) 3 (1,5)
ER Pos Neg Unk 169 (84) 12 (6) 20 (10) Adj HT Yes No Unk 141 (70,1) 46 (22,9) 14 (7)
Characteristics N 201 (%) Characteristics N 201 (%)
T size Tis-T1-T2 T3-T4 T Unk 138(68,6) 40 (20) 23 (11,4) PgR Pos Neg Unk 142(70,6) 33 (16,4) 26 (13)
Nodal Status N0 N+ Nx 54 (26,9) 120 (59,7) 27 (13,4) Adj CT Yes No Unk 105 (51,7) 94 (46,8) 3 (1,5)
ER Pos Neg Unk 169 (84) 12 (6) 20 (10) Adj HT Yes No Unk 141 (70,1) 46 (22,9) 14 (7)

#### Results

ORR were: CR 2%; PR 7%; SD 43%; PD 41%; NE 7%. Eighty-six patients (43%) had visceral metastases. Median duration of F was 7 months (1-42); 8 patients are ongoing and for 1 patient the duration was unknown. 16, 39 and 56 patients received F as 1st, 2nd and 3rd line therapy respectively. For the remaining patients F was administered as more than the 4th line therapy. The possibility to obtain a benefit from F did not seem to correlate with the line of treatment. One of the 2 CR occurred in a patient with visceral disease (lung). Among the patients who progressed under treatment 42% had visceral disease, that was almost the same rate of the whole population. This data suggests that there is lack of correlation between the sites of disease and the probability of benefit. Toxicity was negligible.

#### Conclusions

More than 50% of the patients benefited from F notwithstanding 73% of them received the treatment as third or more line of HT. This study does not show any correlation between the sites of the metastases and the treatment results.

#### Disclosure

All authors have declared no conflicts of interest.

## 341PADJUSTED INDIRECT COMPARISON ANALYSIS DEMONSTRATES SIGNIFICANT BENEFIT IN PROGRESSION-FREE SURVIVAL FOR FULVESTRANT 500MG COMPARED TO ANASTROZOLE IN ADVANCED BREAST CANCER

### Abstract

#### Objectives

Randomised controlled trials (RCTs) provide the highest level of evidence for comparing two treatments. However, with the increasing number of cancer drugs, direct comparison through RCTs is not feasible for all treatment indications and combinations. Bucher et al (1997) have developed a method for an adjusted indirect comparison analysis between RCTs. This method was used to compare the efficacy of fulvestrant 500mg and anastrozole 1mg in advanced breast cancer (ABC).

#### Methods

A systematic literature search on RCTs of fulvestrant 500mg or anastrozole 1mg in ABC was performed in June 2011, using CENTRAL, EMBASE and MEDLINE databases. Published data were used to perform a meta-analysis and an adjusted indirect comparison analysis (Bucher method). The primary endpoint was progression free survival (PFS).

#### Results

Three RCTs with 1023 patients were identified comparing fulvestrant 500mg (F500) with fulvestrant 250mg (F250), and two RCTs with a total of 851 patients were identified comparing anastrozole 1mg (Ana) with fulvestrant 250mg. Meta-analysis demonstrated a significant benefit in PFS for F500 compared to F250 (Hazard ratio (HR) 0.80 95% Confidence Interval (CI):0.69-0.93). There was no significant difference in PFS between Ana and F250 (HR 0.95, 95%CI 0.82-1.1). Using F250 as common comparator, the adjusted indirect comparison analysis demonstrated a significant benefit in PFS for F500 compared to Ana (HR 0.76, 95%CI 0.62-0.94).

#### Conclusions

In the absence of a direct RCT comparison, this adjusted indirect comparison shows that fulvestrant 500mg significantly improves PFS in ABC compared to anastrozole 1mg. Bucher HC, Guyatt GH, Griffith LE, Walter SD (1997). The results of direct and indirect treatment comparisons in meta-analysis of randomized controlled trials. J Clin Epidemiol. 1997 Jun;50(6):683-91.

#### Disclosure

P. Turner: Pauline Turner is a full-time employee of AstraZeneca and is a stockholder in Astra Zeneca.

M. Howlett: Matthew Howlett is a full time employee of Astrazeneca and is a stockholder in Astrazeneca.

All other authors have declared no conflicts of interest.

## 342PFULVESTRANT (FUL) PLUS ENZASTAURIN (ENZA) VS FUL PLUS PLACEBO (PBO) IN AROMATASE INHIBITOR (AI)-RESISTANT METASTATIC BREAST CANCER (MBC): A RANDOMIZED, DOUBLE-BLIND, PHASE 2 TRIAL

### Abstract

AIs are used as first-line treatment for postmenopausal women with hormone-receptor-positive MBC. Overexpression of PKC α has been linked to AI resistance in several studies. We examined whether Enza, a serine/threonine kinase inhibitor that targets PKC, could improve the effect of Ful in pts who progressed following first-line AI treatment for MBC.

Postmenopausal pts with hormone-receptor-positive, HER2-negative, locally advanced or MBC who progressed on prior AI received a loading dose of Ful 500 mg IM on Day (D) 1 and 250 mg on D 15 of Cycle (C) 1 and D 1 of each cycle thereafter. Enza 500 mg or PBO was administered orally once daily (QD) or 250 mg twice daily (BID). Primary endpoint was the clinical benefit rate (CBR). Secondary endpoints were response rate (RR), duration of CB, progression-free survival (PFS) and safety. A total of 156 pts was randomly assigned to therapy; 152 received at least 1 dose of study drug (39 BID; 55 QD; 58 PBO). Baseline disease characteristics were balanced across arms. There was no statistically significant difference in CBR between pts in Ful + Enza vs Ful + PBO. There was no statistically significant difference in CBRs, RRs and PFS between pts on QD and BID dosing schedules. Pts on BID dosing had numerically more TEAEs vs those on QD and PBO (61.5%, 43.6%, and 46.6%, respectively) and numerically more Grade 3/4 TEAEs vs QD and PBO (17.9%, 9.1%, and 5.2%, respectively). Most frequent Grade 3/4 TEAEs in the BID arm were fatigue (n [%]) (4 [10.3%]), dyspnea (2 [5.1%]) and nausea (2 [5.1%]).

8 pts died; 5 due to disease, 3 due to AEs: 1 drug-related hepatic dysfunction (Enza QD arm), 1 non-drug-related myocardial infarction and 1 non-drug-related respiratory failure (both in the Enza BID arm).

Ful + Enza [QD + BID] N = 94 Ful + PBO N = 58 P-value vs PBO
CBR
Number of responders,
n (%), (95% CI)

41 (43.6)
(33.4, 54.2)

24 (44.8) (31.7, 58.5)
0.62
RR, % (95% CI) 5.3 (1.7, 12.0) 5.2 (1.1, 14.4) 0.64
PFS, months (95% CI) 5.2 (3.5, 7.4) 5.5 (3.8, 7.4) 0.59
Median duration of CB, months 9.6 9.7 0.86
Ful + Enza [QD + BID] N = 94 Ful + PBO N = 58 P-value vs PBO
CBR
Number of responders,
n (%), (95% CI)

41 (43.6)
(33.4, 54.2)

24 (44.8) (31.7, 58.5)
0.62
RR, % (95% CI) 5.3 (1.7, 12.0) 5.2 (1.1, 14.4) 0.64
PFS, months (95% CI) 5.2 (3.5, 7.4) 5.5 (3.8, 7.4) 0.59
Median duration of CB, months 9.6 9.7 0.86

CBR = complete response + partial response + stable disease Addition of Enza to Ful does not improve disease outcome in pts with locally advanced or MBC after progression on AI.

#### Disclosure

K. Mansouri: K. Mansouri is an employees of Eli Lilly and Co.

L. Cirri: L. Cirri is an employees of Eli Lilly and Co.

P. Shi: P. Shi is an employees of Eli Lilly and Co.

O. Hamid: O. Hamid is an mployees of Eli Lilly and Co.

All other authors have declared no conflicts of interest.

## 343PNEW IMMUNOHISTOCHEMICAL MARKERS PREDICTIVE OF RESPONSE TO NEOADJUVANT CHEMOTHERAPY PLUS TRASTUZUMAB IN HER2-POSITIVE LOCALLY ADVANCED BREAST CANCER: A SINGLE CENTER EXPERIENCE

### Abstract

We present the results of a prospective pilot study aimed to investigate the value of new immunohistochemichal predictive markers of response to chemotherapy in locally advanced HER-2 breast cancer. pTEN loss, pAKT and HER-3 overexpression cause PI3K activation and induce resistance to trastuzumab in vitro and in vivo, with poorer clinical responses in patients with advanced disease. We studied the espression of EGFR, HER-3, pTEN and pAKT in 31 patients with locally advanced HER-2 positive breast cancer who received neoadjuvant chemotherapy plus trastuzumab. Mean age of patients was 55.7 years (median 55, 95% CI 44,2-63,5 years). Results of immunohistochemistral staining are reassumed in the table.

Mean Median 25% C.I. 75% C.I.
ER% 40,6 40,0 0,0 77,5
PgR% 23,1 10,0 0,0 40,0
Ki-67% 38,2 40,0 25,0 50,0
EGFR Hscore 6,1 0,0 0,0 1,0
HER3 (% positive cells) 50,0 50,0 32,5 70,0
HER3 intensity 1,5 1,0 1,0 2,0
HER3 H score 82,6 70,0 32,5 115,0
PTEN (% positive cells) 48,9 70,0 2,5 70,0
PTEN intensity 1,5 1,0 1,0 2,0
AKT Hscore 56,7 20,0 0,0 70,0
AKT IRS 2,5 2,0 0,0 3,0
Mean Median 25% C.I. 75% C.I.
ER% 40,6 40,0 0,0 77,5
PgR% 23,1 10,0 0,0 40,0
Ki-67% 38,2 40,0 25,0 50,0
EGFR Hscore 6,1 0,0 0,0 1,0
HER3 (% positive cells) 50,0 50,0 32,5 70,0
HER3 intensity 1,5 1,0 1,0 2,0
HER3 H score 82,6 70,0 32,5 115,0
PTEN (% positive cells) 48,9 70,0 2,5 70,0
PTEN intensity 1,5 1,0 1,0 2,0
AKT Hscore 56,7 20,0 0,0 70,0
AKT IRS 2,5 2,0 0,0 3,0

Patients were homogeneous for tumor burden at diagnosis and received anthraciclin-taxane and trastuzumab-based neoadjuvant treatment. 14 out of 31patients achieved pathological complete remission. Ki-67 and HER3 H score were significatively higher in patients who achieved complete remission (medians were 45.5% versus 25%, p = 0.022; 100% versus 50%, p = 0.045 respectively. We found a correlation between EGFR H score and HER3 Hscore (p = 0.02), and an inverse correlation between age and EGFR H score (p = 0.05) and between PgR and AKT intensity (p = 0.03). Ki-67 and HER3 H score maintained significatively different medians in the group of patients who experienced pathological complete response versus the group on incomplete responders with multivariate analysis (p < 0.01).

#### Disclosure

All authors have declared no conflicts of interest.

## 344PPHARMACOKINETICS (PK) OF PERTUZUMAB (P) WITH TRASTUZUMAB (T) AND DOCETAXEL (D) IN HER2-POSITIVE FIRST-LINE METASTATIC BREAST CANCER (MBC): RESULTS FROM THE PHASE III TRIAL CLEOPATRA

### Abstract

#### Introduction

P is a humanized mAb that inhibits heterodimerization of HER2. P and T bind distinct HER2 epitopes, and due to their complementary mechanisms of action they provide a more comprehensive blockade of HER2 signaling. Based on preclinical efficacy models, a steady-state trough P concentration (Ctrough) of 20 ug/ml was selected as target in pts. CLEOPATRA is a Phase III study comparing P + T + D vs placebo (Pla) + T + D in HER2-positive 1L MBC (Baselga NEJM 2012). The objectives of the substudy reported here are to characterize the P PK in the presence of T and D, and to explore potential drug − drug interactions.

#### Methods

P/Pla (840 mg loading, 420 mg maintenance) was administered on Day 1 of each cycle; T (8 mg/kg loading, 6 mg/kg maintenance) was administered on Day 2 of Cycle 1 and on Day 1 of each cycle onward following P; D (75 mg/m2, escalation to 100 mg/m2 if tolerated) was administered on Day 2 of Cycle 1 following T and on Day 1 of each cycle onward following T. All drugs were given q3w iv. Blood samples for P were collected before and after infusion at Cycles 1, 3, 6, 9, 12, 15, 18, and at treatment discontinuation. Samples for T were collected before and after infusion at Cycles 1 and 3. Samples for D were collected at Cycle 1 at 8 serial time points during and following the infusion over a 24 h period to allow calculation of Cmax, CL, Vss, t1/2, AUC0–t, and AUC0-inf.

#### Results

37 pts (17 Pla arm, 20 P arm) were available for PK evaluation. Serum P Ctrough exceeded the target of 20 ug/ml in >90% of pts and there was no impact of T and D on P PK, compared with historical data. Mean serum T Cmax and Cmin at Cycles 1 and 3 were similar in both arms. Ratios of geometric LS means of P + T + D/Pla + T + D x100 for serum T were: Cycle 1 Cmax 90.3; Cycle 3 Cmin 95.9; Cycle 3 Cmax 81.0. D PK parameters were similar in both arms. Ratios of geometric LS means of P + T + D/Pla + T + D x100 for plasma D were: AUC0–t 104.9, AUC0-inf 101.4, Cmax 92.5.

#### Conclusion

P PK parameters were consistent with previous studies, and co-administration of T and D appears not to influence P PK in HER2-positive MBC. There was no evidence of drug − drug interactions between P and T, or between P and D, which have different clearance pathways.

#### Disclosure

J. Cortes: I am an advisory board member for Roche, Celgene and Novartis. I have received research funding from Roche, Celgene, Cephalon and Ferrer.

S. Swain: Advisory Board for Genentech/Roche for EMILIA study and Avastin - uncompensated. Research funding: Genentech/Roche.

T. Patel: I have an advisory board membership for Genentech/Roche to disclose and are a speaker for Genentech/Roche. I have received research funding from Genentech/Roche.

N. Masuda: I have honoraria to disclose received from Chugai Pharmaceutical Co., Ltd.

V. McNally: I am a Roche employee and hold Roche shares.

J. Visich: I am an employee of Genentech.

J. Baselga: Dr Baselga reports the following relationships with relation to the topic of this abstract: Roche, Sanofi Aventis, Consulting/Scientific Advisory Board. Roche, Sanofi Aventis, Honoraria for speaking engagements.

All other authors have declared no conflicts of interest.

## 345PPROGNOSTIC FACTORS INFLUENCING THE SELECTION OF BEVACIZUMAB COMBINED WITH CHEMOTHERAPY IN PATIENTS WITH HER2-NEGATIVE METASTATIC BREAST CANCER IN ROUTINE CLINICAL PRACTICE. ONCOSUR-AVALOX: OBSERVATIONAL CROSS-SECTIONAL STUDY

### Abstract

#### Background

Combining bevacizumab (BEV) with chemotherapy (CT) improves survival in HER2-negative metastatic breast cancer (MBC). We investigated the influence of age, ECOG, hormonal status, number of sites and location of metastases and patient decision on the selection of BEV combined with CT in MBC.

#### Methods

Observational cross-sectional multicenter study in pts with HER2- negative MBC who have received first-line CT with BEV.

#### Results

From November 2010 to November 2011, 124 pts were included: median age 51 (45-64) yr; ECOG: 0 = 50%; 60% pre-menopausic; 23% triple-negative (TN); 77% hormone receptorpositive (HR+). Metastatic disease: ≥3 sites = 42% (TN: 32%; HR + : 45%); location: 44% bone, 35% lung, 30% liver. Most frequent BEV-based combinations were paclitaxel/BEV (53%) and docetaxel/BEV (14.5%); median no. of CT cycles: 6 (5-8). A disease-free survival (DFS) ≥12 months was achieved by 73%; TN: 68%; HR + : 76%. Overall response rate (ORR) was 58%: 51% partial response (PR), 7% complete response (CR); 28% stable disease (SD) and 10% disease progression. TN: ORR 44% (40% PR), clinical benefit 80% (36% SD); HR + : ORR 62% (54% PR), clinical benefit 87% (25% SD). 58% presented at least one toxicity, mainly grade 1-2; 26% BEV-related: only 3 (2.4%) grade 3 toxicities; no grade 4. Receiving adjuvant hormonal therapy was associated to DFS ≥12 months (p < 0.05). ER+ tumors (OR: 0.215; 95% CI: 0.08-0.56; p = 0.002) and one metastatic site, vs. ≥3 sites (OR: 0.309; 95% CI: 0.12-0.83; p = 0.020) were independent factors associated with the selection of paclitaxel-BEV therapy in the overall population (TN or HR+). Metastases in the liver were significantly related to paclitaxel-BEV administration (p < 0.01).

#### Conclusions

Our findings suggest that first-line CT with BEV is an active and tolerable treatment option for pts with TN and HR+ MBC. ER+ tumors and a single metastatic site were identified as independent factors for the selection of a paclitaxel-BEV therapy. The presence of metastases in the liver was significantly associated to the administration of a paclitaxel-BEV regimen.

#### Disclosure

All authors have declared no conflicts of interest.

## 346PCHARACTERISTICS OF PATIENTS WITH HER2-POSITIVE METASTATIC (MBC) OR LOCALLY ADVANCED BREAST CANCER (ABC), TREATED WITH TRASTUZUMAB (T) AS 1ST LINE-THERAPY AND PROGRESSION-FREE FOR AT LEAST 3 YEARS: INTERIM ANALYSIS OF THE LORHA STUDY

### Abstract

#### Background

For more than ten years, treatment of HER2-positive mBC patients (Pts) is based on T plus taxane in 1st line therapy. So far, in numerous studies, we have observed few subsets of Pts (long-term responders) who have not experienced disease progression for several years after T-based regimen in 1st line treatment. This study aims to characterise these Pts in the daily practice from a clinical and biological perspective.

#### Material and methods

This is an ambispective French multicentre non-interventional study. Eligible Pts were women aged ≥18 years with HER2-positive mBC or aBC treated with T as 1st line and who were progression-free for at least 3 years after starting T. The primary objective was to describe the clinical and tumor characteristics of these Pts. Progression Free Survival (PFS), Overall Survival (OS), data on treatment administration and safety were also collected. An exploratory biomarkers analysis is planned on tumor tissue samples. Here we present some preliminary results based on the interim analysis performed at the end of inclusions.

#### Results

159 Pts were enrolled in 2011 and 110 Pts were eligible for data analysis. Median age was 59 years [34-95]. Tumor characteristics were: invasive ductal carcinoma for 96 Pts (88%), positive hormonal receptors in 63 Pts (58%). At initial diagnosis, presenting stages were I-II for 52 Pts (50%). 36 Pts (33%) had a mBC de novo or an aBC. The main metastatic localisations were bone, liver and lung in 51 (47%), 35 (32%) and 22 (20%) Pts respectively. Median T treatment duration was 4.1 years [0.8 - 11.0] in 1st line. T was associated with a taxane-based chemotherapy in 86 Pts (78%). Median PFS was 6.4 years [4.9; - ]. Median OS was not reached. 13 retrospective adverse events related to T (cardiac or leading to discontinuation) were reported. None of these was serious.

#### Conclusions

In this long PFS population treated with a T-based treatment in first line for aBC or mBC Pts, no specific profile in terms of clinical or histological characteristics have been observed. Thus, the exploratory biomarkers analysis will be useful to identify such a profile.

#### Disclosure

O. Tredan: Roche consultant.

P. Beuzeboc: Roche consultant.

D. Coeffic: Roche consultant.

M. Fellous: Roche employee.

L. Arnould: Roche Consultant.

All other authors have declared no conflicts of interest.

## 347PTHE ENCHANTTM TRIAL: AN OPEN LABEL MULTICENTER PHASE 2 WINDOW OF OPPORTUNITY STUDY EVALUATING GANETESPIB (STA-9090) MONOTHERAPY IN WOMEN WITH PREVIOUSLY UNTREATED METASTATIC HER2 POSITIVE OR TRIPLE NEGATIVE BREAST CANCER (TNBC)

### Abstract

#### Background

Hsp90 is a molecular chaperone required for proper folding and activation of many cancer-promoting proteins. Several Hsp90 clients are oncoproteins known to play a key role in the pathobiology of breast cancer, including HER2, p95-HER2, EGFR, ER, PI3K, AKT, and VEGFR. The inactivation of these oncoproteins by Hsp90 inhibition is a promising approach for breast cancer therapy. Ganetespib is an Hsp90 inhibitor which has shown anti-tumor activity in heavily pretreated patients with lung, breast, and other cancers. Ganetespib is well tolerated without severe liver or common ocular toxicities. In a phase 2 trial, 22 breast cancer patients who had received up to 3 prior lines of chemotherapy including trastuzumab were treated with ganetespib monotherapy. In patients with HER2+ disease, the objective response rate (ORR) was 15% (2/13) and the SD rate was 46% (6/13). Only 3 patients presented with TNBC; one of those patients achieved SD with substantial tumor shrinkage on treatment.

#### Methods

This is a single arm international open-label Phase 2 study in patients with HER2 amplified, or triple negative breast cancer. Patients must not have received any prior therapy in the metastatic setting. Prior adjuvant therapy is allowed. Primary endpoint: ORR. Main secondary endpoints include disease control rate, and progression free survival. Additionally, fresh biopsies and serum samples are collected from all patients for determination of predictors of response and mechanisms of resistance to treatment. Patients are treated with ganetespib 150 mg/m2 is given twice weekly of a 4-week cycle for up to 12 weeks. A total of 70 patients are planned for accrual. At the time of submission, the study is receiving IRB approvals in several centers.

#### Disclosure

All authors have declared no conflicts of interest.

## 348PHGFK1 INHIBITS BONE METASTASIS IN BREAST CANCER THROUGH THE TAK1/P38 MAPK SIGNALING PATHWAY

### Abstract

#### Background

Breast cancer metastasis to bone represents a devastating complication of advanced breast cancer, frequently resulting in significant increases in morbidity and mortality. An understanding of the mechanisms that govern breast cancer metastasis at the molecular level should lead to more effective therapies. Recently, the kringle 1 domain of human hepatocyte growth factor (HGFK1) was identified as a candidate metastasis suppressor gene.

#### Methods

Here, we investigated whether HGFK1 is a key regulator of breast cancer bone metastasis.

#### Results

Of the 193 human breast carcinoma tissue samples examined, HGFK1 expression was positive in 82 (42.4%). The positive expression of HGFK1 was significantly associated with a better prognostic value (P < 0.001) and inversely correlated with bone metastasis (P = 0.003). The efficacy of adeno-associated virus carrying HGFK1 (AAV-HGFK1) in osteolytic bone metastasis was then evaluated using an in vivo bone metastasis model. AAV-HGFK1 significantly inhibited osteolytic bone metastasis and prolonged the survival of mice in this model (P < 0.01). In vitro, HGFK1 expression resulted in significant anti-invasion effects, enhanced the phosphorylation of TAK1, p38 MAPK and MAPKAPK2, and decreased the expression of receptor activator of NF-κB (RANK), which was abrogated by the p38 MAPK inhibitor SB203580.

#### Conclusions

This study shows for the first time that HGFK1 significantly inhibits the metastasis of breast cancer to bone by activating the TAK1/p38 MAPK signaling pathway and inhibiting RANK expression. Thus, AAV-HGFK1 treatment represents a potential therapy for bone metastasis in breast cancer.

#### Disclosure

All authors have declared no conflicts of interest.

## 349PIMPACT OF DIABETES AND HYPERGLYCEMIA ON SURVIVAL IN ADVANCED BREAST CANCER PATIENTS

### Abstract

#### Background

Clinical experience and previous studies suggest that women with diabetes and breast cancer have worse outcomes than their non-diabetic counterparts. The purpose of this study was to examine the impact of diabetes and hyperglycemia on cancer-specific survival of patients with metastatic or recurrent breast cancer.

#### Methods

We performed a retrospective analysis of patients with advanced breast cancer receiving palliative chemotherapy from 2006 to 2011 at the National Cancer Institute in Mexico, and compared breast cancer-specific mortality in diabetic and non-diabetic patients, as well as in patients that presented hyperglycemia during palliative treatment.

#### Results

A total of 265 patients receiving palliative therapy were eligible for inclusion. Previous diagnosis or detection of diabetes at recurrence was recorded in 40 patients (15%). No difference was observed between diabetic and non-diabetic patients in terms of OS. A statistically significant difference in OS was observed between patients without diabetes and diabetic patients who had hyperglycemia (p = 0.003). OS in diabetic patients with proper metabolic control was shown to be superior compared to diabetics with hyperglycemia (p = 0.01) Hyperglycemia was identified in 14% of non-diabetics at some point while receiving palliative treatment. For patients that experienced hyperglycemia during treatment or who had a mean glucose level above 130, either in the diabetic or non-diabetic subgroups, a worse outcome was noted compared to normoglycemic patients, with a HR of 1.5 (p = 0.029) and HR of 2.04 (p = 0.006) for death, respectively.

#### Conclusions

Elevated glucose levels confer a poor outcome in diabetic and non-diabetic patients in contrast with patients with normoglycemic levels, conferring an elevated risk of death. According to these results, clinicians must monitor glucose levels during treatment for advanced breast cancer disease, and should take action in order to maintain normal glucose levels.

#### Disclosure

All authors have declared no conflicts of interest.

## 350PANTITUMOR ACTIVITY OF DUAL PI3K AND ER BLOCKADE IN ER POSITIVE BREAST CANCER MODELS

### Abstract

#### Background

Activation of the phosphoinositide 3-kinase (PI3K) pathway, either by receptor tyrosine kinase overexpression or PI3K/Akt/mTOR axis dysregulation, has been implicated in endocrine therapy resistance, prompting combination of PI3K inhibitors and antiestrogen therapy in the clinical setting such as in the recent BOLERO-2 study. We hypothesize that dissecting the molecular crosstalk between the PI3K and estrogen receptor (ER) pathways will help define the subset of patients most responsive to combined PI3K/antiestrogen therapy.

#### Methods

ER+ cell lines MCF7, MCF7-long term estrogen deprived (LTED), MCF7-fulvestrant resistant clones Fslx64 and Fslx70, T47D, ZR75-1, CAMA1, MDA361, KPL-1, BT474, EFM19, HCC1428, UACC812, were treated with the ER degrader, fulvestrant, and the p110α-specific PI3K inhibitor BYL719 in vitro. Cell viability was measured by CellTiter-Glo and Crystal Violet. MCF7 and ER+ patient-derived xenografts were treated with fulvestrant, BYL719 or the combination in vivo. Protein expression was measured by Western blot and immunohistochemistry.

#### Results

Fulvestrant or BYL719 treatment resulted in variable inhibition of cell viability in all ER+ cell lines. Combination treatment was significantly superior to monotherapy in MCF7, MCF7-LTED, MCF7-Fslx64 and MCF7-Fslx70. While MCF7 clones Fslx64 and Fslx70 were resistant to >1 µM of fulvestrant, they were exquisitely sensitive to BYL719. Moreover, PI3K inhibition led to ER upregulation in ER+ cell lines, including those most sensitive to combination treatment. Total ER levels also increased in MCF7 xenografts treated with therapeutic doses of BYL719.

#### Conclusions

Combined treatment with BYL719 and fulvestrant in vitro was superior to single-agent treatment in 4 of 13 ER+ cell lines. Importantly, dual PI3K/ER blockade was effective in cells resistant to ER deprivation and/or degradation. Induced ER levels following PI3K suppression may represent a feedback mechanism by which ER+ cells escape PI3K inhibition. We are currently studying whether this phenomenon predicts sensitivity to dual PI3K/ER blockade in ER+ breast cancer models.

#### Disclosure

J. Baselga: J. Baselga is a consultant/advisory board member for Aragon, AstraZeneca, Sanofi, Bayer, Onyx, Chugai, Constellation, Exelixis, Intellikine, Merck, Novartis, and Roche Genentech.

All other authors have declared no conflicts of interest.

## 351PSAFETY OF EVEROLIMUS FOR WOMEN OVER 65 YEARS OF AGE WITH ADVANCED BREAST CANCER: 18-MO FOLLOW-UP OF BOLERO-2

### Abstract

#### Background

Hormone-receptor–positive (HR+) breast cancer (BC) refractory/resistant to nonsteroidal aromatase inhibitor (NSAI) may be treated with the steroidal AI exemestane (EXE), although there is no approved treatment standard. The BOLERO-2 trial showed that adding everolimus (EVE) to EXE significantly improved clinical benefit beyond that of EXE alone (Hortobagyi et al, SABCS 2011, Abstract S3-7). As many women with advanced BC are elderly (≥ 65 years), the efficacy and tolerability of EVE + EXE in this population are of interest.

#### Methods

BOLERO-2 is a phase 3, randomized trial comparing EVE (10 mg once daily) vs placebo (PBO), both plus EXE (25 mg once daily) in postmenopausal women with advanced HR+ BC progressing or recurring after NSAIs.

#### Results

Baseline disease and prior treatment characteristics were balanced between study arms (N = 724). At 18 months' median follow-up, adding EVE to EXE significantly improved progression-free survival in patients < 65 (n = 449; 8.3 vs 2.9 mo; HR = 0.38; 95% CI = 0.30, 0.47) and ≥ 65 years (n = 275; 6.8 vs 4.0 mo; HR = 0.59; 95% CI = 0.43, 0.80). Overall incidence of adverse events (AEs) was marginally higher in patients ≥ 65 years (n = 272, safety population) compared with those < 65 years. Grade 3/4 AEs occurred in 50% of patients ≥ 65 years compared with 42% of patients < 65 years. Incidences of grade 3/4 stomatitis, rash, pneumonitis, and hyperglycemia in EVE-treated patients ≥ 65 years and those < 65 years were similar. Additional analysis using an age cutoff of 70 years also showed no meaningful differences in the efficacy/safety profile of EVE. Grade 3/4 AEs in patients ≥ 65 years reported among patients receiving EVE (n = 192) but not in those receiving PBO included anemia (9%), hyperglycemia (7%), stomatitis (9%), dyspnea (8%), pneumonitis (5%), neutropenia (3%), and hypertension (2%). These AEs were also reported at similar frequency in EVE-treated patients < 65 years.

#### Conclusions

Adding EVE to EXE was effective and well tolerated overall and, in elderly patients with advanced BC, grade 3/4 AEs were uncommon and manageable. Overall, AEs were consistent with the known safety profile of EVE.

#### Disclosure

M. Gnant: Research support from GSK, Sanofi-Aventis, Novartis, and Roche, consultant to Merrion and Novartis, and received honoraria & travel support from Amgen, Pfizer, Novartis, GSK, Bayer, Sandoz, AstraZeneca, and GenomicHealth.

S. Noguchi: S. Noguchi received grant support from AstraZeneca, BMS, Chugai, GSK, Novartis, Pfizer, Sanofi-Aventis, and Takeda, and honoraria (speaking, advisory boards, etc.) from AstraZeneca, Chugai, GSK, Novartis, Pfizer, Sanofi-Aventis, and Takeda.

M. Piccart: Board for PharmaMar, consultant Sanofi-Aventis, Amgen, BMS, GSK, Boehringer, Roche, & Bayer, grant support Pfizer, Amgen, Bayer, Boehringer, BMS, GSK, Roche, & Sanofi-Aventis, honoraria Bayer, BMS, GSK, Boehringer, Roche, Amgen, & AstraZeneca.

J. Baselga: J. Baselga is a consultant to Novartis, Roche, Merck, Sanofi-Aventis, Verastem, Bayer, Chugai, Exelixis, Onyx, and Constellation.

A. Panneerselvam: Employee of Novartis with stock/stock options.

T. Taran: Employee of Novartis with stock/stock options.

T. Sahmoud: Employee of Novartis with stock/stock options.

G.N. Hortobagyi: Member of the Scientific Advisory Board of Allergan, is a consultant to Allergan, Novartis, Genentech, and Sanofi-Aventis, has received grant support from Novartis, and has received travel expense reimbursement from Novartis, Genentech, and Sanofi-Aventis.

K. Pritchard: Consult sanaven AZE Roche PFE NVR ABR AMG GSK res funding NCICCT Grp contracted AZE BMS SanAven AMG PFE NVR GSK & OrthoBio honoraria & Spkr B SanAven AZE PFE Roche NVR & AMG paid expert test SanAven AZE & GSK AdCom SanAven AZE Roche PFE NVR GSK & AMG.

All other authors have declared no conflicts of interest.

## 352PMTOR EXPRESSION IN BASAL-LIKE BREAST CANCER AND THE ABILITY OF EVEROLIMUS TO INHIBIT THE INVASION CANCER CELL CAPACITY

### Abstract

The introduction of high-throughput technologies in breast cancer enabled the recognition of groups with prognostic value, in which target-therapies can be applied. However, a relevant percentage of patients show no clinical benefit or incur in the development of acquired resistance. A possible solution could be the inhibition of pathways that are common in all tumor subtypes that have a proven role in carcinogenesis. Alterations of the serine-threonine kinase mammalian target of rapamycin (mTOR) signaling pathway are common in cancer and thus mTOR is being pursued as a therapeutic agent. Everolimus, a rapamycin analog, has already an established activity in the treatment of renal cell carcinoma. In this study, we proposed to evaluate the expression of activated mTOR in a large series of invasive carcinoma samples and cell lines, and its association with the four main molecular subtypes (Luminal A, Luminal B, HER2-overexpressing and Basal-like). We also aimed to evaluate the ability of Everolimus to inhibit mTOR expression and function in breast cancer cells. p-mTOR expression was found in 66.7% (231/348) of the invasive breast carcinoma cases analysed. Considering the molecular subtypes of breast carcinomas, p-mTOR was more frequently observed in basal-like breast carcinomas (80.6%). All breast cancer cell lines, representative of distinct molecular subtypes, showed expression of total and activated mTOR. These cells have been treated with RAD001, in order to assess their sensitivity to this drug, and all cell lines showed a decrease of p-mTOR expression after everolimus treatment. Due to the higher prevalence of p-mTOR in basal-like tumors, we treated three basal-like cell lines with Everolimus to assess the effects on cell invasion and aggregation. Cell invasion was significantly inhibited in response to Everolimus. The results revealed that there is a significant higher frequency of p-mTOR in basal-like tumors, compared with the other subtypes. In addition, Everolimus is able to significantly decrease mTOR expression and activity, inhibiting invasion capacity of basal-like breast cancer cells emphasizing the antitumour activity of mTOR inhibitors in breast cancer models.

#### Disclosure

All authors have declared no conflicts of interest.

## 353PTHE PROGNOSTIC MEANING OF PROTEIN TYROSINE PHOSPHATASE NON-RECEPTOR TYPE 12 (PTPN 12) EXPRESSION LOSS IN BREAST CANCER PATIENTS

### Abstract

#### Background

Recent preclinical studies showed that protein tyrosine phosphatase non-receptor type 12 (PTPN 12) appears to be an important tumor suppressor in breast cancer. However, the clinical and prognostic significance of PTPN 12 expression in breast cancer patients has not been elucidated yet.

#### Methods

PTPN 12 expression status was assessed for 183 patients who underwent curative surgery for operable breast cancer between May 2000 and August 2003, using immunohistochemical (IHC) assay of tissue microarray. Clinicopathologic characteristics, and expression status of ER, PR, EGFR, HER-2 were compared according to PTPN 12 expression status. The prognostic significance of PTPN 12 expression on disease-free survival (DFS) and overall survival (OS) was analyzed and the prognosis was also assessed in subgroups defined on the basis of major prognostic factors and ER, PR, HER-2, EGFR expression status.

#### Results

Loss of PTPN 12 expression was observed in 35.5% of the patients in this study. No significant differences were found in the clinicopathological characteristics and ER, PR, HER-2, EGFR expression status between PTPN 12 loss versus PTPN 12 intact patients. At a median follow-up of 9.9 years (range, 0.7-11.1), patients with PTPN 12 loss showed worse 10-year DFS than those with intact PTPN 12 expression (74.8% vs. 83.1%), but without statistical significance (p = 0.220). Ten-year overall survival of patients was not different according to PTPN 12 expression status. In the subgroup analysis performed in HER2 (-) and EGFR (-) patients (n = 128), the DFS was shorter in patients with PTPN 12 loss in univariate analysis (p = 0.057), and the difference was independently significant in multivariate analysis (hazard ratio, 2.97; 95% confidence interval, 1.05-8.37; p = 0.040). In the subgroup of patients who received hormone therapy alone as adjuvant treatment (n = 28), all recurrences were occurred in patients with PTPN 12 loss (5 recurrence in 16 patients) whereas there was no recurrence in patients with intact PTPN 12 expression.

#### Conclusion

PTPN 12 expression loss was associated to poor disease free survival in HER2 negative and EGFR negative breast cancer patients in this study. Our finding suggests that PTPN 12 may have a prognostic meaning predicting recurrence in HER2 negative and EGFR negative breast cancer patients, and future validation is warranted.

#### Disclosure

All authors have declared no conflicts of interest.

## 354PGENETIC INTERACTION PROFILE MAY PREDICT BEVACIZUMAB (BV) EFFICACY IN METASTATIC BREAST CANCER (MBC) PATIENTS (PTS): AN EXPLORATORY RETROSPECTIVE ANALYSIS

### Abstract

#### Background

previous retrospective studies have attempted to identify a possible role of VEGF single nucleotide polymorphisms (SNPs) to predict BV efficacy in terms of OS and PFS in MBC pts with conflicting results (Schneider 2008, Grimaldi 2011, Lambrechts 2011).

#### Methods

on the basis of these preliminary data, we decided to assess in a MBC population if different VEGF, VEGFR-2, IL-8, IL-6, HIF-1alfa, EPAS-1 and TSP-1 genotypes could predict first line BV + Paclitaxel (P) response in terms both of OS and PFS. Analyses were performed on germline DNA obtained from blood samples. Fourteen polymorphisms were investigated by real-time PCR technique. Both single and combinations of SNPs were investigated. The multifactor dimensionality reduction (MDR) methodology was applied to identify a genetic interaction profile for PFS (http://sourgeforge.net/projects/mdr/).

#### Results

102 pts have been enrolled from 8 Oncology Units. Main pts characteristics are: median age 59 years (range 32-81), ECOG-PS 0/1 in 78%/22%, hormone receptor positive 83%, previous adjuvant chemotherapy 68%, disease free interval (DFI) < 12 months 27%. After a median follow up of 17.4 months (1.9-54.7), mPFS was 11.6 months (95% CI: 10.6-12.6) and mOS was 32.4 months (95% CI: 25.9-38.9). None of SNPs were individually associated with PFS. Conversely, a genetic interaction profile consisting of VEGFR-2 rs11133360 and IL-8 rs4073 was significantly associated with PFS. mPFS was 14 months (95% CI: 11.7-16.3) and 10.9 months (95% CI: 9.3-12.4) for the favorable and unfavorable genetic profile, respectively (HR = 0.63, 95% CI: 0.4-0-99, p= 0.046). Furthermore, at the multivariate analysis hormone receptor positive (HR = 0.22, 95% CI: 0.12-0-41, p < 0.0001), DFI >12 months (HR= 0.4, 95% CI: 0.2-0.82, p= 0.011) and BV maintenance (HR = 0.63, 95% CI: 0.25.0.71, p = 0.001) were significantly associated with a better PFS.

#### Conclusions

genetic interaction between VEGFR-2 rs11133360 and IL-8 rs4073 polymorphisms could predict BV response in terms of PFS. With a longer follow up correlations with OS will be investigated. Prospective study is planned. Study supported by the no-profit foundation F.A.R.O.

#### Disclosure

All authors have declared no conflicts of interest.

## 355PFIRST REPORT OF LONG-TERM RESPONDERS TO FIRST-LINE BEVACIZUMAB (BEV) COMBINED WITH CHEMOTHERAPY IN TWO INDEPENDENT COHORTS OF HER2-NEG METASTATIC BREAST CANCER PATIENTS (PTS) WITH HORMONE RECEPTOR –POSITIVE (HR+) AND TRIPLE-NEGATIVE (TN) TUMORS

### Abstract

#### Background

First-line BEV combined with weekly paclitaxel, docetaxel or other chemotherapy significantly improves progression-free survival (PFS) in HER2-negative metastatic breast cancer (mBC), as shown in E2100, AVADO and RIBBON-1 trials. In the ATHENA study, 21% of pts continued BEV for ≥1 year with no new safety outcome and a time to progression of 19.9 months (95% CI 18.9-21.8 months). To further understand and provide insight into the efficacy and safety of long-term responders to first-line BEV, we conducted a descriptive study of 2 different cohorts of pts with mBC: HR+ and TN, treated in routine oncology practice with at least 1 year of first-line BEV.

#### Methods

Pts who had received first-line BEV(≥1 year) associated to chemotherapy were retrospectively included in the 2 independent cohorts and followed up, if they were alive at inclusion, for 18 months. Clinico-pathological characteristics, treatment received, efficacy and safety data were collected.

#### Conclusion

Prolonged BEV-containing therapy in this HR+ cohort is of interest and suggests that some pts achieve sustained disease control with limited side effects.

#### Disclosure

V. Diéras: Advisory boards and Symposium participation.

H. Simon: Member of an advisory board: Roche Mastology Group.

N. Mesnard: Roche employee.

E. Antoine: Membership of an advisory board.

All other authors have declared no conflicts of interest.

## 356PBEVACIZUMAB–CAPECITABINE (BEV–CAP) AFTER INITIAL 1ST-LINE BEVACIZUMAB–DOCETAXEL (BEV–DOC) IN PATIENTS (PTS) WITH HER2-NEGATIVE METASTATIC BREAST CANCER (MBC): SAFETY ANALYSIS OF THE IMELDA TRIAL

### Abstract

#### Background

Combining BEV with 1st-line chemotherapy (CT) significantly improved progression-free survival (PFS) and response rate (RR) vs CT alone in 3 randomised phase III trials (E2100, AVADO, RIBBON-1). In AVADO, DOC 100 mg/m2 was given for a median of 5.5 months (maximum 9 cycles), after which BEV 15 mg/kg was continued as a single agent. IMELDA was designed in the context of BEV–DOC approval to determine whether efficacy is improved by adding CAP to BEV maintenance therapy after discontinuing DOC.

#### Methods

Eligible pts had HER2-negative measurable mBC, ECOG PS 0/1, had received no prior CT for mBC and were eligible for taxane-based CT. After initial 1st-line therapy with 3–6 cycles of BEV–DOC, pts free of progressive disease (PD) were randomised to BEV alone or BEV–CAP (BEV 15 mg/kg q3w; CAP 1000 mg/m2 bid d1–14 q3w) until PD. The primary endpoint is PFS; secondary endpoints include safety, RR and overall survival.

#### Results

Between June 2009 and March 2011 (when enrolment to the study was terminated prematurely after withdrawal of BEV–DOC regulatory approval), 284 pts began BEV–DOC. Median age was 52 y (range 24–80), 17% were aged ≥65 y, 53% had ECOG PS 0 and 30% had triple-negative mBC. Maintenance therapy was administered in 183 pts (64%). Key safety results are below.

Table: 356P

Pts, n (%)  Initial
BEV–DOC phase (N = 284)
Maintenance
BEV ± CAP phase (N = 183)
Discontinued therapy  99 (35) 134 (73)
PD AE Other/unknown 41 (14) 31 (11) 27 (10) 101 (55) 15 (8) 18 (10)
Grade ≥3 AEs, n (%)  138 (49) 63 (34)
Neutropenia Febrile neutropenia Diarrhoea Mucosal inflammation Asthenia Hand-foot syndrome 44 (15) 30 (11) 3 (1) 9 (3) 6 (2) 6 (2) 2 (1) 0 2 (1) 2 (1) 1 (1) 25 (14)
Grade ≥3 BEV AEs of special interest Hypertension Proteinuria Bleeding Arterial TE Venous TE Wound-healing complication GI perforation 5 (2) 1 (<1) 9 (3) 2 (1) 2 (1) 2 (1) 5 (2) 9 (5) 3 (2) 2 (1) 2 (1) 2 (1) 0 0
Pts, n (%)  Initial
BEV–DOC phase (N = 284)
Maintenance
BEV ± CAP phase (N = 183)
Discontinued therapy  99 (35) 134 (73)
PD AE Other/unknown 41 (14) 31 (11) 27 (10) 101 (55) 15 (8) 18 (10)
Grade ≥3 AEs, n (%)  138 (49) 63 (34)
Neutropenia Febrile neutropenia Diarrhoea Mucosal inflammation Asthenia Hand-foot syndrome 44 (15) 30 (11) 3 (1) 9 (3) 6 (2) 6 (2) 2 (1) 0 2 (1) 2 (1) 1 (1) 25 (14)
Grade ≥3 BEV AEs of special interest Hypertension Proteinuria Bleeding Arterial TE Venous TE Wound-healing complication GI perforation 5 (2) 1 (<1) 9 (3) 2 (1) 2 (1) 2 (1) 5 (2) 9 (5) 3 (2) 2 (1) 2 (1) 2 (1) 0 0

AE = adverse event; TE = thromboembolic event

#### Conclusions

Most grade ≥3 AEs during BEV–DOC were typical DOC-related AEs. Apart from hypertension and proteinuria, BEV AEs occurred predominantly in early cycles, suggesting that long-term BEV is well tolerated. Efficacy results are expected in 2013.

#### Disclosure

J. Gligorov: JG has sat on Advisory Boards and received speaker honoraria from Roche.

J. Bines: JB has served on Advisory Boards for Roche.

P.A. Cortes: PC has served on Advisory Boards for Roche, BMS and Sanofi-Aventis.

U. Freudensprung: UF works as a Contractor for F Hoffmann-La Roche Ltd.

G. Mustacchi: GM has acted as a Consultant for Roche, Agendia, Celgene, Novartis, Merck, Glaxo, Eisai.

All other authors have declared no conflicts of interest.

## 357PNON PEGYLATED LIPOSOMAL DOXORUBICIN BEYOND THE FIRST LINE TREATMENT OF METASTATIC BREAST CANCER PATIENTS: A RETROSPECTIVE ANALYSIS

### Abstract

#### Background

The aim of this retrospective study was to access the benefit of a non-pegylated liposomal doxorubicin (NPLD) treatment beyond the first line therapy in patients with metastatic breast cancer (MBC) after previous exposure to an anthracycline containing regimen.

#### Patients and methods

A pharmacy prospectively collected database in the Franche-Comte region, was used to identify a cohort of patients with MBC treated by NPLD regimen between 2003 and 2010.In total, 140 patients were included in the analysis. Progression-free survival (PFS) was chosen as the primary efficacy endpoint. Cox proportional hazard models were fitted to identify factors that could influence both PFS and overall survival (OS) lenght. Survival data were computed according to Kaplan Meier method.

#### Results

Primary tumours characteristics were oestrogen receptor-positive, progesterone receptor positive and HER2 positive in 77%, 63% and 20% respectively. Median PFS length was 4 months [95% CI: 2-5] and 33% of patients experienced a PFS longer than 6 months. Overall response (OR) was observed in 27 patients (19.3%, 95%CI [10.2-28.4]). Median OS after NPLD was 13 months [95% CI: 10-14]. In multivariate analysis, prognostic factors significantly related to longer OS from NPLD first exposure were age younger than 50 years old (HR = 0.57 [0.35-0.93], p = 0.02) and HER2 positive tumour (HR = 0.53 [0.30-0.95], p = 0.03). 17 patients (12.1%) discontinued NPLD treatment due to toxicity. In those, 9 (6.4%) patients experienced cardiac toxicity (grade 2 or 3) . There were 2 (1.4%) deaths related to cardiac toxicity.

#### Conclusion

Re-challenge by an anthracycline-containing regimen should been considered in patients with MBC taking into account this encouraging ORR and the length of PFS and response duration. However cardiac safety of patients must be carefully investigated during the treatment taking into account the number of treatment cessation due to cardiac damage. Of interest the subset of patients with HER2 positive tumour seemed to highly benefit from the NLPD. This finding emphasizes the sensitivity to anthracycline in this subset of HER2 positive tumours and oncologist should not forgot anthracycline containing regimen as a possible option in MBC after failure to numerous lines including chemotherapy and HER2 targeted agents.

#### Disclosure

All authors have declared no conflicts of interest.

## 358PPHARMACOKINETICS OF ERIBULIN MESILATE IN COMBINATION WITH CAPECITABINE IN PATIENTS WITH ADVANCED/METASTATIC CANCER: RESULTS FROM A PHASE IB DOSE-ESCALATION STUDY

### Abstract

#### Background

Eribulin is a microtubule dynamics inhibitor EMA approved for certain patients (pts) with locally advanced (LA)/metastatic breast cancer (MBC) who have received ≥2 prior chemotherapeutic regimens for advanced disease. Prior therapy should have included an anthracycline and taxane. This Phase Ib, open-label dose-escalation study assessed maximum tolerated dose (MTD), safety and pharmacokinetics (PK) of eribulin in combination with capecitabine in pts with advanced/metastatic cancer.

#### Methods

Pts received eribulin mesilate (2–5-min IV) by Schedule 1 (1.2, 1.6 or 2.0 mg/m2 on Day [D] 1) or Schedule 2 (0.7, 1.1 or 1.4 mg/m2 on D1 and D8) in combination with twice-daily oral capecitabine 1000 mg/m2 on D1–14 every 21 days. To assess PK and drug-drug interaction of eribulin, capecitabine and capecitabine metabolites in Cycle 1 and 2, samples were taken on D1 (pre-, 0, 0.25, 0.5, 1–4, 6 and 8 h post-dose), any time on D2–3 and D4–6, and D8 (Schedule 1: pre-, 0.5, 1–4 and 6 h post-capecitabine; Schedule 2: pre-dose). PK was examined by non-compartmental analysis, and cardiac repolarization by 12-lead ECGs at screening, D1 (pre- and 4 h post-dose), D2–3, D8 and D15. Correlation of eribulin and capecitabine plasma concentrations with change from baseline (Δ) QTc was explored.

#### Results

Schedule 1 (n = 19) and Schedule 2 (n = 15) MTDs were 1.6 and 1.4 mg/m2 eribulin mesilate, respectively. Dose-limiting toxicities (all n = 1) were: Grade (G) 4 neutropenia, G3 febrile neutropenia, G3 fatigue, G3 lethargy (Schedule 1); G4 neutropenic sepsis, G3 neutropenia (Schedule 2). There were no unexpected toxicities. Eribulin PK was independent of schedule and had dose-related increases in exposure. No accumulation occurred upon multiple dosing; at each dose, eribulin PK was comparable in Cycles 1 and 2. Exposure to capecitabine and metabolites was variable and independent of schedule. Co-administration had no effect on ΔQTc.

#### Conclusions

No drug-drug interaction of eribulin and capecitabine was observed. From these results, the combination appears to be tolerated without effect on cardiac repolarization. A Phase II LA/MBC study evaluating Schedule 2 MTD is ongoing.

#### Disclosure

C.J. Twelves: The author declares the following conflicts of interest: consultant/advisory role (Eisai) and honoraria/speakers bureau (Eisai).

C. Savulsky: The author declares the following conflicts of interest: employee (Eisai Ltd).

C. Johnston: The author declares the following conflicts of interest: employee (Eisai Ltd).

L. Reyderman: The author declares the following conflicts of interest: employee (Eisai Inc.).

J. Wanders: The author declares the following conflicts of interest: employee at the time of study (Eisai Ltd).

R. Plummer: The author declares the following conflicts of interest: research funding (Eisai Ltd).

T.R..J. Evans: The author declares the following conflicts of interest: research funding (Eisai Ltd).

All other authors have declared no conflicts of interest.

## 359PACTIVITY AND SAFETY OF A COMBINATION OF EPIRUBICIN, DOCETAXEL AND CISPLATIN AS NEOADJUVANT TREATMENT FOR LOCALLY ADVANCED BREAST CANCER (LABC): A PRELIMINARY REPORT

### Abstract

#### Aim and background

Efficacy of cisplatin-containing triplet chemotherapy regimens for neoadjuvant therapy of LABC has not been investigated extensively. The aim of this study was to evaluate activity and safety of a combination of epirubicin, docetaxel and cisplatin (ETC) in the neoadjuvant setting.

#### Patients and methods

Forty-two patients with LABC (T2-T4, N0-N2, M0) were enrolled the study from March 2010 to April 2011. These patients received epirubicin (60 mg/m2 intravenously [I.V.] day 1), docetaxel (60 mg/m2 I.V. day 1) and cisplatin (60 mg/m2 I.V. day 1) every 21 days for at least 4 cycles, plus a final 2 additional cycles. Upon completion of therapy, the primary tumor was resected when not contraindicated. The primary endpoint was the pathological complete response (pCR) rate; seconday endpoints included response rate and toxicity.

#### Results

Median patients age was 48 years (range, 23-73 years). Median tumor size was 3.2 cm. Thirty-seven patients (88%) received 6 cycles of ETC. The overall clinical response rate was 78.6%. Twenty of 42 patients (47.6%) achieved a complete pathological response (CPR). All tumors became operable after neoadjuvant chemotherapy. The most common side effect was myelotoxicity. WHO grade 4 neutropenia developed in 30 (73%) patients. No recurrence was observed in any patient after a mean follow-up of 17 months (13-21 months).

#### Conclusion

Although our study group was small and the follow-up period relatively short-term, the considerably high rates of CPR in this preliminary series suggest that ETC regimen may be a promising option in neoadjuvant treatment of LABC. >

#### Disclosure

All authors have declared no conflicts of interest.

## 360PSEQUENTIAL DOCETAXEL AS ADJUVANT CHEMOTHERAPY FOR NODE-POSITIVE OR/AND T3 OR T4 BREAST CANCER: CLINICAL OUTCOME (MANSOURA UNIVERSITY)

### Abstract

#### Purpose

This trial compared six cycles of fluorouracil, epirubicin, and cyclophosphamide (FEC) with a sequential regimen of three cycles of FEC followed by three cycles of docetaxel (FEC-D) as adjuvant treatment for women with node-positive or/and T3 or T4 breast cancer.

#### Patients and methods

Between January 2006 and January 2010, 657 patients with operable breast cancer were randomly assigned to either FEC every 21 days for six cycles, or three cycles of FEC followed by three cycles of docetaxel, both given every 21 days. Radiotherapy was mandatory for all patients who had undergone breast conserving surgery. Radiation to the chest wall, supraclavicular area, was recommended following mastectomy. Hormone-receptor–positive patients received tamoxifen for 5 years after chemotherapy. The primary end point was 5-year disease-free survival (DFS).

#### Results

Median follow-up was 61 months. Five-year DFS rates were 74% with FEC and 78% with FEC-D (P= .013). Multivariate analysis adjusted for prognostic factors showed a 17% reduction in the relative risk of relapse with FEC-D. Five-year overall survival rates were 85% with FEC and 89.4% with FEC-D, demonstrating a 27% reduction in the relative risk of death (P= .014). The incidence of grade 3 to 4 neutropenia, the need for hematopoietic growth factor and incidence of nausea /vomiting were higher with FEC. Docetaxel was associated with more febrile neutropenia, stomatitis, edema, and nail disorders. Though rare overall, there were fewer cardiac events after FEC-D, attributable mainly to the lower anthracycline cumulative dose.

#### Conclusion

Sequential adjuvant chemotherapy with FEC followed by docetaxel significantly improves disease free and overall survival in node-positive or/and T3 or T4 breast cancer patients and. Although the magnitude of the benefit observed with FEC-D, differences in the toxicity profiles of FEC and FEC-D may influence the choice of treatment for patients

#### Disclosure

All authors have declared no conflicts of interest.

## 361PA RETROSPECTIVE ANALYSIS OF PLATINUM-BASED NEOADJUVANT CHEMOTHERAPY FOR LOCAL ADVANCED TRIPLE NEGATIVE BREAST CANCER

### Abstract

#### Purpose

We retrospectively analyzed platinum-based neoadjuvant chemotherapy for LATNBC to compare survival outcomes between patients with PCR and with non-PCR. Furthermore, the disease free survival of LATNBC patients with PCR continuously receiving primary regimen as adjuvant setting had comparative advantage concerning that of “PCR” patients switching to other regimens as adjuvant setting as well as those without any chemotherapy after sugery.

#### Patients and methods

124 women with stage II or III TNBCs experienced platinum-based regimens as neoadjuvant chemotherapy from Nov 1, 2007 to Dec 31, 2011. All patients were divided into the two groups, who were with and without PCR in the pathological reports after surgery. According to the adjuvant settings for LATNBC patients with PCR, the three arms were determined as continuous primary regimen (the same as neoadjuvant) arm, no more chemotherapy arm and switching arm. Disease free survival was computed using the Kaplan-Meier product limit method.

#### Result

We presented a retrospective chart review of 124 LATNBC patients who underwent platinum-based neoadjuvant chemotherapy in our hospital. Fifty (40.32%) of those patients receiving neoadjuvant chemotherapy had PCR when they underwent surgery. After controlling for covariates associated with survival, patients undergoing neoadjuvant chemotherapy with residual tumor had significantly worse survival than patients with PCR (HR = 0.37,P < 0.05). Of 50 patients with PCR confirmed by surgery, the disease free survival of 24 cases switching to other regimens in the adjuvant setting was significantly better than that of 24 cases continuously receving primary regimens in the adjuvant setting (HR= 0.51, P = 0.025)and that of 2 cases with no more chemotherapy(HR= 0.58, P = 0.017)

#### Conclusion

Patients with PCR had statistically significantly better clinical survival than those with non-PCR after platinum-based neoadjuvant settings.So far, if LATNBC patients with PCR after platinum-based neoadjuvant chemotherapy, they might have better clinical survival if they receive switching regimens than to receive primary regimens and to continue with no additional chemotherapy after surgery. A randomized prospective study needs to be carried out to strengthen the results because of statistical bias.

#### Disclosure

All authors have declared no conflicts of interest.

## 362PMIGHT EARLY METABOLIC RESPONSE BY 18F-FDG-PET/CT BE USEFUL TO SELECT PATIENTS (PTS) WITH BREAST CANCER (BC) WHO WILL NOT OPTIMALLY RESPOND TO PREOPERATIVE CHEMOTHERAPY (PCT)?

### Abstract

#### Purpose

To evaluate 18F-2-fluoro-2-deoxy-D-glucose positron emission tomography/computed tomography (18F-FDG PET/CT) changes between baseline and after 2 cycles of PCT in pts with early/locally advanced (E/LA) BC, with the aim to verify whether early metabolic assessment of response during PCT may have a role in clinical practice to enable early changes in the therapeutic strategy.

#### Patients and methods

Sixty pts with newly diagnosed E/LA BC received 6-8 cycles of anthracycline and taxane-based PCT. Fifty-eight pts underwent surgery, which consisted in breast conserving surgery or radical mastectomy; axillary node dissection was performed in all cases. Optimal pathologic response (pR) to PCT was defined as absence of cancer cells in breast and ipsilateral axillary lymph nodes. All other conditions were defined as nonresponders (pNR). Maximum standardized uptake value (SUV max) with 18F-FDG PET/CT was measured for each pathologic lesion at baseline and after 2 cycles of PCT. SUV max percentage changes (Δ-SUV) were compared with pR rate according to immunohistochemical (IHC) BC characteristics. Δ-SUV >50% defined a metabolic response (MetR).

#### Results

Thirteen (22%) of 60 pts achieved pR. According to IHC, pR rates where 16% in ERpositive/HER2negative (ER + /HER2-) pts, 29% and 27% in HER2-positive (HER2+) and triple negative (TN) pts respectively. Sensitivity of metR to identify pR was 100% in all three subgroups, but the specificity was low: 38% in ER + /HER2- pts (38%) was the highest value. In this subgroup of pts PET had a low positive predictive value (24%), while the negative predictive value was 100%, showing, compared to HER2+ and TN pts, the highest ability to correctly predict pNR (32%). At a median follow-up of 36.6 months, recurrence rate was higher in metabolic non-responders, particularly in the ER + /HER2- subgroup, in which Kaplan-Meier analysis of disease-free survival confirmed a significant difference (p = 0.0490).

#### Conclusions

PET assay after 2 cycles of PCT correctly predicted pNR in 32% of ER + /HER2- pts, identifying a subgroup of BC pts with worse prognosis who might benefit from an early change of the therapeutic strategy.

#### Disclosure

All authors have declared no conflicts of interest.

## 363PBREAST CANCER RECURRENCES AT THE CHEST WALL (BCRCW) WHEN STANDARD TREATMENTS (TX) HAVE FAILED: LYSO-THERMOSENSITIVE LIPOSOMAL DOXORUBICIN (LTLD) + MILD LOCAL HYPERTHERMIA (MLH)

### Abstract

#### Background

BCRCW has a poor prognosis, with disfigurement, pain, and restriction of movement. Study treatment consisted of LTLD that releases high concentrations of doxorubicin (Dox) in areas treated with mild hyperthermia at > 39.5°C. MLH kills tumor cells, selectively increases liposomal permeability in tumor microvasculature, releases Dox from LTLD, and promotes Dox tumor uptake.

#### Methods

We conducted a phase I study of LTLD + MLH in patients (pts) with BCRCW tumors < 3 cm deep who had failed all standard Tx including surgery, radiation, and chemotherapy (CTx). Pts received up to 6 LTLD/MLH Txs every 21 days. Dosing cohorts started at 40 mg/m2 and stopped escalation at 50 mg/m2. LTLD was infused IV over 30 minutes (min); then MLH was given by microwave or ultrasound. The thermal dose goal was 40°C-42°C for 60 min. Pharmacokinetic samples for total plasma Dox and doxorubicinol (Doxol) were taken at 0.5, 5, 10 and 24 hours after starting infusion.

#### Results

Eleven pts with a median of 4 prior CTx (range 2 – 12) were enrolled; 10 had recurred after prior anthracycline (AC). All pts received > 2 cycles. The within subject variability in Dox and Doxol exposure was small with mean Cycle 2 vs Cycle 1 ratios ranging from 0.99 to 1.06.

Cmax/dose (ng/ml)/(mg/m2Dox
Doxol
Cycle 1
499.82
0.46
Cycle 2
512.00
0.45
AUClast/dose ((ng*hr/ml)/(mg/m2Dox
Doxol
1,338.18
7.96
1,381.82
8.04
Cmax/dose (ng/ml)/(mg/m2Dox
Doxol
Cycle 1
499.82
0.46
Cycle 2
512.00
0.45
AUClast/dose ((ng*hr/ml)/(mg/m2Dox
Doxol
1,338.18
7.96
1,381.82
8.04

Two types of grade 3/4 toxicity were seen in > 5% of 42 cycles given: reversible neutropenia in 17 (40.5%) and reversible leukopenia in 9 (21.4%). One case (each) of mucositis (grade 1), chest wall thermal burn, and chest wall cellulitis (both grade 4) occurred, and no cases of cardiomyopathy or hand-foot toxicity were seen. The rate of clinically significant (> 6 point) QoL improvement on the FACT-B after 2 cycles was 54.5% (95% CI: 25.1% - 83.9%), including 1 lasting > 3 months. The local objective response rate was 45.5% (95% CI: 16.1% - 74.9%), with 1 complete and 4 partial local responses.

#### Conclusion

LTLD + MLH is safe and active in BCRCW pts with prior radiation and AC exposure. A phase II study is underway.

#### Disclosure

All authors have declared no conflicts of interest.

## 364PIMMUNOHISTOCHEMICAL PREDICTORS OF THE CLINICAL AND PATHOLOGICAL RESPONSE TO NEOADJUVANT CHEMOTHERAPY IN LOCALLY ADVANCED BREAST CANCER

### Abstract

#### Purpose

To determine predictive immunohistochemical characteristics of the tumor correlated with the response to neoadjuvant chemotherapy in patients with locally advanced breast cancer.

#### Material and methods

A prospective study of 87 breast cancer patients (Т2-4N0-M0) treated in N.N.Blokhin Russian Cancer Research Center 1997 - 2009. Tumor samples were taken prior to neoadjuvant chemotherapy (cor-biosy). Pathological criteria assessed were: histological variant, tumor grade, HR and Her-2/neu status, Ki67 expression and glycoprotein Pgp-170 status. After 4-6 cycles of chemotherapy all the patients underwent radical surgery. We quantified the response clinically and pathologically (Lavnikova's system of therapeutic pathomorphism evaluation).

#### Results

The experimental results showed the following response rates: complete response – 11 (12.6%), partial response – 54 (62.1%), stable disease – 21 (24.1%) patients. One patient had progression of the disease. Stable disease correlated significantly with positive Pgp-170 status (41.7% vs. 17.7%, р = 0.04), whilst negative Pgp-170 – with objective response (82.4% vs. 54.2%, р = 0.02). Pathomorphism was achieved as follows: lack of pathomorphism - 10(11.5%), first-degree pathomorphism 17(19.5%), second-degree – 28(32.2%), third-degree – 18(20.7%), fourth-degree – 14(16.1%). Rare histological variants had more often third-degree pathomorphism than ductal carcinoma (р = 0.02); G1 had more often fourth-degree pathomorphism than G2 (p = 0.002); high-degree pathomorphism was more frequently observed in patients with N2 status than with N0 (p = 0.006) and N1 (p = 0.009). Her-2/neu + tumors showed lower degree of pathomorphism that those with HR status (23.7% vs. 45.6%, р = 0.03). Pgp 170+ tumors had significantly more often no (p = 0.004) or first-degree pathomorphism (p = 0.004), and as a result – lower rate of high degree pathomorphism than in Pgp-170 negative tumors (12.5% and 47.1%, respectively).

#### Conclusions

Some pathological and immunohistochemical characteristics of the tumor can be assessed preoperatively and predict the likelihood of the clinical and pathological response to neoadjuvant chemotherapy.

#### Disclosure

All authors have declared no conflicts of interest.

## 365PIS POSTMASTECTOMY RADIOTHERAPY FOR PATIENTS DIAGNOSED WITH BREAST CANCER WHO HAVE RECEIVED NEOADJUVANT THERAPY REALLY NECESSARY?

### Abstract

#### Introduction

The role of adjuvant radiotherapy (RTA) after mastectomy in patients who have received neoadjuvant treatment is controversial and it has not been shown benefit in ramdomized clinical trials (ref ASCO-Guide lines). On the other hand, the low rates of local and distant recurrences after complete pathological response (RPc) question the value of local control disease.

#### Objective

To analyze the rate and profile of recurrence in breast cancer patients who have received neoadjuvant chemotherapy, mastectomy with and without radiotherapy.

#### Patients and methods

From the Department of Medical Oncology, University Hospital Reina Sofia we have identified 171 patients between 1997 and 2010 who have received neoadjuvant treatment. 36 patients have not received radiotherapy treatment after mastectomy. We present clinical and biological features of patients and their evolution.

#### Results

The mean age is 53 years, 100% are stage II and III (55.6% and 44.4%), infiltrating ductal histology (88.9%) and poorly differentiated tumors (44.4%). Regarding to hormonal status, 58.3% are hormone receptor positive and Her-2 33.3% and 21% are triple negative. The rate of pathological complete response (T and N) is 22.2%. at T of 41.7%.

With a follow-up mean of 19.92 months (SD: 13.9), we have found one recurrence of the disease localized in liver. At this moment we do not have any local recurrence.

#### Conclusions

Based on our experience and the lack of data supporting radiotherapy in these patients, we believe that radiotherapy should be assessed on individual basis, pending randomized studies provide direct benefit.

#### Disclosure

All authors have declared no conflicts of interest.

## 366PTHE IMPACT OF ADDITIONAL PROGNOSTIC INFORMATION OBTAINED FROM KI-67 CHANGES AFTER NEOADJUVANT CHEMOTHERAPY VARIES IN SUBTYPE OF BREAST CANCER

### Abstract

#### Introduction

A reduction in the Ki-67 index after neoadjuvant chemotherapy has been reported to be associated with a favorable prognosis. The present study investigates whether a reduction in Ki-67 may be a predictive surrogate marker of favorable prognosis in each subtype of breast cancer.

#### Methods

A total of 385 patients who received neoadjuvant anthracycline followed by taxane chemotherapy and subsequent surgery for invasive breast cancer were analyzed retrospectively. By immunohistochemistry (IHC), patients were divided into 4 subtypes (Luminal A, Luminal B, Triple negative and HER2). Ki-67 was examined by IHC in pre-treatment core needle samples and post-treatment surgical excisional specimens. The relapse-free survival (RFS) rate was compared among each subtype.

#### Results

The median follow-up period was 56 months. The rate of pathological complete response was higher for HER2 (34.8%) and Triple negative (24.3%) subtypes than for Luminal B (8.3%) and Luminal A (3.8%) subtypes (p < 0.0001). A reduction in Ki-67 was observed in 58.5%, 83.4%, 70.2% and 74.2% of patients in the Luminal A, Luminal B, Triple negative and HER2 subtypes, respectively. The differences in RFS between patients whose Ki-67 was reduced and those not reduced were statistically significant for Luminal B (81.4% vs. 50.0%, p = 0.003), Triple negative (74.8% vs. 43.5%, p = 0.003) and HER2 (82.7% vs. 59.0%, p = 0.008). However, for Luminal A, the difference in RFS was not observed depend on the changes of Ki-67 (78.8% vs. 75.3%, p = 0.78).

#### Conclusions

The reduction in Ki-67 after neoadjuvant chemotherapy is a predictive surrogate marker of a favorable RFS. However, this conclusion could not be applied to Luminal A subtype.

#### Disclosure

All authors have declared no conflicts of interest.

## 367PELECTROCHEMOTHERAPY FOR CHEST WALL RECURRENCE FROM BREAST CANCER IN OLDER WOMEN: ANALYSIS OF 55 PATIENTS

### Abstract

#### Background

The incidence of chest wall recurrence (CWR) after mastectomy for breast cancer (BC) ranges from 5 to 40%. It is a common finding that a number of patients are not suitable for radical resection or full-dose radiotherapy. When resistance to systemic therapies occurs, electrochemotherapy (ECT), an electroporation-based local drug delivery system, could represent a valuable treatment option for older women.

#### Methods

We analyzed a prospectively maintained database of 55 BC patients (median 70 years, range 38-88) with irresectable CWR who experienced disease progression after at least 2 lines of systemic therapies. Tumor response, response duration and toxicity profile were analyzed according to patients' age (<70 vs >70 years).

#### Results

The patients received a median of 2 ECT courses (range, 1-5). Younger (n = 27) and elderly (n = 28) patients were comparable for clinico-pathological features, except for the number of CW metastases (median 15 vs 8, respectively, P = .040). Two-month objective response was: complete 22 patients (40.0%), partial 26 (47.3%), no change 7 (12.7%). The complete response rate was significantly higher in the elderly group (57.7 vs 28%, P = .02). Pain and local toxicity scores were similar, but worsened with the increasing number of ECT applications. Median follow-up was 32 months (range, 6–53), 3-year local control rate was 70%. Thirty-three patients (60%) developed new lesions (NL) in non-electroporated areas (median, 6.6 months). Less than 10 metastases (P < .001), the narrower area of tumor spread on the CW (P = .022), endocrine- instead of chemotherapy (P = .025) and complete response after ECT (P = .019) were associated to longer NL-free survival. Older women showed a trend to a lower local tumor control compared to younger patients (2-year local progression-free survival, 83 versus 88%, P = .120). On the contrary, elderly patients reported a better 2-year NL-free survival (45 versus 22%, P = .095).

#### Conclusions

Older patients with CWR present fewer skin metastases and are more likely to achieve complete response after ECT. The satisfactory CW control without severe toxicity makes elderly women with refractory CWR suitable candidates for ECT application.

#### Disclosure

All authors have declared no conflicts of interest.

## 368PIS SURGERY OF THE PRIMARY TUMOR CONVENIENT IN METASTATIC BREAST CANCER?

### Abstract

#### Introduction

Stage IV breast cancer is treated primarily with systemic therapies. Excision of the primary breast cancer tumor in presence of synchronous distant metastases is a controversial argument, and a standard recommendation is not proposed by current guidelines. We performed a meta-analysis with the aim of pooling the existing survival data of surgery of the breast primary tumor in stage IV disease.

#### Materials and methods

We searched PubMed for publications including female, with histologically confirmed stage IV breast cancer at presentation and an intact primary tumor. Primary outcome was overall survival (OS) in patients treated with resection of primary breast cancer in presence of synchronous distant metastases. Secondary endpoints were PFS or TTP (and local and/or distant PFS or TTP, whichever reported). Hazard ratios (HRs) for survival when reported after multivariate analysis (with 95% confidence intervals) were obtained from publications and aggregated in a meta-analysis. A meta-regression weighted for extent of disease, ER/HER2 status, age, visceral or bone disease, rate of radiotherapy, and systemic therapies offered was also performed.

#### Results

15 articles were included in this meta-analysis (all retrospective case series), for a total of 15.378 patients. Surgery of the primary breast cancer appeared to be an independent factor for an improved survival in the multivariate analyses from the individual studies, with an HR of 0.69 (p < 0.00001). According to meta-regression, the survival benefit was independent of age, extent, site of metastatic disease and HER2 status, but was directly proportional to rate of patients exposed to systemic therapies and radiotherapy and inversely correlated to ER+ status of the population included.

#### Conclusions

Our pooled-analysis, reveals that surgery of an intact primary tumor, although associated with distant metastases, reduces the risk of death by 30%. This results are particular significant if local surgery is associated with systemic therapy and radiotherapy into a multimodality strategy. The surgical excision of a primary breast cancer in patients with stage IV disease – if feasible - should be discussed with and proposed to patients.

#### Disclosure

All authors have declared no conflicts of interest.

## 369PBEING THERE FOR WOMEN WITH METASTATIC BREAST CANCER. A PAN-EUROPEAN PATIENT SURVEY

### Abstract

#### Background

This two-part European patient survey was designed to help identify needs of women with metastatic breast cancer (MBC), through understanding their experiences of diagnosis, treatment and care.

#### Material and methods

The survey, initiated in March 2011, was conducted using a two-stage methodology. The first stage collected views on standards of MBC care and unmet needs of patients from 47 MBC-related patient groups in 8 European countries. This information was used to develop the second stage patient survey. The patient survey was designed to capture personal experiences of MBC diagnosis, treatment, information provision and to determine insights into the ‘trade-off’ between extending overall survival and side effects associated with MBC treatment. This online survey was open to women with locally advanced or MBC or their carers. All data were collected using anonymised local language questionnaires with responders recruited through local patient groups.

#### Results

A total of 230 responses were received from 17 identified European countries (94% of whom had locally advanced or MBC, and 6% of whom were adult carers). Although the overall experience of care was generally good to excellent (77%), there were still gaps in terms of treatment choice and information provision. Specifically, findings indicate that 32% of patients perceived treatment choice to be lacking. Overall, results showed that 67% of women with locally advanced or MBC believed life-extending treatment to be important so that they could spend more time with their family and friends; the same proportion judged their treatment worthwhile if it prolonged survival, irrespective of potential side effects. Additionally, 68% of responders would have liked more information about future medical treatments and research, with 57% wishing to receive this information from their oncologist.

#### Conclusions

These new survey findings highlight the unmet needs of women with MBC in Europe with respect to treatment choice and provision of information. Prolonging survival is a priority for most women, even if associated with toxicity.

#### Disclosure

C.J. Twelves: The Being There survey was supported by an educational grant from Eisai Europe Limited.

J. Stebbing: The Being There survey was supported by an educational grant from Eisai Europe Limited.

## 370PCNS METASTASES, SUBTYPES AND SURVIVAL IN BREAST CANCER: A POPULATION BASED STUDY IN EASTERN SWITZERLAND

### Abstract

#### Purpose

To examine tumor characteristics and outcomes associated with central nervous system (CNS) metastases in patients with metastatic breast cancer (MBC).

#### Methods

Patients listed in the regional cancer registry of St. Gallen-Appenzell with MBC between 2003-2009 were included. Estrogen- (ER), progesterone receptor (PR) and HER2 status were collected from pathology reports. Biologic subtypes were approximated using standard immunohistochemical markers. Survival status was assessed in January 2012. Multivariate logistic regression models were used to identify factors associated with the risk of developing CNS metastases and with survival >12 months (mt) after the diagnosis of CNS involvement.

#### Results

Overall, CNS metastases were observed in 170 (22%) of 773 patients with MBC included in the study. In the multivariate model, factors associated with CNS metastases were age <65 (odds ratio (OR) 3.40; 95% confidence interval (CI) 2.32-4.98) and biologic subtype. Compared to patients with ER- and/or PR-positive and HER2-negative tumors, the risk to develop CNS metastases was two-fold higher for patients with triple negative (TN) tumors (OR 2.10; 95% CI 1.17-3.76) and 72% higher for patients with HER2-positive tumors (OR 1.72; 95% CI 1.14-2.61). CNS involvement occurred rarely as first metastatic manifestation. In most patients (n = 137, 81%), CNS metastases were observed metachronically: in TN disease after a median time of 12 mt, and in the other subtypes after a median time of 19 mt (p < 0.01). Median survival of all patients after diagnosis of CNS metastases was 6 mt (interquartile range 2-14 mt) with no significant differences among subtypes (log-rank test p = 0.28). After diagnosis of CNS metastases, 48 patients (28%) lived longer than 12 mt and 14 (8%) longer than 24 mt. Age <65 (OR 3.50), surgical excision (OR 3.30) and radiotherapy of CNS (OR 4.10) were independently associated with survival >12 mt.

#### Conclusion

Patients with TN or HER2-positive tumors showed increased risk for CNS metastases. However, after diagnosis of CNS metastases only surgery and radiotherapy favorably influenced survival. New approaches to control CNS disease in young patients with these subtypes need to be developed.

#### Disclosure

All authors have declared no conflicts of interest.

## 371PMALE BREAST CANCER: THE EXPERIENCE OF AN ONCOLOGICAL CENTER

### Abstract

#### Background

Male Breast Cancer (MBC) is a rare condition comparing to Female Breast Cancer (FBC). Because of that, its understanding is weak and its treatment has often been extrapolated from FBC in spite of many important differences already suggested by many studies. Objective: To evaluate risk factors, recurrence and survival in male breast cancer.

#### Material and methods

Review of clinical files of 115 male patients diagnosed with breast cancer and treated in the IPO- Porto from 1976 to 2011. Evaluation of demographic, tumor- related variables, treatment and survival were done. Survival curves were calculated by the Kaplan-Meier method and compared using the log-rank test. Statistical significance was set at p < 0.05.

#### Results

The median age of the treated patients was 63 years-old (33-94 years-old). Prior or active tobacco use was observed in 36,1% and alcohol use 44,8%. Over-weight or obesity was registered in 66,7% of the cases. 48,8% of the patients reported a positive family history of cancer (17,4% of breast or ovarian cancer in a first-degree member), 8,3% were BRCA2 positive and 12,5% developed secondary malignancies. Most tumors were located in left breast (55,7%), were ductal carcinomas (91,2%), G2 (54,6%), stage III (46,4%) or II (26,8%), with positive hormonal receptors (90,4%) and with limphovascular invasion (67,4%). The majority of patients underwent surgery (95,6%), followed by chemotherapy, radiotherapy and hormonotherapy (26,1%) or radiotherapy and hormonotherapy (26,1%). 32,4% relapsed, in visceral non-hepatic sites (32,4%) or hepatic (17,6%). Median recurrence-free survival was 27,3% at five years. 5-years cancer- specific survival was 71,3%. In univariate analysis were prognostic factors the advanced stage (p < 0,001) and hepatic recurrence versus other sites (p = 0,014). The last factor was the only independent prognostic factor. 33,9% of the patients died for breast cancer.

#### Conclusions

This study confirmed many differences between MBC and FBC namely: older age at diagnose, high prevalence of family history, BRCA2 mutations and second malignancies. Cancer-specific survival is similar when adjusted to other factors such as stage and local of recurrence.

#### Disclosure

All authors have declared no conflicts of interest.

## 372PINCIDENCE OF BONE METASTASES AND SURVIVAL AFTER A DIAGNOSIS OF BONE METASTASES (BM) IN BREAST CANCER PATIENTS

### Abstract

#### Objectives

To measure crude and cumulative incidence of BM and cumulative survival after diagnosis of BM. BM were grouped by (i) BM only (ii) BM followed by visceral metastases (iii) visceral metastases followed by BM.

#### Methods

Kaplan-Meier and Cox regression database analysis of women with breast cancer diagnosed 1975-2006 and treated at Guys Hospital London, whose details were prospectively updated regularly till end 2010.

#### Results

Of 7064 women, 1589 (22%) developed BM by end follow-up (mean 8.4 years); 2254 (32%) were diagnosed with breast cancer < 50 years, and 4810 (68%) ≥50 years; 735 (14.4%) were classified as grade I, 2303 (45.3%) grade II and 2049 (40.3%) grade III; 1530 (27.8%) were estrogen receptor (ER) –ve and 3982 (72.2%) ER +ve. Of all BM, 535 (33.7%) patients were in group i; 871 (54.8%) in group ii and 183 (11.5%) in group iii. Incidence of all BM within 0-3 years from breast cancer diagnosis was highest in 1980 (64/1000 person years) and lowest in 2006 (11/1000 person years), with the decline most pronounced in 1985-1990. Cumulative incidence of BM after 5 years follow up was highest in 1976-1982 (0.25 [95% CI 0.23-0.27]), falling to 0.22 (0.20-0.24) in 1983-1989, 0.18 (0.10-0.14) in 1990-1997 and 0.095 (0.08-0.12) in 1998-2006.Risk of BM was significantly influenced by: calendar period of breast cancer diagnosis, HR was 0.77 (0.68-0.86) 1983-1989, 0.46 (0.40-0.54) 1990-1997 and 0.33 (0.28-0.40) 1998-2006, all vs 1975-1982; tumour grade, HR 1.23 (1.08 - 1.40) grade 3 vs 1-2; nodal status, HR 1.88 (1.60-2.21) 1-3 nodes vs 0 nodes and 3.95 (3.36-4.64) 4+ nodes vs 0 nodes; and tumour size, HR 2.00 (1.75-2.29) 2-5 cm vs <2cm and HR 3.31 (2.73-4.01) 5+ cm vs <2cm. Amongst women with BM diagnosed within 10 years of breast cancer treatment, 25% were diagnosed within 1 year, 50% within 2.5 years and 75% within 4.7 years. Median survival after BM diagnosis was 2.3 years in group i compared with 0.96 years in group ii and 0.91 years in group iii.

#### Conclusion

Incidence of BM has decreased in recent years and risk of BM is significantly affected by clinical tumour characteristics. Women with bone only metastases survive longer with their disease. BM remain an important target for prevention, treatment and palliation in breast cancer.

#### Disclosure

A. Taylor: I own shares with Amgen and work at Amgen.

All other authors have declared no conflicts of interest.

## 373PPREGNANCY-ASSOCIATED BREAST CANCER: CHEMOTHERAPY DURING PREGNANCY

### Abstract

#### Background

Pregnancy-associated breast cancer is a rare disease and treatment for these patients is a challenging problem. There are limited data about diagnosis and treatment for this disease, and outcome of fetuses who were exposed to chemotherapy during pregnancy.

#### Patients and methods

Forty-two pregnancy-associated breast cancer patients were treated from November 1999 to May 2012. We reviewed the data, including clinicopathological characteristics, therapeutic management, and fetal outcome of twenty-eight patients who received chemotherapy during pregnancy.

#### Disclosure

All authors have declared no conflicts of interest.

## 374PASSOCIATION OF CYP1A1 A4889G AND T6235C POLYMORPHISMS WITH INCREASED RISK AND AGGRESSIVENESS OF BREAST CANCER

### Abstract

#### Background

Estrogen and its metabolites, activated by cytochrome P450 enzymes (CYP1A1), participate in the origin and progression of breast cancer (BC). The CYP1A1 gene is highly polymorphic in humans. G and C variant alleles of CYP1A1 A4889G and T6235C polymorphisms encode enzymes with increased activity in activation of these compounds than the respective wild alleles. This study aimed to clarify the biological characteristics of the above-mentioned polymorphisms and their roles in the risk of BC.

#### Materials and methods

CYP1A1 A4889G and T6235C genotypes of 742 BC patients (median age: 52 years, 638 Caucasians, 104 non-Caucasians) and 742 healthy women (median age: 40 years, 638 Caucasians, 104 non-Caucasians) were obtained in genomic DNA by PCR and enzymatic digestion. The differences between groups were analyzed by the χ2 test. Power of analysis (PA) was used to determine the effect of sample size on the results obtained in the study.

#### Results

Patient and control samples were in Hardy-Weinberg equilibrium at CYP1A1 A4889G (χ2= 0.15, P= 0.70; χ2= 1.15, P= 0.28) and T6235C (χ2= 2.65, P= 0.10; χ2= 1.93, P= 0.16) loci. The frequency of CYP1A1 4889AG + GG genotypes was higher in patients than in controls (29.0% versus 23.2%, P= 0.004; PA= 93.0%). Carriers of the variant G allele were under a 1.50-fold (95% CI: 1.14-1.97) increased risk for BC than those with the wild AA genotype. CYP1A1 4889AG + GG (80.1% versus 19.9%, P= 0.01; PA= 99.0%) and 6235TC + CC (81.4% versus 18.6%, P= 0.02; PA= 99.0%) genotypes were more frequent in BC patients with histological grade III tumors when compared with those with grades I + II tumors. The CYP1A1 4889AG + GG genotypes was also more common in BC patients with histological grade III tumors than in controls (80.1% versus 23.2%; P= 0.04; PA= 72.0%). Carriers of the variant G allele were under a 1.35-fold (95% CI: 1.01-1.82) increased risk of occurrence of histological grade III tumors than others.

#### Conclusion

Our data suggest that CYP1A1 A4889G and T6235C polymorphisms alter the risk and aggressiveness of BC in Brazilian women. We believe that women with variant alleles of the above-mentioned genes should receive additional medical attention for disease prevention and early diagnosis.

#### Disclosure

All authors have declared no conflicts of interest.

## 375PTARGETING BREAST CANCER STEM CELLS FOR TREATMENT FAILURE CASES OF LATE STAGE BREAST CANCER

### Abstract

Targeting breast cancer stem cells (BCSCs) present in bone marrow and breast tissues is an attractive alternative for ‘event-free survival’ in breast cancer as it is an extremely heterogeneous disease. The invasive and mesenchymal properties of BCSCs with CD44 + /CD24low/ALDH1+ phenotype has made them a promising target for eliminating the metastatic capacity of primary tumors. We hypothesize that the ability to therapeutically attack stem cells will be decisive for the development of specific target therapies. Ten chemo-failure late stage patients with biopsy-proven triple-negative metastatic breast cancer were selected randomly. Breast cancer cells were isolated from whole tumor and cultured for in vitro drug sensitivity towards platinum, anthracycline and docetaxel and correlation was drawn between cell differentiation and drug response. Accordingly chemotherapy was designed for a particular patient. BCSCs were also isolated from the whole tumor, cultured and chemotherapy was designed. We detected chemo-failure in 65% cases for whole cell chemo-predictive assays, while BCSCs isolated from those non-responders responded to chemotherapy. Among positive samples, 89% of patients showed platinum sensitivity and rest were found to be anthracycline sensitive. No sensitivity towards docetaxel was observed. In lieu of this, cisplatin was applied ex vivo and percentage of BCSCs came down to 6.58% from an initial 11.16% (for a representative case). Thus the primary aim to target BCSCs to control metastasis and relapse of disease was somewhat obtained. We further plan to correlate ratio of selected markers present in patients in pre- and post-chemotherapeutic condition with time to recurrence, mortality, morbidity and progression-free survival.

#### Disclosure

All authors have declared no conflicts of interest.

## 376GENETIC POLYMORPHISM IN AURORA-A AS A RISK FACTOR IN EGYPTIAN WOMEN WITH BREAST CANCER

### Abstract

#### Introduction

Although many risk factors have been identified, the cause of any individual breast cancer (BC) is most often unknowable. Aurora A, the gene, encoding serine/threonine kinase, has been shown to be over expressed in many tumors, noticeably in BC. The aim is to study the single nucleotide polymorphism in Aurora-A, and compare the frequency distributions of different genotypes between patients with BC and those with fibroadenoma to define its tumorigenic contribution to BC development.

#### Patients and methods

This study was conducted on 60 pathologically confirmed BC patients, 20 patients with fibroadenoma and 40 frequency matched controls. There were no age, stage or histology restrictions.Serum CA-15-3&estradiol were measured. DNA was isolated from peripheral blood lymphocytes for genotyping of Aurora A at the T91A (Phe31Ile) site and were analyzed by PCR–RFLP assay.

#### Results

This study showed that women carrying the Ile/Ile genotype had an increased risk of developing BC (P = 0.04). Logistic regression analysis showed that subjects having IIe/IIe genotype had a 9.3fold increased risk of developing BC compared with those with the Phe/Phe genotype (OR 9.3;95% CI 1.12-77.69). The heterozygous Phe/IIe genotype was not significantly associated with the risk of cancer, suggesting a possible recessive effect of the polymorphism. No significant association was observed between the polymorphism and the risk of fibroadenoma. Patients with negative ER &PR were more likely to carry the IIe/IIe genotype compared with positive ER &PR positive patients (34.8% vs 5.4%). Serum estradiol showed significant increase in the BC category (p= 0.016), while, no significant difference was observed between the benign and the control categories (p = 0.41). The presence of +ve family history (FH) of BC and history of sex hormone intake showed significant increase in BC category when compared to control group and benign group. In BC group, patients with +ve FH had 7.5 times the chance of carrying IIe/IIe genotype compared to patients with –ve FH (OR7.5; 95% CI 1.53-36.71, P0.01).

#### Conclusion

This study provides the evidence that the Aurora-A Ile/Ile genotype is associated with an increased risk for the occurrence but not progression of BC.

#### Disclosure

All authors have declared no conflicts of interest.

## 377CLINICAL, MOLECULAR PROFILES AND RESPONSE TO NEOADJUVANT CHEMOTHERAPY IN PATIENTS WITH LOCALLY ADVANCED BREAST CANCER: AN INDIAN EXPERIENCE

### Abstract

#### Background

Breast cancer is now the most common cancer in many parts of India and Locally advanced breast cancer (LABC) accounts for nearly one thirds of cases. Advanced stages at presentation are attributed to late diagnosis and biologically aggressive disease in Indian women.

#### Materials and methods

Patients diagnosed with inoperable LABC at the Gujarat Cancer and Research Institute, Ahmedabad, India were included in the study. Baseline evaluation included clinical assessment, testing for Estrogen Receptor(ER), Progesterone Receptor (PgR), Her2Neu and serum Vascular Endothelial Growth Factor (S. VEGF). First line NACT was FAC or FEC. Patients were assessed clinically and radiologically at the end of 3 cycles for response evaluation (mammography or ultrasonography of breast). S. VEGF levels were repeated at the end of 3 cycles of NACT or before surgery. All resectable patients underwent a Modified Radical Mastectomy. Unresectable patients were offered taxanes based chemotherapy or supportive care.

#### Results

Fifty seven patients with LABC were included. Fifty Four patients (95%) received NACT. Mean age at diagnosis was 49 years. ER was positive in 51%, PgR in 35% and Her2Neu in 47%. Mean baseline S. VEGF was 463.4pg/ml. Baseline S. VEGF was higher in patients who were hormone receptor or Her2Neu positive. The overall clinical response rate [complete response (cCR) + partial response (cPR)] to the initial anthracycline based chemotherapy was 63%, cCR was 0%. Resectability rate after 1st line NACT was 77%. Pathological CR (pCR) rate was 4%. There was no significant reduction in S. VEGF levels after NACT, irrespective of clinical or pathological responses. Nearly 17% of patients were lost to follow up at various stages of treatment including three patients who refused any kind of treatment.

#### Conclusions

This study analyzes clinical and molecular profile of patients with LABC in India which differs from that seen in developed nations. High treatment drop-out rates may reflect treatment related toxicities, socio-cultural and logistical issues in a big developing country. Responses to anthracycline based NACT are comparable to the results reported from similar centers.

#### Disclosure

All authors have declared no conflicts of interest.

## 378RESPONSE TO NEOADJUVANT THERAPY AND DISEASE FREE SURVIVAL AND OVERALL SURVIVAL IN PATIENTS WITH TRIPLE-NEGATIVE BREAST CANCER (TNBC): SINGLE CENTER EXPERIENCE

### Abstract

#### Background

Triple negative breast cancer (TNBC) is a distinct subtype of BC wich is characterized by the absence of expression of estrogen receptor (ER), progesterone receptor (PR) or HER2/neu. TNBC is a very heterogeneous disease, that accounts for 15 to 20% of breast cancers. Despite initial chemosensitivity, patients with TNBC had a poorer outcome in terms of disease-free and overall survival, compared to non-triple-negative breast cancers. Neoadjuvant chemotherapy is commonly used for the initial treatment for TNBC, allowing for a higher rate of breast-conserving surgery and giving clues about the individual responsiveness of a particular cancer to chemotherapy.

#### Methods

We retrospectively reviewed 66 TNBC patients treated with neoadjuvant chemotherapy diagnosed at our institution between 2001-2011. They were divided into three subgroups: complete pathological response after neoadjuvant chemotherapy (group A), residual tumor smaller than 1 centimeter wide (group B), and residual tumor bigger than 1 centimeter wide or involving lymph nodes.

#### Results

(One patient died before surgery and another patient refused it. Twenty (31%) of the other 64 patients achieved a complete pathological response. Mean DFS was 2,15 years; recurrent disease was higher for C (20,64%) vs B (3pts, 23%) vs A (2 pts, 10%) (p = 0,0003). Most common recurrent sites were local (18) and bone (9). Mean OS was 4,5 years (95% IC 3,6 – 5,4); 3 (15%) pts died in A, 3 (23%) in B vs 16 (51%) in C (p = 0,03).

#### Conclusions

The pathologic complete response rate seen in our series is consistent with the one reported in other studies involving neoadjuvant chemotherapy for TNBC. The pathologic response after a neoadjuvant therapy correlates with disease-free and overall survival rates.

#### Disclosure

All authors have declared no conflicts of interest.

## 379ULTRASOUND ESTIMATION OF THE BLOOD FLOW IN BREAST CARCINOMAS AS A PREDICTIVE FACTOR OF EFFICIENCY OF NEOADJUVANT POLYCHEMOTHERAPY (NPCT)

### Abstract

#### Objective

Using ultrasound (US) techniques to determine the role of the intensity of the blood flow (BF) of breast tumors as a criterion of efficiency ACnPCT.

#### Methods

30 patients aged 50-62 years with verified BC, who were candidates for 3 courses of AC-nPCT, were carried out color-coded US in the triplex mode scanning and power doppler.

Luminal A Luminal B HER2(+) Triple-negative
T2N0M0
T2N1M0
T3N0M0
T3N1M0
Average systolic blood flow velocity, ACBFL (cm/sec) 14,2 ± 2,6 15,9 ± 3,9 14,6 ± 2,5 16,2 ± 2,8
Luminal A Luminal B HER2(+) Triple-negative
T2N0M0
T2N1M0
T3N0M0
T3N1M0
Average systolic blood flow velocity, ACBFL (cm/sec) 14,2 ± 2,6 15,9 ± 3,9 14,6 ± 2,5 16,2 ± 2,8

Measurements were made of resistance index (IR), pulsating index (IP).

#### Results

ACBFV is increased and corresponds to the malignancy process, while higher rates were in patients with the triple-negative breast cancer (16,2 ± 2,8 cm/sec) and luminal B type (15,9 ± 3,9 cm/sec). While luminal A type had ACBFL 14,2 ± 2,6 cm/sec and HER2(+) had 14,6 ± 2,5 cm/sec. IR in all subgroups was 0,73 ± 0,12, that matches the sharply increased values, and shows an increase peripheral resistance in vessels. IP was 12,35 ± 1,75, that reflects the amplified elastic and flexible properties of the newly formed blood vessels. Effect of AC-nPCT was higher in patients with the luminal B type (CR – 5, PR – 3). In patients with triple-negative BC effect of nPCT was worse, despite the high rate of BF in the tumor (PR – 1, stabilization – 6, progression – 2). In patients with low rate of BF in the tumor (luminal A, HER2(+) type) effect was the worst (PR – 2, stabilization – 7, progression – 4 patients in both groups). The correlation coefficient was statistically significant for BF as increased ACBFV, IR, IP, and the frequency of CR + PR to nPCT of luminal B type.

#### Conclusions

There is a correlation between the velocity of BF and degree of destruction of tumor by cytostatics. While the rate of BF depended on the molecular type of the tumor and did not depend on the size of the primary tumor.

#### Disclosure

All authors have declared no conflicts of interest.

## 380IMPACT OF BODY MASS INDEX (BMI) ON DISEASE FREE SURVIVAL AND LIKELIHOOD OF PATHOLOGIC COMPLETE RESPONSE IN PATIENTS WITH LOCALLY ADVANCED BREAST CANCER TREATED WITH NEOADJUVANT CHEMOTHERAPY

### Abstract

#### Background

The BMI is a clinical parameter that although not perfect is often used to measure adiposity. In breast cancer there are multiple studies indicating that overweight/obesity (O/O) is related with lower survival and increased risk of relapse. It has been also reported less likely to achieve a pathological complete response (pCR) after neoadjuvant chemotherapy (NAC) in O/O patients.

#### Methods

We retrospectively reviewed the records of 108 patients diagnosed with invasive locally advanced breast cancer (ILABC) who had been treated with NAC (anthracycline + taxane ± trastuzumab). The aim of our study was to review the impact of BMI on the pCR and the possibility of recurrence. pCR was defined as the criterion of strict pCR breast+nodes.

#### Results

From 2004 to 2011, 108 patients received NAC. Based on their weight and height at baseline we divided into two groups; group 1: underweight/normal (BMI <25 kg/m2) and group 2: O/O (BMI ≥25 kg/m2). Fifty one (47.2%) patients were in group 1 and 57 (52.8%) were in group 2. Median age was 46 years in group 1 and 52 in group 2 (p = 0.012). There were no differences at cTNM stage in both groups (p = 0.269). The NAC dose intensity was not different between groups (p = 0.399) Hormonal receptor negative tumors were more frequent in group 2 [27 (71.1%)] than in group 1 [11 (28.9%)] (p = 0.005). Likewise the triple negative receptor status was more frequent in group 2 [20 (83.3%)] than in group 1 [4 (16.7%)] (p = 0.001). Thirty-four patients had HER2 overexpression but was not different in both groups. (p = 0.695). There was a pCR in 18 patients (35.3%) from group 1 vs 11 patients (19.3%) from group 2 (p = 0.061). O/O patients were less likely to have pCR (OR 0.43 CI 95% 0.18-1.0). Median follow-up was 26 months (7-84). Twelve patients relapsed (3 local, 8 systemic or both 1). There were more relapses in group 2 (10 patients vs 2) (log rank 0.014). In an adjusted Cox regression analysis the BMI ≥25 kg/m2 was an independent factor for disease recurrence (HR 5.3 CI 95% 1.1-26.1).

#### Conclusions

Despite the retrospective nature of this study we can conclude that the patients with ILABC and O/O have a high risk for relapse and decreased response to NAC. It requires well-designed prospective studies to control confounding factors (dose intensity, chemotherapy regimens, changes in BMI) to get clear answers of these associations.

#### Disclosure

All authors have declared no conflicts of interest.

## 381STUDY OF TRIPLE NEGATIVE BREAST CANCER AND LYMPH NODE INVOLVEMENT AMONG EASTERN INDIAN PATIENTS- AN INSTITUTIONAL EXPERIENCE

### Abstract

#### Background

Breast Cancer is the most commonly diagnosed cancer among Indian women. Recent reports from India suggest a significant increase in the incidence of Estrogen and Progesterone receptor negative (ER-/PR-) breast cancer. Recently the targeted therapies improving the prognosis, so, know about receptor status becoming important. Yet very little is known about the specific histological subtypes or receptor status of breast cancer in women of eastern India. So a hospital based study was undertaken.

#### Objectives

This study was conducted to understand the frequency of occurrence of specific breast cancer subtypes and risk factors associated with the age, histologic subtypes among the patients and find the clinical features in relation to receptor status.

#### Methods

224 breast cancer patients, who are operated during the period of Mar 2009 to Nov 2011 from our cancer hospital, were selected for the study. Personal history (i.e. Age, age of Menarche and menopause status etc), case report and histopathology report were analyzed. Statistical analysis was performed using Software SPSS version 16.0

#### Results

Average age of the Patients 46.9 ± 1.32
Invasive- Ductal 66.51%
Invasive- Lobular 10.71%
Others (Fibrodenoma, infiltrating duct carcinoma etc.) 22.76%
Average Lymph Node Size 1.85 c.m ± 0.09
ER/PR Breast Cancer 45.53%
ER/PR/Her2neu (Triple Negative) 23.21%
Average lymph node number of Triple negative cancer patients 12 ±0.54
Among the triple negative patients average age 44.9 ± 3.24
Average lymph node number of without Triple negative cancer patients 5.71 ±0.44
Her2neu+ 56.25%
Average age of the Patients 46.9 ± 1.32
Invasive- Ductal 66.51%
Invasive- Lobular 10.71%
Others (Fibrodenoma, infiltrating duct carcinoma etc.) 22.76%
Average Lymph Node Size 1.85 c.m ± 0.09
ER/PR Breast Cancer 45.53%
ER/PR/Her2neu (Triple Negative) 23.21%
Average lymph node number of Triple negative cancer patients 12 ±0.54
Among the triple negative patients average age 44.9 ± 3.24
Average lymph node number of without Triple negative cancer patients 5.71 ±0.44
Her2neu+ 56.25%

#### Conclusions

Invasive Ductal carcinoma is very common among the eastern Indian patients. Maximum patients contains Garde II tumor. Among ER, PR negative breast cancer patient 50% does not possess Her2Neu receptor (triple negative). The Triple negative cancer patients contain a significantly high number of Lymph node (p < 0.001) then other patients.

#### Disclosure

All authors have declared no conflicts of interest.

### Abstract

#### Background

It is well known that neoadjuvnat chemotherapy is acceptable for women with locally advanced breast cancer. However, it is not achieving consensus that what kind of regimen is most effective and tolerable, although lots of regimens and dosages were clinically used.

#### Materials and methods

We compared the patients who were received adriamycine and doxetaxel (AD) and adriamycin, cyclophosphamide followed by paclitaxel (AC-T) as neoadjuvant chemotherapy and then received operation from 1 January 2006 to 30 September 2011. The group of AD regimen was scheduled for 3 cycles of AD (50mg/m2 and 75mg/m2, respectively) with 3 weeks interval and then completes resection. The group of AC-T was scheduled for 4 cycles of AC regimen (50mg/m2 and 500mg/m2, respectively) and then 4 cycles of paclitaxel (175mg/m2) with 3 weeks interval and then completes resection.

#### Results

The patients who were enrolled in this study were totally 78(AD and AC-T were equally 38.) The significant differences of patients' characteristics between two groups were not observed. However, the significant differences were identified in hematologic toxicity including neutropenia more than grade 3 p < 0.001), neutropenic fever p < 0.001), dose reduction rate due to hematologic toxicity (p = 0.012) and chemotherapy induced anemia (p = 0.012)), although chemotherapy induced thrombocytopenia (p = 1.0) was not different between two groups. No differences were identified in non-hematologic toxicity including hepatic toxicity, gastrointestinal toxicity and peripheral neuropathy. The response of chemotherapy was no difference between two groups, which was estimated by conversion rate of breast conserving surgery, clinical response of chemotherapy (p = 0.148), clinically downstaging rate (p = 0.464) and pathologic complete response rate (p = 1.0). There is no factor to predicting pathologic complete response or conversion to breast conservation in this study.

#### Conclusions

The neoadjuvant AC-T regimen is more tolerable with similar clinical outcome compared to AD regimen.

#### Disclosure

All authors have declared no conflicts of interest.

## 383WHY DO WOMEN WITH BREAST CANCER IN SARAWAK, MALAYSIA PRESENT LATE?

### Abstract

#### Introduction

Breast cancer (BC) is the most common cancer in Sarawak. We explore the reasons of late stage presentation in this study.

#### Methods

Based on a 175 questions questionnaire, information on barriers to late presentation were collected as a prospective study on 626 cases (2009 to 2011). Descriptive statistics and statistical tests were performed using the SPSS ver 17.0.

#### Results

The stage at diagnosis differed significantly with 71% of the Chinese being diagnosed at early stage compared to only 50% in Malay and 45% in Natives (p < 0.0004).

The delay between first symptom and first medical consultation (DELAY 1) was more than one month in 57% of the patients and it differed significantly among ethnic groups (50% for Chinese, 64% for Malay and 65% for Natives, p < 0.002). The highest delay were: women aged <30 years (75%) vs >50 years (53%, p < 0.037); from rural area (65%) vs urban areas (55%, p < 0.04).The main variables affecting this delay were knowledge about BC (p < 0.004), lack of interest in one's health ( p < 0.0005) and choice of first professional consulted (doctor/nurse vs traditional healer, p < 0.03).

The delay between first medical consultation and effective diagnosis (DELAY 2) for > 1 month was 14% of the patients and it differed significantly among ethnic groups (14% for Chinese, 23% for Malay and 13% for Natives, p < 0.0008). The main reasons: the number of doctors consulted before diagnosis (less the better, p < 0.0003); Malays (66.4%) and natives (64.4%) consulted more than two doctors when compared to Chinese (42.6%, p < 0.0001) and were less likely to follow the recommendations given by the doctors (p < 0.06).

The impact of teaching Breast self-examination (BSE) on DELAY 1 and 2: Seventy-eight percent of patients were taught BSE and by 80% of government nurses. The age group (30-40 years) had been taught BSE more than other age groups. For DELAY 1, 72% were not taught BSE, p < 0.0001. There was no difference for those with DELAY 2. However for both DELAY 1 and 2, 42% were not taught BSE, p < 0.0001.

#### Conclusions

More than 50% of the patients had DELAY 1 and 14% for DELAY 2. Learning BSE had an impact on DELAY 1 but not for DELAY2. Those in age groups (>40 to 60 years) had less BSE taught and this finding is crucial for public health education as most BC occur after 45 years.

#### Disclosure

All authors have declared no conflicts of interest.

## 384THE ROLE OF ANTIESTROGENS IN BREAST CANSER CELLS' INVASIVENESS

### Abstract

#### Aim

The current study investigates the effect of two selective antagonists of the estrogen receptor (ER), fulvestrant (FL) and tamoxifen (Tam), on the invasive ability of breast cancer cells. ER blockage using anti-estrogens is associated with a small induction of invasiveness. Moreover, matrix metalloproteinases (MMPs) contribute to this procedure by cleaving components of the extracellular matrix, forming a path for the migrating cells. In addition, the focal adhesion kinase (FAK) plays a central role in invasiveness.

#### Methods

We used two ER+ breast cancer cell lines, MCF7 and T47D. Cells were stimulated by estradiol (E2) and then treated with FL, Tam and the metabolites endoxifen (End) and 4OH-tamoxifen (4OHT). The invasiveness was evaluated using the matrigel assay and MMPs expression with gelatin zymography assay. Using immunofluoresence, we study the expression and localization of phospho FAK and its correlation with F-actin.

#### Results

We found that E2 exerts a protective role in invasiveness. On the contrary, the anti-estrogens, as well as their metabolites reversed the effect of E2, a finding that was more obvious with Tam. The effect of the agents on MMPs expression was more pronounced at 48 h, when Tam and 4OHT were found to reduce the levels of both MMP-2 and MMP-9 compared to FL. Following E2 exposure, the maximum autophosphorylation of FAK (Y397) was observed at 10 min. At the same time point we assessed the effect of our agents, regarding the spatial organization of actin fibers. The concurrent administration of E2 with FL, Tam or End was associated with rearrangement of cytoskeleton, a finding that was not observed in untreated cells, as well as after the use of 4OHT and E2 or E2 alone.

#### Conclusions

Our results indicate that the metabolite 4OHT is superior to the pro-drug Tam as well as the pure anti-estrogen FL with respect to invasiveness, MMPs induction and actin rearrangement. Regarding the comparison of Tam with FL, the effect on MMPs and rearrangement was similar. However, FL was found to be slightly superior to Tam concerning invasiveness.

#### Disclosure

All authors have declared no conflicts of interest.

## 385CANCER STEM CELL-LIKE CELLS: A THERAPEUTIC MODEL IN BREAST CANCER PATIENTS, WHERE ANY OTHER RECOMMENDED THERAPY HAS FAILED

### Abstract

#### Introduction

Nowadays, the difficulty to treat metastatic breast cancers is high. Many clinical therapeutic lines have failed and there is no suggested therapeutic approach (in literature) concerning this type of tumors. The last decades, circulating tumor cells (CTCs) are the state – of – the –art in cancer therapy. In the present study, CTCs were isolated and identified. CSCs (cancer stem cells) were isolated from the above population of CTCs and their gene pattern was compared with those from the primary tumor as well as with those from the metastatic tumor. This study attempts to find a correlation between the CTCs and the metastatic regions in comparison with the primary tumor as well as to find out if all the CSCs have the same hallmarks in the selected breast cancer stem cell populations.

#### Materials and methods

In order the protocol to be performed, CTCs from patient's blood samples were isolated and then cultured in appropriate conditions. CSCs were identified and isolated from the population of CTCs. mRNA was extracted and was used for Microarray hybridization assays. The same procedure was repeated for primary tumor as well for metastatic tumor samples. The expression pattern of primary tumor cells was compared with those in the metastatic tumor. Finally, the therapeutic approach which was based on the above findings, was designed.

#### Results

The results showed that the gene expression pattern of metastatic sites is similar to that of metastatic sites and less to that of the primary site. Then, the patients followed a therapeutic approach based on the data of the clinical results which were evaluated within a six- month period showing that the patients showed an objective response rate.

#### Conclusion

The results showed that the entity that determines the clinical outcome in breast cancer, is the sub-population of CTCs, the circulating CSCs. Moreover, there are various types of CSCs in one patient, and not only one, as they have different growth mechanisms in primary and secondary tumors.

#### Disclosure

All authors have declared no conflicts of interest.

## 386BREAST CANCER PATIENTS WITH HER2NEU OVEREXPRESSION: RELATIONSHIP AMONG CLINICOPATHOLOGICAL CHARACTERISTICS AND LOCAL/DISTANT RECURRENCES AND SURVIVAL

### Abstract

#### Objectives

The aim of this study was to correlate the overexpression of HER2 with clinical pathologic characteristics and its influence on local/distant recurrence and survival.

#### Methods

From 1998 to 2010, prospective data of 146 patients with invasive breast cancer with HER-2 overexpression was studied. The sample was divided into three groups: Negative hormonal receptors (NHR); Luminal B1 (Estrogen receptor +/-, Progesterone receptor +), Luminal B2 (Estrogen receptor +, Progesterone receptor -). Histological type (HT), size tumor (T), differentiation grade (DG), and nodal status were determined. Correlation with local and distant recurrence, and 5-year overall survival was done.

#### Results

NHR: 33.5%, luminal B1: 47.5%; Luminal B2: 19% Clinical pathologic characteristics:

Age (%)

Size(%)

HT (%)

<50 50-69 >70 T1 T2 T3 T4 Ductal Lobular
NHR 23 51 26 39 37 10 98
Luminal B1 33 51 16 51 39 99
Luminal B2 22 64 14 53 43 30 96
Age (%)

Size(%)

HT (%)

<50 50-69 >70 T1 T2 T3 T4 Ductal Lobular
NHR 23 51 26 39 37 10 98
Luminal B1 33 51 16 51 39 99
Luminal B2 22 64 14 53 43 30 96

DG (%)

Nodal Status

G1 G2 G3 Negative Positive
NHR 21 79 59 41
Luminal B1 11 37 52 48 52
Luminal B 28 73 57 43
DG (%)

Nodal Status

G1 G2 G3 Negative Positive
NHR 21 79 59 41
Luminal B1 11 37 52 48 52
Luminal B 28 73 57 43

Local and distant recurrence and overall survival at 5 years follow-up:

Local Recurrence(%) Metastasis Recurrence(%) Overall Survival (%)
NHR 20 81.5
Luminal B1 88.5
Luminal B2 15 86.5
Local Recurrence(%) Metastasis Recurrence(%) Overall Survival (%)
NHR 20 81.5
Luminal B1 88.5
Luminal B2 15 86.5

#### Conclusions

Patients with NHR and patients with PR- had more aggressive differentiation grade, less axillary infiltration and increased risk of distant metastasis. At 5-year follow-up, overall survival was better in the PR+ group.

#### Disclosure

All authors have declared no conflicts of interest.

## 387CRANIAL MAGNETIC RESONANCE IMAGING (MRI) IN THE STAGING OF HER2-POSITIVE BREAST CANCER PATIENTS

### Abstract

#### Purpose

The aim of the current study was to evaluate whether early detection of brain metastases could improve survival outcomes in HER2-positive breast cancer patients.

#### Patients and methods

HER2-positive breast cancer patients without neurological symptoms within 12 months from current stage of diagnosis were included in the study. The factors affecting the development of brain metastasis in the entire group and the metastatic patients were investigated, respectively. Survival durations after first metastasis and brain metastasis were compared between the asymptomatic and symptomatic patients

#### Results

Eleven of the 96 patients (11.5%) had occult brain metastasis. Of whom 9 patients were in the metastatic group, only 2 patients were in the non-metastatic group, (22% vs 3.6%, respectively, p = 0.008). Although median survival durations from first metastasis (27.4 vs 20.2 months, respectively, p = 0.858) and brain metastasis (5.2 vs 4.4 months, respectively, p = 0.710) similar, cerebral death was numerically different (22% vs 46%, p = 0.271) in the asymptomatic and symptomatic patients.

#### Conclusion

Present study suggests that routine cranial imaging do not bring any benefit to the non-metastatic HER2-positive breast cancer patients. Furthermore, although early detection of brain metastasis provides reduction in cerebral deaths, it does not prolong survival in metastatic patients.

#### Disclosure

All authors have declared no conflicts of interest.

## 388DIAGNOSIS OF BREAST CANCER METASTASES WITH PET/TC IN PATIENTS WITH ELEVATION OF TUMOR MARKERS: FINAL DATA

### Abstract

#### Background

Breast cancer is one of the most common cancers worldwide. PET / CT is more accurate than traditional methods for the detection of distant metastases or local recurrence and enables for early assessment of treatment response in patients after surgery for breast cancer undergoing primary chemotherapy. PET/CT is not considered a conventional examination during follow-up of patients with breast cancer, but recent data indicate its usefulness both in cases of asymptomatic increase of tumor markers that uncertain conventional imaging results. This study investigates the potential role of PET / CT to detect clinically occult metastases in patients with suspected recurrence of breast cancer during follow-up.

#### Methods

The authors studied 67 patients in breast cancer follow-up after primary surgery and chemotherapy and/or external radiotherapy. All patients were in remission without any other clinical or instrumental signs of relapses, except for the progressive elevation of CA 15.3 and/or CEA, tested during the follow-up. In 47 patients conventional imaging provided uncertain results and increase of CA 15.3 that was not correlated to any evidence of metastatic disease. The final diagnosis was obtained by histopathology (n.13) or by combined follow-up (n.54), including imaging at least 6 months later.

#### Results

Disease relapse was proven in 55 out of 67 patients and successfully PET/CT has identified clinically occult disease with an excellent sensitivity. In 21 cases the anatomical distribution of metastasis sites was in the bone, 16 in the lymphonode, 11 in the lung and 7 in the liver. We found 2 false-negative, 4 false-positive and 6 true-negative.

#### Conclusions

PET/TC may be more sensitive than the serum tumor markers in detecting relapse of breast cancer. This study demonstrated the clinical utility of tumor marker-guided PET in the follow-up of breast cancer patients.

#### Disclosure

All authors have declared no conflicts of interest.

## 389SURGICAL RESECTION OF LOCALLY ADVANCED PRIMARY TUMOR IN PATIENTS WITH DISTANT METASTATIC BREAST CANCER AT DIAGNOSIS: RESULTS OF A RETROSPECTIVE COMPARATIVE STUDY

### Abstract

#### Background

Women with metastatic breast cancer (MBC) with intact locally advanced primary tumors (LAT) present aggressive local symptoms that may warrant palliative surgery to the breast. However, it is unclear if such surgery otherwise improves clinical outcome. The aim of this study is to demonstrate if surgery of the breast may avoid uncontrolled chest wall disease and may improve survival.

#### Methods

We reviewed the records of all MBC patients presented with intact LAT, treated at our institution between 2007 and 2011.We compared two groups of patients: surgical group versus nonsurgical group. Clinical outcome was assessed in the two groups. Prognostic factors affecting loco regional relapse, were evaluated.

#### Results

75 patients were identified. The mean patient age was 49 ± 12.15 years. 52% were pre menopausal women. 87.1 % of tumors were hormone receptors positive. Her2 was assessed in 59 cases and was positive in 33.9%. Inflammatory breast cancer presented 16 %. Clinical lymph node involvement was noted in 58.7% cases. 69.6% had visceral metastasis and 5.3 % had brain metastasis. 89.3% have good Performans Status ≤ 1. All women received systemic therapy. First-line therapy consisted of anthracycline-based regimen (95.6%) and Taxane (39.7%). Among patients with HER2 positive 36.4% received Trastuzumab. Only 14% of patients with bone metastasis received bisphosphonate. 49.3% underwent mastectomy while 50.7% had intact LAT. The two groups were well balanced regarding demographics, clinicals, and tumors, characteristics. Among women who underwent mastectomy 48.5% had axillary lymph node dissection, and excision margins were positive in 25% cases. Loco regional radiotherapy (LRRT) was given to 8 women. pCR occurred in 7 patients among those who were operated. Local recurrence (LR) occurred in 9 patients (28.1%). Median time to local relapse was 3 months (2-19). LR was related to excision margin (p = 0.0001) and LRRT (p = 0.04). Median PFS was 10 months in nonsurgical group patients versus 16.5 months in surgical group.

#### Conclusions

MBC patients with locally advanced breast cancer who underwent mastectomy had improvement in local symptoms when the excision margin was in sano. However, the impact of this surgery in survival is not clear.

#### Disclosure

All authors have declared no conflicts of interest.

## 390CISPLATIN, CYCLOPHOSPHAMIDE, METHOTREXATE, AND 5-FLUOROURACIL (PCMF) AS FIRST LINE THERAPY OF METASTATIC TRIPLE NEGATIVE BREAST CANCER (MTNBC) IN PATIENTS PREVIOUSLY TREATED WITH ADJUVANT ANTHRACYCLINES AND TAXANES: RETROSPECTIVE ANALYSIS

### Abstract

#### Background

The purpose of this study was to evaluate the efficacy and toxicity of applied PCMF chemotherapy every four weeks (q4w) in the treatment of MTNBC patients (pts) previously treated with adjuvant anthracyclines and taxanes.

#### Methods

We administered cisplatin 30 mg/m2 on day 1 to 3 with regular hydration, cyclophosphamide 400 mg/m2 on day 1 to 5, 5-fluorouracil 500mg/m2 on day 2 and 4, and methotrexate 30mg/m2 on day 1, 3, and 5.The evaluation was performed after 4 cycles. We included 76 pts (median age 58 years, range: 32-71 years). PS medium was 1 (0-2); 21 pts (27.6%) were premenopausal and 55 (62.4%) postmenopausal. Adjuvant anthracyclines (AC or FAC x 4) and taxanes (paclitaxel/w x 12, or docetaxel/3w x 4) therapy was applied in all patients. DFS <12 months was obtained in 47 patients (61.8%) and >12 months in 29 pts (38.2%). Adjuvant radiotherapy was applied in 48 pts (63.2%). Visceral metastases were observed in 51 pts (67.1%), bone metastases in 34 (44.7%), and soft tissue metastases in 11 pts (14.5%). Only one organ was affected in 15 pts (19.7%), in 41 (53.9%) two organs, and in 5 pts (6.6%) three or more.

#### Results

We evaluated 76 pts with MTNBC after 4 cycles of PCMF chemotherapy. CR was found in 3 pts (3.9%), PR in 16 pts (21.0%), SD in 11 pts (14.5%), and PD in 39 pts (51.3%); with ORR in 19 pts (24.9%) and TCR in 30 pts (39.5%). Hematological toxicity: grade 3 and 4 anemia was found in 4 pts (5.3%) and leukopenia in 9 pts (11.8%). Non-hematological toxicity: renal grade 3 awas observed in 2 pts (2.6%). Not a single therapy was interrupted due to toxicity, but therapy prolongation was present in 15% of applied cycles.

#### Conclusions

The application of PCMF chemotherapy q4w in the treatment of patients previously treated with anthracyclines and taxanes as adjuvant treatment appeared to be effective with ORR in 24.9% and TCR in 39.5% and can be administered with good tolerance.It is today our standard treatment in this pts population.

#### Disclosure

All authors have declared no conflicts of interest.

## 391TRASTUZUMAB-BASED CHEMOTHERAPY (CT) FOR HER2-POSITIVE METASTATIC BREAST CANCER (MBC): WHICH OPTIMAL PARTNER BEYOND THE FIRST LINE? A SINGLE-CENTRE RETROSPECTIVE STUDY

### Abstract

#### Background

The activity of trastuzumab-based chemotherapy (CT) in HER2+ MBC is well established, but the question of the optimal CT partner remains a relevant issue. We performed a retrospective comparison of the clinical outcomes associated with different trastuzumab-CT regimens.

#### Patients and methods

Patients (pts) for this analysis were selected from a monoinstitutional database of HER2+ MBC pts receiving trastuzumab-based CT for the metastatic disease (February 2005-December 2008). Treatment activity and safety were assessed by the WHO criteria, time to progression (TTP) and overall survival (OS) were calculated by the Kaplan Meier method.

#### Results

A total of 147 pts with measurable disease were evaluated: 57 received trastuzumab with weekly vinorelbine (T/VNR), 48 weekly or 3–weekly trastuzumab plus docetaxel (T/Doc), 42 a triple combination of 3-weekly trastuzumab plus oral vinorelbine and capecitabine (T/osVNR/Cape) as 1st line therapy. Objective RR was 76% in the T/VNR group, 68% in the T/Doc regimen, and 72% in the T/osVNR/Cape combination. There was no significant difference in median TTP according to treatment type (14, 11 and 12 months, respectively, p = 0.62); median OS was 36 months, 32 and 34 months, respectively (p = 0.48). More treatment-related grade 3-4 toxicities were recorded in the T/Doc population. Following progression, pts received trastuzumab-based subsequent lines of treatment; according to our Institution police, shifting to a different CT partner: all had a 2nd line, 86 a 3rd line, 41 a 4th line, 7 a 5th line. In univariate analysis no visceral involvement, T/Doc 1st line CT, low primary tumor grading were correlated with better PFS and OS; in multivariate analysis the number of trastuzumab-based regimens was the only factor with statistically significant impact on OS (p = 0.02).

#### Conclusions

These results confirm the high activity of the tested regimens as 1st line therapy of HER2+ MBC, without significant differences in clinical outcomes, also suggesting the benefit of multiple lines of trastuzumab-based CT in a significant subset of pts, since each subsequent line may contribute to a longer OS.

#### Disclosure

All authors have declared no conflicts of interest.

## 392PHASE II STUDY OF SINGLE AGENT ORAL VINORELBINE (OV) AS FIRST-LINE CHEMOTHERAPY (CT) IN PATIENTS (PTS) WITH HER-2 NEGATIVE METASTATIC BREAST CANCER (MBC): A SINGLE CENTER EXPERIENCE

### Abstract

#### Background

Quality of life and pts' preferences play an important role in treatment decision-making in the metastatic setting. Previous studies indicated that Oral CT is convenient and a preferred option by many pts. We hereby report the efficacy and safety of OV as first-line CT for MBC.

#### Patients and methods

31 pts were enrolled between January 2007 and December 2010. All pts had measurable disease, a majority (84%) relapsing after anthracyclines (ANT) and/or taxanes (TXN) adjuvant treatment, WHO PS ≤ 2, adequate bone marrow, hepatic and renal functions and no adjuvant CT within the last 6 months. Pts were treated every 3 weeks with OV 60 mg/m2 D1 and D8 for the 1st cycle and thereafter 80 mg/m2 D1 and D8 every 3 weeks in the absence of G4 neutropenia and/or febrile neutropenia. Treatment was administered until disease progression or unexpected adverse event or pt refusal to continue. Primary endpoint (EP) was Objective Response Rate (ORR); secondary EPs were TTP, OS and safety. Follow-up results until April 2012 are reported.

#### Results

Median age was 42 years (range, 33 -75); median WHO PS 1 (range, 0-2). Previous adjuvant therapy: ANT-based alone: 29%, TXN-based alone: 19%, ANT plus TXN: 36%, other: 16%. Median disease-free interval from end of previous CT was 7 months. 26 pts (84%) had 2 or more metastatic sites, liver (61%), bone (58%), lung (58%) being the most frequent sites. A median of 6 cycles were administered (range, 2-20). ORR was achieved in 9 pts (29%), including 1 complete and 8 partial responses. 12 pts (39%) had stable disease, resulting in a clinical benefit rate (CBR) of 68%. In pts pretreated by ANT, ORR was 35% and CBR was 70%. Median TTP was 3.7 months [95% CI: 2.2-5.2]. Median survival was 16 months [95% CI: 11.4-20.6]. 3 pts (10%) developed G 3-4 neutropenia. No events of febrile neutropenia, cardiac, renal toxicities or alopecia were recorded. G 3 thrombocytopenia was reported in 2 pts (6%). 5 pts (16%) developed G 3 nausea-vomiting.

Conclusions: Results show a good efficacy and tolerance profile of OV as first line CT for HER-2 negative MBC pts. Similar activity was observed in the sub-group of pts pretreated by ANT.

#### Disclosure

All authors have declared no conflicts of interest.

## 393VINORELBINE WITH OR WITHOUT TRASTUZUMAB IN METASTATIC BREAST CANCER

### Abstract

#### Background

Vinorelbine is one of the most widely used drugs in metastatic breast cancer. We report a single center experience with vinorelbine with or without trastuzumab in patients with metastatic breast cancer.

#### Patients and methods

All patients with metastatic breast cancer receiving vinorelbine with or without trastuzumab during a six years period were retrospectively reviewed. Demographic data and data on response, time to progression (TTP) and survival were collected. Patients received vinorelbine IV 25-30 mg/m2or PO 60-80 mg/m2 in days 1 and 8 of a 21 days cycle. In patients who received concomitant trastuzumab a standard dosing schedule with 8 mg/kg loading dose followed by 6 mg/kg in subsequent administrations every three weeks was used.

#### Results

Eighty seven women were included. The median age was 63 years (range 32 to 85). Sixty two patients received vinorelbine alone and 25 patients received vinorelbine with trastuzumab. In 67 patients this was the first line treatment for metastatic disease and in 20 patients it was 2nd or later line of treatment. Seventy patients were evaluable for response while the remaining seventeen patients were not evaluable due to early progression (n = 6) or early termination of treatment for adversary effects (n = 11). The response rate of evaluable patients was 37.1% [1.4% Complete Response (CR) and 35.7% Partial Response (PR)]. Eighteen additional patients (25.7%) had Stable Disease (SD) for three or more months resulting in a Disease Control Rate of 62.8%. Twenty four of 54 (44.4%) patients receiving first line treatment had a response while in the second and subsequent lines setting two of 16 (12.5%) patients responded (x2 = 9.66, p = 0.001). A response was obtained in 63.6% of patients receiving concomitant trastuzumab and in 25% of patients receiving vinorelbine alone (x2 = 13.63, p = 0.0002). The median TTP was six months (range 1-45). Sixty six patients of the cohort have died and the median overall survival was 11.5 months (range 1-83). Adverse effects necessitating interruption of treatment were observed in 18.5% of patients.

#### Conclusion

This retrospective study of vinorelbine in metastatic breast cancer confirms a high disease control rate. Response rate is higher in first line treatment compared to subsequent lines and with the combination with trastuzumab.

#### Disclosure

All authors have declared no conflicts of interest.

## 394A RETROSPECTIVE STUDY OF CISPLATIN/VINORELBINE VERSUS CAPECITABINE/VINORELBINE AS SECOND-LINE OR THIRD-LINE TREATMENT IN ADVANCED BREAST CANCER

### Abstract

#### Purpose

To compare the efﬁcacy and safety of cisplatin or capecitabine, both with vinorelbine, as second-line or third-line treatment in advanced breast cancer previously treated with anthracyclines and/or taxanes.

#### Methods

From June 2004 to November 2011, 62 advanced breast cancer patients were eligible. Patients (38) enrolled in group NP received VIN 25mg/m2 on day 1 and 8 combined with Cisplatin 75mg/m2 on day 1 of a 21-day cycle. Patients (24) enrolled in group NX received VIN 25mg/m2 on day 1 and 8 of a 21-day cycle combined with CAP 1000mg/m2 twice daily for 14 consecutive days followed by 7 days of rest. Tumor assessment was performed every 2 cycles according to RESIST criteria. Toxicity was assessed according to National Cancer Institute Common Toxicity Criteria (version 3.0).

#### Results

The overall response rate (ORR) in group NP was 47.4%, all were partial responses (PRs). In group NX, ORR was 33.3% (P = 0.275), with 4.2% CRs and 29.2% PRs. Median time to progression (TTP) was 6.1 months (range, 3.2-9.0 months) in group NP and 6.3 months (range, 4.1-8.5 months) in group NX(P = 0.783). COX regression showed no statistically signiﬁcant difference (P = 0.782, OR = 0.920). Median overall survival (OS) was 28.8 months (range, 21.6-36.0 months) in group NP and 15.1 months (9.6-20.6 months) in group NX(P = 0.027, OR = 0.495). COX regression showed a statistically signiﬁcant difference(P = 0.045). Neutropenia was the most frequent hematologic toxicity, with 57.9% grade 3/4 neutropenia observed in group NP and 38.1% in group NX(P = 0.145). 13.2% grade 3/4 vomiting was seen in group NP and no grade 3/4 vomiting in group NX. No grade 3/4 nephrotoxicity or hand-foot syndrome was noted in both groups.

#### Conclusion

Better OS was seen in group NP than in group NX. Treatment-related toxicity in both groups was manageable.

#### Disclosure

All authors have declared no conflicts of interest.

## 395DOCETAXEL (D), GEMCITABINE (G) AND BEVACIZUMAB (BEV) AS SALVAGE CHEMOTHERAPY (CT) FOR HER-2 NEGATIVE METASTATIC BREAST CANCER (MBC)

### Abstract

#### Introduction

The combination of D and G is a valid salvage CT regimen for the treatment of mBC. When combined with CT, Bev has increased the response rate and progression-free survival (PFS) in both first and second line treatment of mBC. In this multicenter study, we evaluated the combination of D, G, and Bev, administered biweekly, as second or further line CT in patients (pts) with HER-2 negative mBC.

#### Methods

Women with HER-2 negative mBC, who had received at least one prior line of CT, were treated with D 50 mg/m2, G 1500 mg/m2 and Bev 10 mg/Kg, in cycles every two weeks. Bev was continued until disease progression. This was a study with Simon's optimal two step statistical design.

#### Results

Forty-eight pts have been enrolled. Median age was 61 years, performance status was 0-1 in 95.8% and 2 in 4.2% of pts, 83.3% had hormone receptor positive disease, and 47.9% had received one prior line of CT. All pts were evaluable for toxicity and 43 for response. Objective response rate was 44.2% (95% confidence interval [CI], 29.3-59%), median PFS was 6.8 months (95% CI, 4.5-9.1 months), and median overall survival 20.1 months (95% CI, 9.9-30.3 months). Grade 3-4 hematologic toxicity consisted of neutropenia (39.6%) and febrile neutropenia (4.2%). Most common grade 2-3 non-hematologic adverse events included nausea (10.4%), diarrhea (10.5%), neurotoxicity (12.5%), fatigue (31.3%), allergic reactions (6.3%), hemorrhage (4.2%), and hypertension (4.2%). No grade 4 non-hematologic toxicities or toxic deaths were recorded.

#### Conclusion

The combination of D, G and Bev has promising activity and a manageable toxicity profile as salvage CT for HER-2 negative mBC. Updated results will be presented at the congress.

#### Disclosure

All authors have declared no conflicts of interest.

## 396EFFICACY OF TRASTUZUMAB CONTAINING RETREATMENT AFTER PROGRESSION ON LAPATINIB THERAPY IN JAPANESE PATIENTS WITH HER2-POSITIVE METASTATIC BREAST CANCER

### Abstract

#### Background

Lapatinib has been approved for HER2 positive metastatic breast cancer patients with refractory to trastuzumab (T) therapy in Japan. Currently, it has been proposed that lapatinib can resensitize trastuzumab–mediated antibody-dependent cellular cytotoxicity (ADCC). Recent clinical data suggest the efficacy of T containing retreatment after progression on lapatinib therapy in patients with HER2-positive metastatic breast cancer. Here, we present a retrospective review of data from 25 patients who received T containing retreatment after progression on lapatinib therapy.

#### Methods

We reviewed the data of 50 patients with HER2-positive metastatic breast cancer who received lapatinib therapy in our institution from August 2004 through March 2012. Of these, 25 patients received T containing retreatment after progression on lapatinib therapy. We retrospectively assessed the clinical benefit of this treatment regimen in these patients.

#### Results

Luminal-HER2 and HER2-enriched subtypes were identified in 13 (52%) and 12 (48%) of these cases, respectively. The median duration of lapatinib therapy was 5.9 months (range, 1.8-20.2 months). The median number of preceding regimens was 3 (range, 2-8) in metastatic setting. Seven patients (31.8%) responded to T containing retreatment; all 7 patients achieved PR and none achieved CR. There were no significant differences in subtype, number of preceding regimens and brain metastases; however, responders achieved higher clinical response from lapatinib therapy than non-responders (response rate of 84% versus 13%, respectively). The median time to progression of T containing retreatment was 3.0 months (95% CI, 2.4-3.5 months). Among seven responders to T containing retreatment, one patient responded to refractory T containing regimen. All patients tolerated T containing retreatment with no occurrence of Grade 3/4 toxicities.

#### Conclusion

T containing retreatment could be a favorable treatment regimen which can achieve clinical response in patients with HER2-positive metastatic breast cancer who experienced progression on prior trastuzumab and following lapatinib therapy.

#### Disclosure

All authors have declared no conflicts of interest.

## 397LOW DOSES OF GEMCITABINE (G) WITH CISPLATIN (C) IN TREATMENT OF METASTATIC BREAST CANCER (MBC) PROGRESSING AFTER ANTHRACYCLINES, TAXANES, CAPECITABINE AND OTHER ANTINEOPLASTIC AGENTS

### Abstract

#### Background

The highest synergism of G was registered with C, as G suppresses DNA reparation after its damage from C, which leads to apoptosis of tumor cell. The study of optimal doses of G and C combination continues for intensive pretreated MBC.

#### Purpose

Evaluate efficiency and tolerability of the low doses of G and C in 131 patients with MBC progressing after anthracyclines, taxanes, capecitabine and other antineoplastic agents.

#### Methods

G 600-750 mg/m2 and C 30 mg/m2 were administered on days 1 and 8 every 3 weeks in the 2nd line for 28 patients, in the 3rd for 54 patients and in the 4th for 49 patients. Groups were comparable in age, performance status (PS), ER/PgR and HER2 expressions, localization and number of metastatic sites prior to treatment.

#### Results

Total amount of cycles: 549, median 4.2 per patient (range: 2-12). The overall response (OR) of all patients was 27%, clinical benefit (OR + stable disease) – 71.7%, with median time to progression (TTP) 4.8 months (95% CI 3.9-5.7 months) and median overall survival (OS) 14.6 months (95% CI 12.1-17.1 months). The OR in the 2nd line was observed in 39.3%, in the 3rd – 27.8% and in the 4th– 18.4% (р < 0.05), clinical benefit – in 85.7, 59.3 and 77.6% (р > 0.05) respectively. Pain intensity reduction and improvement of PS were noted in patients in spite of the lines. Linear discriminant and regression analyses showed that OR, TTP and OS didn't depend on the ER/PgR and HER2 expressions, but ECOG 2-3, pain syndrome, metastases in the lymph nodes and liver were the factors of poor prognosis. The treatment-related adverse events were neutropenia, anemia, thrombocytopenia, alopecia, nausea, vomiting, peripheral neuropathy and fatigue. Toxicities with grade 3-4 intensity were neutropenia (31.6% of patients), vomiting (0.8%), anemia (2.3%), thrombocytopenia (3.1%). No febrile neutropenia and deaths related to the study treatment occurred.

#### Conclusions

The low doses of G with C are effective and tolerable in treatment of MBC progressing after antracycline, taxanes, capecitabine, other antineoplastic agents not only in the 2ndand 3rd, but also in the 4th lines and are suitable for outpatient therapy.

#### Disclosure

All authors have declared no conflicts of interest.

## 398LONG TERM USE OF CAPECITABINE IN PATIENTS WITH LOCALLY RECURRENT OR METASTATIC BREAST CANCER

### Abstract

#### Background

Capecitabine monotherapy is considered standard treatment in anthracycline- and taxane-pretreated locally recurrent or metastatic breast cancer and has proven efficacy in this setting. Here we report the findings of a retrospective study to evaluate the impact of continuing capecitabine treatment beyond the standard 6 cycles until disease progression, on efficacy and outcomes in patients with locally recurrent or metastatic breast cancer. In addition we sought to consolidate the information known about capecitabine dose modification and efficacy outcomes in this patient population.

#### Patients and methods

Clinical databases in two institutions were used to identify all patients with locally advanced or metastatic breast cancer treated with capecitabine over the period July 2006 – July 2011. Baseline characteristics, progression-free (PFS) and overall survival (OS) were recorded.

#### Results

In total, 150 patients met the inclusion criteria. Of these patients 67 (45%) received less than 6 cycles, 17 (11%) received 6 cycles only and 66 (44%) received more than 6 cycles of capecitabine. 65 (43%) patients commenced treatment at a dose less than 1250mg/m2 twice daily and 60 patients (40%) received a dose reduction during treatment. Median overall survival was longer in those patients who received more than 6 cycles of capecitabine treatment (20.9 months) compared to those who stopped at 6 cycles by clinical decision (16.3 months) although this did not reach significance (P = 0.088). PFS and OS were similar among patients who received lower vs. full-dose capecitabine (PFS P = 0.32, OS P = 0.71).

#### Conclusions

In this retrospective analysis, patients receiving more than 6 cycles of treatment had a better OS than those who stopped at 6 cycles (by clinical decision). Continuing capecitabine monotherapy beyond the standard 6 cycles if well tolerated, should be considered in patients with locally recurrent or metastatic breast cancer as this may prolong PFS. In addition, these data support the practice of dose-reducing capecitabine, including the possibility of starting at a lower dose (<1250 mg/m2 twice daily), to reduce the incidence of adverse events without compromising efﬁcacy.

#### Disclosure

All authors have declared no conflicts of interest.

## 399ORAL CAPECITABINE AND VINORELBINE IN METASTATIC BREAST CANCER. A RETROSPECTIVE ANALYSIS OF TOLERABILITY AND ACTIVITY

### Abstract

#### Background

The purpose of this study was to retrospectively analyze toxicity profile and activity of an all-oral combination schedule of Capecitabine (Cape) and Vinorelbine (VNR) in metastatic breast cancer (MBC) patients (pts).

#### Methods

All pts treated had a histological confirmed diagnosis of breast cancer (BC). Each 3-week cycle of treatment consisted of 500 mg/m2 cape twice daily (2 weeks on, 1 week off), and 60 mg/m2 VNR on days 1 and 8.

#### Conclusions

Oral combination of Cape and VNR is a well-tolerated regimen and has an activity profile even in pre-treated MBC pts.

#### Disclosure

All authors have declared no conflicts of interest.

## 400MAMASTEAM: BEVACIZUMAB (BVZ) IN THE TREATMENT OF BREAST CANCER IN ADVANCED LINES OF TREATMENT, A NEW MODEL FOR THE ASSESSMENT OF ACTIVITY IN ADVANCED CANCER

### Abstract

#### Introduction

Bevacizumab is the first anti-angiogenic approved in the first-line treatment in advanced breast cancer. However, little is known about his activity in advanced lines. In the treatment of metastatic cancer, the response rate and time to progression decreases as we move forward in subsequent therapeutic lines [1]. Objective: To compare in each patient and in a paired way BVZ activity in advanced lines of treatment, achieved with the previous treatment.

#### Method

Coming of 5 Spanish hospitals have been selected 129 consecutive patients who received treatment with BVZ second-line treatment or later. For the efficacy analysis were used tests of Wilcoxon signed-rank and sign test.

#### Results

When analyzing the data of 114 patients (88.4%) had progressed. The responses obtained with the previous line of treatment (without BVZ) were: RC: 4 (3.1%), RP: 31 (24%), EE: 54 (41.9%), PROG: 36 (27.9 %) and NV: 4 (3.1%). The line obtained during treatment with BVZ: RC: 2 (1.6%), RP 31 (44.2%), EE: 37 (28.7%), PROG: 28 (21.7%) and NV: 5 (3.9%). With respect to time to progression (TP), 76 patients achieved a TP longer than they had achieved with the previous line of treatment (p = 0.011). The overall median time to progression was also higher: 4.06 vs 5.83 months (p = 0.032).

#### Conclusion

Opposite to it is expected, treatment with bevacizumab in advanced linesimproves the results obtained with previous treatment lines. This suggests that bevacizumab is active in advanced lines and produces favorable changes in the natural history of patients with metastatic breast carcinoma 1. Dufresne, A. et al. Impact of chemotherapy beyond the first line in patients with metastatic breast cancer. Breast Cancer Res Treat, 2008. 107(2): p. 275-9.

#### Disclosure

All authors have declared no conflicts of interest.

## 401PATTERN OF BONE METASTASIS AND TREATMENT OUTCOMES OF CHINESE TRIPLE NEGATIVE BREAST CANCERS

### Abstract

#### Objective

To investigate the characteristics of bone metastasis, prognosis and treatment strategies of triple negative breast cancer (TNBC) patients in China.

#### Methods

269 metastatic TNBC between Jan 1, 2005 and Dec 31, 2011 were retrospectively reviewed. Bone metastases were identified as primary metastatic site in 83 patients (30.8%). Among them, 25 patients (30.1%) had bone metastasis as the only relapse and the remaining were accompanied by involvement of other locations. Clinical pathologic characteristics and treatment outcomes were evaluated in these 83 cases.

#### Results

Median age was 48.2 (22-84) years. Median disease-free survival was 30.3 month (95% CI: 8.9-58.9 months). In these patients with bone metastasis as the primary distant site, most of them had multiple osteolytic bone lesions (78.3%). The incidence of skeletal related events was 24.1%. Chemotherapy was administered as the first choice in majority of patients (75 out of 83, 90.4%). Overall response rate was 44.0%, including 1 CR (1.3%) and 32 PR (42.7%). Median time to progression (TTP) of first-line chemotherapy was 8.0 (95%CI 8.7-16.0) months and median overall survival (OS) was 30.0 (95%CI 30.8-47.4) months. No differences were seen when comparing taxanes-based with non-taxanes regimen, platinum-based with non-platinum regimen and capecitabine-contained with non-capecitabine regimen in terms of response rate, TTP and OS.

#### Conclusions

Although the incidence of bone metastasis in mTNBC were not as high as in Luminal type, it was also common in Chinese patients, different from previous literatures reported in western countries. The survival of metastatic TNBC with bone metastases is poor. Chemotherapy was without doubt the primary treatment options, however, no the superiority of any chemo-agent have been proved now.

#### Disclosure

All authors have declared no conflicts of interest.

## 402RISK OF MENINGEAL INVOLVEMENT IN METASTATIC BREAST CANCER PATIENTS

### Abstract

#### Background

Meningeal Carcinomatosis (MC) occurs in 4-15% of patients with solid tumors, usually as a late event in the natural history of disease. Improvement in cancer treatments and patient survival and advances in neuro-imaging techniques have recently led to an increased incidence of MC, mostly in metastatic breast cancer. About 2-5% of breast cancer patients develop MC: median overall survival is 4-6 weeks without treatment and about 2-6 months when specific therapies are used. Objective: The aim of our study is to determine clinical, histological and immunological patterns of primary breast cancer which increase risk of MC.

#### Patients and methods

We performed a retrospective analysis of a cohort of 1037 consecutive patients treated for breast cancer from April 1994 to May 2010 at San Giovanni Battista Hospital in Turin. Diagnosis of MC was carried out in presence of neoplastic cells at cerebrospinal fluid citology. In case of negative or inconclusive citology, despite the presence of a characteristic clinical picture of MC, diagnosis was made using neuro-radiological findings, obtained by magnetic resonance imaging.

#### Results

We found 29 of 1037 breast cancer patients (2,8%) who developed MC. Among continuous variables associated with increased risk of MC (p < 0,001) there are young age (median age 46 years for MC positive versus 60.7 years for MC negative), tumor diameter (greater than 3.5 cm) and number of metastatic lymph-nodes (greater than 8). Discrete variables significantly associated with MC were: high tumor grade (p = 0,001), presence of vascular invasion (p = 0,038), greater tumor staging (pT3, pT4 p < 0,001) and axillar staging (pN3 p = 0,002), hormonal receptor negativity (p < 0,001) and Her-2 overexpression (p = 0,001). Multivariate analysis showed that only the presence of other distant metastases (p < 0,005) was significantly associated with high risk of develop MC with an odds ratio of 9.8.

#### Conclusions

We identified breast cancer patients at high risk to develop MC. These patients may benefit from an early diagnosis of MC to obtain better results with current treatments and to improve quality of life.

#### Disclosure

All authors have declared no conflicts of interest.

## 403QUALITY OF LIFE AND SYMPTOMS IN BREAST CANCER PATIENTS UNDERGOING CONVENTIONAL CHEMOTHERAPY

### Abstract

Quality of life (QoL) and symptom assessment are of increasing importance to evaluate treatment outcomes in breast cancer patients. We aimed to study feasibility of using standardized QoL and symptom assessment tools to determine benefits and risks of conventional chemotherapy (CT) in breast cancer patients. One hundred and seven breast cancer patients (Stages I-IV) were included in the study (58% patients with metastatic breast cancer). Mean age/SD – 53/10 y.o. All the patients underwent taxane based CT with the previous treatment including CT (89%), surgery (69%), radiotherapy (39%), hormone therapy (24%) or biotherapy (9%). QoL was assessed using generic QoL questionnaires, the SF-36 and the EQ-5D; symptoms – by using Comprehensive Symptom Profile in Patients with Breast Cancer (CSP-Br). The CSP-Br is a self-reported tool which allows to assess the severity of 57 symptoms specific for breast cancer patients. Distribution of patients according to the grades of QoL impairment was analyzed using the SF-36. For comparison of means at different time-points Wilcoxon's matched pairs test was used. The patients reported the usefulness of PROs tools to facilitate patient-physician communication. The following distribution of patients receiving taxane based CT according to the grades of QoL impairment was observed: 23% of patients had no QoL impairment; 15% patients - mild QoL impairment, 33% - moderate or severe QoL impairment, and 19% - critical QoL impairment. Health index measured by the EQ-5D decreased during the CT cycle and improved by the next CT cycle. The most prevalent and disturbing symptoms were the following: hair loss (>90%), fatigue, feeling of constant tiredness (>80%) and psychological symptoms (>70%). Definite deterioration of symptom profile and severity was observed in a week after CT cycle completion: the severity of 21 out of 47 symptoms increased significantly (p < 0.05). The usefulness of patient-reported outcome measures to distinguish patients in terms of QoL impairment as well as in terms of severity and number of symptoms experienced was demonstrated. The SF-36, EQ-5D and CSP-Br are robust and feasible tools to measure benefits and risks of breast cancer treatment from patient perspective.

#### Disclosure

All authors have declared no conflicts of interest.

## 404INTRATHECAL TRASTUZUMAB IN THE TREATMENT OF LEPTOMENINGEAL METASTASES FROM HER2-POSITIVE CANCER

### Abstract

#### Introduction

Up to 8% of all cancer patients develop leptomeningeal metastases (LM). Median overall survival after diagnosis is approximately 1 month. Trastuzumab, a monoclonal antibody against the HER2 receptor, is used in the treatment of HER2 positive cancer patients. It is still not clear if systemic trastuzumab can penetrate the intact blood brain barrier, because of its high molecular weight. Given intrathecally, trastuzumab could achieve higher concentrations in the cerebrospinal fluid (CSF).

#### Purpose

Determine safety, response and overall survival of patients treated with intrathecal (IT) trastuzumab, in a single centre population.

#### Patients and methods

Clinical data of patients treated with IT trastuzumab have been reviewed. Survival was defined as time since beginning of IT trastuzumab until death or last follow-up.

#### Results

A total of 4 patients have been treated with IT trastuzumab (Table 1). Table 1. Patient characteristics, previous treatment and survival. .

Table: 404

Patient Sex Age Primary tumor Histology Previous treatment Survival (months)
41 breast IDC, HR- CT, S, RT, sT 1,6 +
44 breast IDC, HR+ CT, RT, HT, S, sT 23,7
31 breast IDC, HR+ S, RT, HT, sT 2,4
Patient Sex Age Primary tumor Histology Previous treatment Survival (months)
41 breast IDC, HR- CT, S, RT, sT 1,6 +
44 breast IDC, HR+ CT, RT, HT, S, sT 23,7
31 breast IDC, HR+ S, RT, HT, sT 2,4

All HER2 positive, IDC invasive ductal carcinoma, ADC adenocarcinoma, S surgery, RT radiotherapy, HT hormonal treatment, sT systemic trastuzumab. LM were confirmed by lumbar puncture in patients B, C and D. In patient A LM were diagnosed by MRI (CSF cytology persistently negative). Patients A, B, and C received weekly IT trastuzumab 25 mg. Patient D received weekly IT (trastuzumab 25 mg + methotrexate 12 mg). Toxicity related to IT trastuzumab was not observed. The CSF cytology remained positive in patient D and became negative after 1 and 3 weeks for patient C and B, respectively. Patient A was still receiving treatment at last follow-up.

#### Conclusions

In our group of patients, IT trastuzumab was well tolerated and had encouraging results. Ours is a small and somewhat heterogeneous population and we can not extrapolate on the efficacy of trastuzumab in this setting, but a study with a larger number of patients is warranted and may help identify patients appropriate for this therapy.

#### Disclosure

All authors have declared no conflicts of interest.

## 405TiPA PHASE II STUDY OF ALBUMIN-BOUND PACLITAXEL COMBINED WITH EPIRUBICIN AND CYCLOPHOSPHAMIDE AS NEO-ADJUVANT THERAPY IN BREAST CANCER WOMEN

### Abstract

#### Purpose

To observe whether the efficacy and safety of Albumin-Bound Paclitaxel has advantage over Cremophor-Formulated Paclitaxel as neo-adjuvant therapy in breast cancer

#### Patients and methods

Twenty-six breast cancer patients were enrolled into Albumin-Bound Paclitaxel group. Albumin-Bound Paclitaxel 260mg/m2, Epirubicin 60 mg/m2, Cyclophosphamide 500mg/m2 was administrated as neo-adjuvant therapy. Twenty-six patients who were treated with Cremophor-Formulated Paclitaxel 175mg/m2, Epirubicin 60 mg/m2, Cyclophosphamide 500mg/m2 as the control group. The objective observation include the pathologic complete response rates, clinical response rates, safety and toxicity. The PI3K, pAKT, mTOR, BAD protein were examined before and after chemotherapy in two groups.

#### Results

34.6% patients (9/26) in Albumin-Bound Paclitaxel group achieved a clinical complete response, this rate was 11.5% higher than control group which rate was 23.1%(6/26). 15.4% patients (4/26) in Albumin-Bound Paclitaxel group achieved a pathologic complete response, this rate was 11.6% higher than control group which rate was 3.8%(1/26). None patients (0/26) in Albumin-Bound Paclitaxel group occurs neutropenia but 15.4% patients (4/26) in control group occurs neutropenia during chemotherapy. All patients have normal left ventricular ejection fraction range (LVEF > 50%) before and after therapy. The positive rate of PI3K, mTOR, pAKT expression decreased 45.6%,42.5%,41.1% respectively in Albumin-Bound Paclitaxel group after chemotherapy and the control group decreased 13.2%,14.1%,10.1%. The positive rate of BAD expression increased 21.6% in Albumin-Bound Paclitaxel group after chemotherapy and the control group only increased 6.2%.

#### Conclusion

Albumin-Bound Paclitaxel has advantage in clinical response rate over Cremophor-Formulated Paclitaxel in breast cancer. Albumin-Bound Paclitaxel has no more additional adverse events than Cremophor-Formulated Paclitaxel. Albumin-Bound Paclitaxel has lower incidence of neutropenia than Cremophor-Formulated Paclitaxel. Albumin-Bound Paclitaxel decrease the expression of PI3K, pAKT, mTOR protein and increase the expression of BAD protein more significantly than Cremophor-Formulated Paclitaxel after chemotherapy

#### Disclosure

All authors have declared no conflicts of interest.

## 406TiPPHASE II TRIAL OF PRIMARY SYSTEMIC THERAPY (PST) WITH DOCETAXEL (D) FOLLOWED BY HIGH-DOSE EPIRUBICIN IN COMBINATION WITH CYCLOPHOSPHAMIDE (EC) PLUS CONCURRENT TRASTUZUMAB (T) FOR STAGE II-III HER-2 POSITIVE BREAST CANCER PATIENTS (PTS)

### Abstract

#### Background

PST with taxanes, anthracycline-containing regimens and T showed a high percentage of pathologic complete response (pCR); T administered concurrently with chemotherapy is more effective, but the combination with anthracyclines may be at risk of cardiotoxicity. The present study evaluates efficacy and toxicity of T administered concurrently with a sequential regimen of D followed by EC in neoadjuvant setting.

#### Materials and methods

This phase II single stage trial is enrolling pts with cT2-T4, N0-2, M0, Her-2 positive (IHC 3+ or 2+ amplified) breast cancer. A core biopsy is required prior treatment start to evaluate hormonal receptors, Ki67, topoisomerase II, Her-2, with re-evaluation of these parameters, whenever possible, at definite surgery. LVEF is evaluated before and during treatment. Blood samples are collected at baseline for evaluation of 9 genetic polymorphisms related to higher risk of developing cardiac toxicity. We are quantitatively evaluating gene copy number by multiple ligand probe amplification (MLPA), PTEN, p-Akt, p-MAPK, and PIK3CA mutations. Pts receive 4 cycles of D (100mg/m2) plus T (loading dose 8mg/kg followed by 6mg/kg), followed by 4 cycles of EC (120/600mg/m2) plus T, every 3 weeks. Definite surgery is planned at the end of PST, and standard radiotherapy and hormonal adjuvant treatment in case of positive hormonal receptors are given; adjuvant T is given for 6 months. The primary objective of the trial is pCR (absence of invasive tumor cells in the breast and axilla); secondary objectives are cardiac safety, toxicity, disease-free survival. The planned sample size is 42 pts.

#### Results

To date, 29 pts have been enrolled; median age is 45 years, pre/postmenopausal 22/7, ER and/or PgR positive in 16 pts. 27 pts are evaluable for pathological response: we observed 20 pCR (74%, 95%C.I, 57.5-90.6). Grade 3-4 neutropenia was the most frequent toxicity, encountered in 75% of the pts, other toxicities were mild to moderate. No clinical cardiotoxicity was observed.

#### Conclusions

At preliminary analysis, PST with sequential administration of D followed by EC, given concurrently with T, appears to be very active and safe in stage II-III breast cancer pts.

#### Disclosure

All authors have declared no conflicts of interest.

## 407TiPTREATMENT OF PATIENTS WITH HORMONE RECEPTOR (HR)-POSITIVE AND HER2-POSITIVE LOCALLY ADVANCED OR METASTATIC BC WITH A COMBINATION OF PERTUZUMAB AND TRASTUZUMAB PLUS AN AROMATASE INHIBITOR (AI): AN OPEN-LABEL RANDOMISED PHASE II STUDY (PERTAIN)

### Abstract

#### Background

Development of resistance to endocrine therapy in patients (pts) with breast cancer (BC) is a major clinical concern. Preclinical studies suggest that blockade of HR with HER2 inhibition may be key to overcoming resistance and improving clinical outcomes. The combination of pertuzumab (P) and trastuzumab (H) with docetaxel significantly improves outcomes by a comprehensive blockade of HER2 signalling. However, initial results suggest that this benefit may be less apparent in HR-positive pts. PERTAIN is the first study to assess whether a fuller blockade of HER signalling with P and H in conjunction with an AI may resensitise HER2-positive tumours to endocrine therapy and provide an effective first-line treatment option in pts with HR-positive and HER2-positive advanced BC.

#### Methods

PERTAIN is a multicentre, open-label, Phase II trial for postmenopausal women with HER2-positive and HR-positive locally advanced/metastatic BC to assess the efficacy of P plus H with an AI as first-line therapy. Pts will be randomised 1:1 to Arm 1 (P: 840 mg initial dose, 420 mg q3w iv; H: 8 mg/kg initial dose, 6 mg/kg q3w iv; AI [anastrozole 1 mg or letrozole 2.5 mg qd po]) or Arm 2 (H + AI at same dose as Arm 1). Induction chemotherapy (CT) (docetaxel or paclitaxel) may be given to pts for up to 18 weeks at the investigator's discretion. Administration of study medications will continue until disease progression or unacceptable toxicity. Pts must not have been treated with anti-HER2 agents, except H and/or lapatinib in the (neo)adjuvant setting, and must have no CNS involvement or clinically significant cardiovascular disease. The primary endpoint is progression free survival (PFS); secondary endpoints include overall survival, overall response rate, clinical benefit rate, duration of response, time to response, safety and tolerability, and QoL. The study opened in April 2012 and will recruit 250 pts. The main analysis for the primary endpoint (after 165 PFS events) will include a treatment group variable and be stratified by whether pts were chosen to receive induction CT (Yes/No) and time since adjuvant hormone therapy (<12 months, ≥12 months, and no prior hormone therapy).

#### Disclosure

G. Arpino: I declare that I have participated in an advisory board for GSK.

C. Poole: I declare that have served on a advisory board and received research finding from F. Hoffmann-La Roche.

L. Mitchell: I am currently employed by F. Hoffmann-La Roche.

C. Pelizon: I am currently employed by F. Hoffmann-La Roche.

M.F. Rimawi: I declare that have received research funding from Genentech Inc and GSK.

All other authors have declared no conflicts of interest.

## 408TiPPERTUZUMAB AND TRASTUZUMAB IN COMBINATION WITH VINORELBINE FOR FIRST-LINE TREATMENT OF PATIENTS WITH HER2-POSITIVE LOCALLY ADVANCED OR METASTATIC BREAST CANCER (LABC/MBC): A SINGLE-ARM, TWO-COHORT, PHASE II STUDY (VELVET)

### Abstract

#### Background

The humanised monoclonal antibody pertuzumab binds to the dimerisation domain of HER2, preventing heterodimerisation and downstream signalling. As pertuzumab is directed against a different epitope to trastuzumab, a more comprehensive HER2 blockade is achieved by combining the agents. The CLEOPATRA study showed significantly improved efficacy for pertuzumab and trastuzumab plus docetaxel, but pertuzumab with trastuzumab has not been tested with other chemotherapies in MBC. Trastuzumab plus vinorelbine (V) has comparable efficacy to trastuzumab plus docetaxel but with fewer adverse events. VELVET will assess the overall response rate (ORR) of pertuzumab with trastuzumab + V in first-line treatment of HER2-positive MBC. Administration of pertuzumab and trastuzumab in the same infusion bag will also be assessed.

#### Methods

VELVET is a multicentre, open-label, two-cohort, Phase II trial in patients (pts) with HER2-positive LABC/MBC not previously treated in the metastatic setting with non-hormonal anticancer therapy. Pts must have LVEF ≥55% and ECOG PS of 0/1. Enrolment began in April 2012 and 210 pts will be recruited. Cohort 1 (first 105 pts enrolled) will receive pertuzumab and trastuzumab sequentially, and Cohort 2 (next 105 pts) will receive pertuzumab and trastuzumab in the same infusion bag at Cycle 2 onwards assuming Cycle 1 was tolerated. V will be given in both cohorts. Treatment will continue until disease progression/unacceptable toxicity. Study doses (all iv): pertuzumab: 840 mg initial dose, 420 mg q3w; trastuzumab: 8 mg/kg initial dose, 6 mg/kg q3w, and V: 25 mg/m2 Day 1 and 8 (first cycle) then 30 − 35 mg/m2 Days 1 and 8 q3w (dose escalation at investigator discretion). Assuming best overall response of 70–80% and aiming at a distance from the estimated proportion to the CI limits of 8–11%, a total of 95 pts need to be evaluable per cohort (assuming withdrawal rate ∼10%). Primary endpoint is ORR by independent assessment. Secondary endpoints include investigator ORR assessment, time to response, duration of response, PFS, TTP, OS, safety/ tolerability and QoL.

#### Disclosure

T. Petit: I declare that have served on a advisory board for F. Hoffmann-La Roche.

L. Mitchell: I am currently an employee of F. Hoffmann-La Roche.

C. Pelizon: I am currently an employee of F. Hoffmann-La Roche.

E.A. Perez: I declare that my institute has received research finding from Genentech, Sanofi Onc. and GSK.

All other authors have declared no conflicts of interest.

## 409TiPPERTUZUMAB IN COMBINATION WITH TRASTUZUMAB AND A TAXANE FOR THE FIRST-LINE TREATMENT OF PATIENTS WITH HER2-POSITIVE ADVANCED BREAST CANCER: A SINGLE ARM PHASE IIIB STUDY (PERUSE)

### Abstract

#### Background

Pertuzumab (P) inhibits downstream signaling of HER2 by preventing its heterodimerization with other HER family members. The epitope recognized by P is distinct from that bound by trastuzumab (H) and their complementary mechanisms of action lead to a more comprehensive HER2 blockade when combined. Data from the phase III trial CLEOPATRA showed significantly improved PFS in pts with HER2-positive 1L MBC given P + H + docetaxel (D). As H was not widely available in the (neo)adjuvant setting prior to CLEOPATRA recruitment, a relatively low proportion of pts in CLEOPATRA had previously received H. PERUSE will assess the safety and tolerability of P + H + one of a choice of taxanes (T) as 1L therapy for pts with HER2-positive metastatic or locally advanced BC. Efficacy endpoints will also be recorded in PERUSE, in a pt population likely to have experienced wider exposure to prior H therapy.

#### Methods

PERUSE is a phase IIIb, multicenter, open-label, single-arm study in pts with HER2-positive BC who have not been treated with systemic non-hormonal anticancer therapy for MBC. The planned sample size is 1500. Pts will receive, P: 840mg initial dose, 420mg q3w IV; H: 8mg/kg initial dose, 6mg/kg q3w IV, T: D, paclitaxel or nab-paclitaxel according to local guidelines. A planned protocol amendment will allow HR-positive pts to receive endocrine therapy in conjunction with P + H following completion of T in line with clinical practice. Treatment will be administered until disease progression or unacceptable toxicity. At baseline, pts must have an LVEF of ≥50%, ECOG PS of 0, 1 or 2 and must not have received prior anti-HER2 agents for MBC. Prior H and/or lapatinib in the (neo)adjuvant setting is allowed, provided there was no disease progression on-treatment. A disease-free interval of ≥6 months is required. The primary endpoint is safety and tolerability. Secondary endpoints include PFS, OS, ORR, CBR, duration of response, time to response and QoL. The final analysis will be carried out when pts have been followed up for ≥12 months. Interim analyses are planned after enrollment of ∼350, 700 and 1000 pts. Regular interim safety assessments by a DSMB will take place.

#### Disclosure

D. Miles: I have an interest in relation with one organisation that could be perceived as a possible conflict of interest in the context of the subject of this abstract. I have served on Advisory Board Meetings for Roche/Genentech.

F. Puglisi: I have an interest in relation with one organisation that could be perceived as a possible conflict of interest in the context of the subject of this abstract. I participate in Advisory Board Meetings for Roche.

A. Schneeweiss: I have an interest in relation with one organisation that could be perceived as a possible conflict of interest in the context of the subject of this abstract. I serve on Roche Advisory Boards and am in volved with corporate-sponsored research with Roche.

L. Mitchell: I am currently an employee of F. Hoffmann-La Roche.

A. Dünne: I have an interest in relation with one organisation that could be perceived as a possible conflict of interest in the context of the subject of this abstract. I am an employee of hoffmann-La Roche.

T. Bachelot: I have an interest in relation with one organisation that could be perceived as a possible conflict of interest in the context of the subject of this abstract. I am involved with Roche sponsoredresearch and Roche advisory Board Meetings.

All other authors have declared no conflicts of interest.