TP53-associated early breast cancer: new observations from a large cohort

Abstract

Background

A recent large, well-annotated international cohort of patients with Li-Fraumeni syndrome and early-stage breast cancer was examined for shared features.

Methods

This multicenter cohort study included women with a germline TP53 pathogenic or likely pathogenic variant and nonmetastatic breast cancer diagnosed between 2002 and 2022. Clinical and genetic data were obtained from institutional registries and clinical charts. Descriptive statistics were used to summarize proportions, and differences were assessed using χ² or Wilcoxon rank sum tests. Metachronous contralateral breast cancer risk, radiation-induced sarcoma risk, and recurrence-free survival were analyzed using the Kaplan-Meier methodology.

Results

Among 227 women who met study criteria, the median age of first breast cancer diagnosis was 37 years (range = 21-71), 11.9% presented with bilateral synchronous breast cancer, and 18.1% had ductal carcinoma in situ only. In total, 166 (73.1%) patients underwent mastectomies, including 67 bilateral mastectomies as first breast cancer surgery. Among those patients with retained breast tissue, the contralateral breast cancer rate was 25.3% at 5 years. Among 186 invasive tumors, 72.1% were stages I to II, 48.9% were node negative, and the most common subtypes were hormone receptor-positive/HER2-negative (40.9%) and hormone receptor positive/HER2 positive (34.4%). At a median follow-up of 69.9 months (interquartile range = 32.6-125.9), invasive hormone receptor–positive/HER2-negative disease had the highest recurrence risk among the subtypes (5-year recurrence-free survival = 61.1%, P = .001). Among those who received radiation therapy (n = 79), the 5-year radiation-induced sarcoma rate was 4.8%.

Conclusion

We observed high rates of ductal carcinoma in situ, hormone receptor–positive, and HER2-positive breast cancers, with a worse outcome in the hormone receptor–positive/HER2-negative luminal tumors, despite appropriate treatment. Confirmation of these findings in further studies could have implications for breast cancer care in those with Li-Fraumeni syndrome.

Li-Fraumeni syndrome is an autosomal-dominant cancer predisposition syndrome first described in 1969 (1). A high burden of diverse cancers at younger ages forms the main characteristic of classic families (2). TP53 germline pathogenic or likely pathogenic variants underlie Li-Fraumeni syndrome; germline genetic testing is the gold standard for a Li-Fraumeni syndrome diagnosis (3).

Cancer surveillance in Li-Fraumeni syndrome can be challenging because of the wide spectrum of tumors and the variable cancer risks according to age, sex, and TP53 mutation type (4-6). Cancer treatment confers additional cancer risks on carriers, including sarcomas in the radiation field (7,8) and therapy-related hematologic malignancies (9).

Breast cancer is the most common malignancy in women with Li-Fraumeni syndrome. Lifetime risk of breast cancer has been estimated at 85% by age 60 years (4). The median age of first breast cancer diagnosis varies from 32 to 40 years, depending on the TP53 variant location and its functional impact on p53 (10,11). The increased risks of radiation-induced sarcomas and second primary breast cancers may inform local therapy decisions for early-stage disease. Mastectomy instead of breast-conserving surgery may allow patients with Li-Fraumeni syndrome and breast cancer to avoid adjuvant radiation therapy (RT) (8,12-14). Multiple studies have shown that breast tumors are most often hormone receptor positive, and a high proportion show HER2 overexpression or amplification, which should influence systemic treatment (11,15-20).

Given the rarity of Li-Fraumeni syndrome, most studies of breast cancer in Li-Fraumeni syndrome have been small cohorts (11,16-19). In this study, we assembled a larger cohort of individuals with Li-Fraumeni syndrome–associated breast cancer from international institutional registries to improve the description of clinical features and early outcome data for women with Li-Fraumeni syndrome diagnosed with early breast cancer.

Methods

Women (age ≥18 years) diagnosed with a first breast cancer, in situ or nonmetastatic invasive disease, between 2002 and 2022 were identified from international Li-Fraumeni syndrome registries: in the United States at the Dana-Farber Cancer Institute (Boston); in Brazil at the Hospital Sírio-Libanês (Sao Paulo and Brasilia); and in France at the Institut Gustave Roussy (Paris), Montpellier Cancer Institute (Montpellier), and the Institut Paoli-Calmettes (Marseille). Institutional review boards from each academic center approved the study and provided a waiver of consent for the deidentified data.

All data, including type and year of genetic testing, and test results were collected directly from genetic test reports and institutional databases. TP53 variants were considered pathogenic or likely pathogenic according to at least 1 of the following criteria: 1) ClinVar classification (21), 2) TP53-ACMG/AMP specifications from the ClinGen TP53 Variant Curation Expert Panel (22), and 3) test report by 1 or more major laboratories.

Patients with confirmed constitutional mosaic TP53 variants were included. We excluded TP53 carriers with any metastatic disease (breast cancer or non–breast cancer) at first breast cancer diagnosis, carriers of mosaic TP53 variants considered likely acquired aberrant clonal expansions, and carriers of a second pathogenic variant in another breast cancer susceptibility gene.

Patient demographics, breast tumor characteristics, clinical treatment, and additional cancer diagnoses were reviewed. Cancer family history (first-, second-, and third-degree relatives) was assessed through pedigrees collected in genetics clinic or through patients’ self-reported questionnaires. Each person or family was classified according to the recently proposed Li-Fraumeni syndrome spectrum classification (23).

Breast cancer diagnoses were confirmed using hospital and pathology records. Date of first breast cancer diagnosis was established from diagnostic biopsy date. Pathologic tumor stages were evaluated according to the American Joint Committee on Cancer Anatomic Stage Group. For patients who received neoadjuvant therapy, initial clinical staging was used; otherwise, pathology data from the final surgical specimen was collected. Estrogen receptor, progesterone receptor, and HER2 status by immunohistochemistry or in situ hybridization was performed as a routine diagnostic by local pathology departments. Estrogen receptor–positive or progesterone receptor–positive tumors were defined as 10% or greater (hormone receptor-positive). HER2-positive tumors were defined using American Society of Clinical Oncology–College of American Pathologists guidelines (24). When HER2 immunohistochemical testing was 2+, in situ hybridization testing was required to determine HER2 status.

Synchronous bilateral breast cancer was defined as breast cancer diagnosed within 6 months. When breast cancer was bilateral at presentation, the tumor with the higher risk of recurrence, as determined by invasive status, stage, or tumor biology, was designated as “first” for subsequent analyses.

Radiation-induced sarcomas were defined as those developing in the field of adjuvant breast therapeutic radiation. First breast cancer recurrences were locoregional (breast, chest wall, axilla, ipsilateral, or supraclavicular nodes), distant, or both. Recurrence-free survival (RFS) was defined as the time of first breast cancer diagnosis to development of any recurrence (locoregional, distant, or both) or death. Patients without an event were censored at the time of last follow-up.

All collected data were summarized using count and percentage or median and range, when appropriate. The difference of clinical characteristics between ductal carcinoma in situ (DCIS) and invasive tumor types were assessed using χ² test or Wilcoxon signed-rank test, when appropriate. Five-year metachronous contralateral breast cancer risk, radiation-induced sarcoma risk, and RFS by breast cancer subtype were estimated using the Kaplan-Meier method. The survival rates among breast cancer subtypes were compared using the log-rank test.

Results

Cohort characterization

Among 288 women identified with breast cancer and a TP53 germline pathogenic variant, 227 from 217 pedigrees met criteria for this analysis (Supplementary Table 1, available online). Cohort characteristics are described in Table 1. Only 18 patients (8%) were known to have a TP53 pathogenic variant before their first breast cancer diagnosis. TP53 germline status was available within 1 year after their breast cancer diagnosis for most patients (n = 147 [64.8%]), and most were identified with multigene panel testing (n = 142 [62.6%]) followed by TP53 single-gene analysis (n = 48) and family variant testing (n = 30). Patients with DCIS only were more likely to have had testing only for their known family variant than those with invasive disease (31.7% vs 9.1%, P = .002).

Table 1.

Cohort characteristics

Variable	Overall cohort, No. (%), N = 227	Ductal carcinoma in situ only, No. (%), n = 41	Invasive breast cancer, No. (%), n = 186	P
Timing of genetic testing				.6
TP53 genetic testing within 1 y of breast cancer diagnosisa	147 (64.8)	25 (61.0)	122 (65.6)
Type of genetic testing				.002
Multigene panel	142 (62.6)	21 (51.2)	121 (65.1)
Single gene	48 (21.1)	6 (14.6)	42 (22.6)
Family variant	30 (13.2)	13 (31.7)	17 (9.1)
Unknown	7 (3.1)	1 (2.4)	6 (3.2)
Type of TP53 variant				.07
Missense, DNA-binding domain	104 (45.8)	18 (43.9)	86 (46.2)
Missense, Tetramerization domain	64 (28.2)	11 (26.8)	53 (28.5)
Nonsense	16 (7.0)	2 (4.9)	14 (7.5)
Frameshift	13 (5.7)	2 (4.9)	11 (5.9)
In-frame	3 (1.3)	1 (2.4)	2 (1.1)
Intronic	18 (7.9)	2 (4.9)	16 (8.6)
Copy number variation	9 (4.0)	5 (12.2)	4 (2.2)
Constitutional mosaicism for the TP53 variant	13 (5.7)	0 (0)	13 (7.0)
Li-Fraumeni syndrome clinical criteria				.5
Classic or revised Chompret	142 (62.6)	29 (70.7)	113 (60.8)
Attenuated Li-Fraumeni syndrome	84 (37)	12 (29.3)	72 (38.7)
Other criteriab	1 (0.4)	0 (0)	1 (0.5)
Close relativesc affected by breast cancer
Any age	154 (67.8)	30 (73.2)	124 (66.7)	.4
Age <50 y	108 (47.6)	24 (58.5)	84 (45.2)	.1
Additional cancers (non–breast cancer)
Any	65 (28.6)	15 (36.6)	50 (26.9)	.2
Concomitant or before first breast cancer	28 (12.3)	8 (19.5)	20 (10.8)	.4
After first breast cancer	37 (16.3)	7 (17.1)	30 (16.1)

a

18 patients were tested in the previous year or earlier from their first breast cancer diagnosis.

b

TP53 pathogenic variant carrier with a personal history of cancer before age 18 years who does not meet classic or revised Chompret criteria.

c

Close relatives were defined as first-, second-, and third-degree relatives.

In this cohort, 101 different TP53 variations were noted (Supplementary Table 2, available online). Missense variants were the most common type (n = 168 [74.0%]), of which 104 (62.0%) were in the DNA-binding domain and 64 (38.0%) affected the tetramerization domain. Thirteen patients (5.7%) were considered constitutional mosaics for the TP53 variant (Table 1).

Classic or revised Li-Fraumeni syndrome Chompret criteria were met by 142 (62.6%) patients, whereas 84 (37.0%) patients met attenuated Li-Fraumeni syndrome criteria (Table 1). There were 154 patients (67.8%) who had at least 1 first-, second-, or third-degree relative affected by breast cancer, and 108 patients (47.6%) had at least 1 relative diagnosed with breast cancer before age 50 years.

Year of first breast cancer diagnosis, clinical and pathological breast cancer features, and treatment regimens are described in Table 2 and Supplementary Table 3 (available online). Most patients were premenopausal at their first breast cancer diagnosis (n = 190 [83.7%]), the median age at diagnosis was 37 years (range = 21-71), and almost one-quarter (n = 55) were diagnosed at age 30 years or younger. Bilateral synchronous disease was present among 11.9% of women (n = 27) at their first breast cancer diagnosis. Fourteen patients (6.2%) received a breast cancer diagnosis while pregnant or breastfeeding. Eighteen patients (8.0%) had genetic testing before their first breast cancer diagnosis. Among patients with information about clinical presentation (n = 151), most presented with breast symptoms (n = 97 [64.2%]), and 50 (33.1%) had a screen-detected tumor (15 with DCIS only and 35 with invasive breast cancer). Among screen-detected tumors, most were identified by mammography (n = 32), 10 by breast magnetic resonance imaging, and 5 by breast ultrasound alone. Three patients were incidentally diagnosed with breast cancer during risk-reducing mastectomy, and 1 patient was diagnosed during reduction mammoplasty.

Table 2.

Breast cancer features and treatment

Variable	Overall cohort, N = 227	DCIS only, n = 41	Invasive breast cancer, n = 186
Year of first breast cancer diagnosis, No. (%)
2002-2006	32 (14.1)	4 (9.8)	28 (15.1)
2007-2011	37 (16.3)	7 (17.1)	30 (16.1)
2012-2016	66 (29.1)	18 (43.9)	48 (25.8)
2017-2022	92 (40.5)	12 (29.3)	80 (43.0)
Age at first breast cancer diagnosis, median (range), y	37 (21-71)	41 (26-71)	35 (21-70)
Breast cancer diagnosis, by age group, No. (%)
≤30 y	55 (24.2)	3 (7.3)	52 (28.0)
31-45 y	122 (53.7)	25 (61.0)	97 (52.2)
>45 y	50 (22.0)	13 (31.7)	37 (19.9)
Breast cancer presentation, No. (%)
Symptomatic	97 (42.7)	2 (4.9)	95 (51.1)
Asymptomatic, breast screening	50 (22.0)	15 (36.6)	35 (18.8)
Incidental findinga	4 (1.8)	4 (9.8)	0 (0)
Unknown	76 (33.5)	20 (48.8)	56 (30.1)
Bilateral synchronous disease	27 (11.9)	1 (2.4)	26 (14.0)
Breast cancer diagnosis while pregnant or breastfeeding, No. (%)	14 (6.2)	1 (2.4)	13 (7.0)
Stage, No. (%)
In situ disease	41 (18.1)	41 (100)	—
I	68 (30.0)	—	68 (36.6)
II	66 (29.0)	—	66 (35.5)
III	41 (18.1)	—	41 (22.0)
Non-metastatic, unspecified	11 (4.8)	—	11 (5.9)
Histology, No. (%)
DCIS	41 (18.1)	41 (100)	—
Invasive ductal carcinoma	162 (71.3)	—	162 (87.1)
Invasive lobular carcinoma	1 (0.5)	—	1 (0.5)
Mixed (invasive ductal carcinoma and invasive lobular carcinoma)	4 (1.8)	—	4 (2.2)
Other invasive typesb	6 (2.6)	—	6 (3.2)
Invasive unspecified	13 (5.7)	—	13 (7.0)
Hormone receptor and HER2 status, No. (%)
DCIS, hormone receptor positive	20 (8.8)	20 (48.8)	—
DCIS, hormone receptor negative	4 (1.8)	4 (9.7)	—
DCIS, hormone receptor status unavailable	17 (7.5)	17 (41.5)	—
Invasive hormone receptor-positive/HER2-negative	76 (33.5)	—	76 (40.9)
Invasive hormone receptor-positive/HER2-positive	64 (28.2)	—	64 (34.4)
Invasive hormone receptor-negative/HER2-positive	23 (10.1)	—	23 (12.4)
Invasive hormone receptor-negative/HER2-negative	12 (5.3)	—	12 (6.5)
Invasive subtype unspecified	11 (4.8)	—	11 (5.9)
Type of surgery, No. (%)
Breast-conserving surgery	52 (22.9)	10 (24.4)	42 (22.6)
Mastectomy	166 (73.1)	30 (73.2)	136 (73.1)
Unknown	9 (4.0)	1 (2.4)	8 (4.3)
Adjuvant radiation therapy, No. (%)
Yes, after breast-conserving surgery	40 (17.6)	6 (14.6)	34 (18.3)
Yes, after mastectomy	39 (17.2)	1 (2.4)	38 (20.4)
No	130 (57.2)	33 (80.5)	97 (52.2)
Unknown	18 (8.0)	1 (2.4)	17 (9.1)
Adjuvant endocrine therapy, No. (%)
Yes	131 (57.7)	4 (9.7)	127 (68.3)
No	78 (34.3)	35 (85.4)	43 (23.1)
Unknown	18 (8.0)	2 (4.9)	16 (8.6)
Follow-up, median (IQR), mo	69.9 (32.6-125.9)	69.9 (37.6-108.8)	72.7 (32.3-133.1)
5-year recurrence-free survival, % (95% confidence interval)	81.4 (74.7 to 86.4)	97.1 (80.9 to 99.6)	77.9 (70.1 to 83.9)

a

Incidental finding during breast surgery (risk-reduction mastectomy or reduction mammoplasty). DCIS = ductal carcinoma in situ.

b

Other invasive types include mucinous carcinoma, medullary carcinoma, papillary carcinoma, tubular carcinoma, and undifferentiated carcinoma.

Surgical treatment included mastectomies in 166 patients (73.1%), of which 69 were unilateral and 97 bilateral. Most patients with invasive disease (146/186 [78.5%]) received neoadjuvant or adjuvant chemotherapy. There were 79 patients (34.8%) who received adjuvant RT: 40 after breast-conserving surgery and 39 after mastectomy. Among 200 patients with unilateral disease (in situ or invasive) as first breast cancer presentation, 112 (56.0%) underwent contralateral risk-reducing mastectomy, including 67 with their first breast cancer surgery. Patients with genetic testing within 1 year of breast cancer diagnosis had a higher uptake of contralateral risk-reducing mastectomy (67.7% vs 36.8%, P < .0001) (Table 3).

Table 3.

Contralateral risk-reducing mastectomy, by timing of genetic testing

Variable	Total (%)	Genetic testing ≤1 y before breast cancer diagnosis	Genetic testing >1 y after breast cancer diagnosis	P
Unilateral breast cancer at first presentation	200 (100)	124 (100)	76 (100)	—
Contralateral risk-reducing mastectomy uptakea
Any time	112 (56)	84 (67.7)	28 (36.8)	<.0001
Concomitant with first breast cancer surgery	67 (33.5)	56 (45.2)	11 (14.5)	.01

a

One patient was excluded from this analysis because of missing data related to date of contralateral risk-reducing mastectomy.

In situ disease

Forty-one tumors (41/227 [18.1%]) were DCIS only, including 20 hormone receptor-positive (48.8%) and 4 hormone receptor-negative (9.7%) (Table 2). Most DCIS was intermediate (10 [24.4%]) or high nuclear grade (17 [41.5%]) (Supplementary Table 3, available online). Median age at DCIS diagnosis was 41 years (range = 26-71 years). Among the patients with DCIS, 10 (24.4%) had breast-conserving surgery and 7 (17.1%) received RT (Table 2). Thirty patients (73.2%) underwent mastectomy: unilateral (n = 8), bilateral mastectomy for synchronous DCIS (n = 2), and unilateral mastectomy plus simultaneous contralateral risk-reducing mastectomy (n = 20).

Invasive disease

In our cohort, most patients had an invasive breast cancer diagnosis (186/227 [81.9%]) at a median age of 35 years (range = 21-70 years) (Table 2). Invasive ductal carcinoma was the most common histology (162/186 [87.1%]). By subtype, hormone receptor-positive/HER2-negative (n = 76 [40.9%]) was the most frequent, followed by hormone receptor-positive/HER2-positive (n = 64 [34.4%]), hormone receptor-negative/HER2-positive (n = 23 [12.4%]), and hormone receptor-negative/HER2-negative (n = 12 [6.5%]). Most disease was stage I (68 [36.6%]), followed by stage II (66 [35.5%]), and stage III (41 [22.0%]). Tumor size was 2 cm or smaller in 88 cases (47.3%), and 91 (48.9%) were node negative (Supplementary Table 3, available online).

Regarding surgical treatment, 136 (73.1%) patients underwent mastectomy, 42 (22.6%) underwent breast-conserving surgery, and data were unavailable for 8 patients with invasive breast cancer. RT was administered to 34 of the 42 (81%) patients who had breast-conserving surgery. Among the 136 patients whose first surgery was mastectomy, 61 underwent unilateral mastectomy, 25 underwent bilateral mastectomy for bilateral synchronous breast cancer, and 50 patients elected unilateral mastectomy plus contralateral risk-reducing mastectomy. Postmastectomy radiation was administered to 38 (27.9%) of the 136 mastectomy patients; among these, 33 (86.8%) had node-positive disease. Among the 92 patients who underwent mastectomy and did not receive RT, 34 (36.9%) had nodal involvement.

Among the 186 patients with invasive disease, 78.5% received neoadjuvant or adjuvant chemotherapy (Table 2). A majority (n = 120 [82.8%]) of the 145 patients with invasive hormone receptor–positive disease received adjuvant endocrine therapy (Supplementary Table 4, available online). Endocrine therapy included tamoxifen alone for 65 patients (54.2%), tamoxifen plus ovarian suppression or ablation for 15 (12.5%) patients, tamoxifen followed by aromatase inhibitor therapy for 5 (4.2%) patients, aromatase inhibitor therapy alone in 14 (11.7%) patients, and aromatase inhibitor therapy plus ovarian suppression or ablation in 12 (10.0%) patients; 9 (7.5%) patients received endocrine therapy, for which the regimen was unavailable.

Outcomes

Survival analyses by Kaplan-Meier testing according to clinical outcomes are illustrated in Figure 1. The median follow-up of the whole cohort was 69.9 months. A total of 35 patients developed metachronous contralateral breast cancer, including 5 patients who had undergone risk-reducing mastectomies. Among patients with unilateral invasive disease, 5-year metachronous contralateral breast cancer risk was 25.3% (95% confidence interval [CI] = 17.5 to 35.6). Among patients with unilateral DCIS only, 5-year metachronous contralateral breast cancer risk was 17.5% (95% CI = 5.5 to 48.3).

Figure 1.

Survival analysis by Kaplan-Meier analysis according to clinical outcomes. A) Metachronous contralateral breast cancer–free survival. B) Radiation-induced sarcoma–free survival. C) Recurrence-free survival by breast cancer subtype.

^a One patient was excluded from the metachronous contralateral breast cancer–free survival analysis because of missing information regarding date of contralateral risk-reducing mastectomy.

Six (7.6%) of 79 patients whose first breast cancer treatment included RT developed a sarcoma in the radiation field. Five-year risk of radiation-induced sarcoma was 4.8% (95% CI = 1.6 to 14.2) (Figure 1, B). The 5-year RFS was 97.1% for patients with DCIS only and 77.9% for patients with invasive breast cancer (Table 2).

Invasive hormone receptor–positive/HER2-negative disease had the lowest 5-year RFS, at 61.1% (Table 4). This rate was statistically significant compared with the other breast cancer subtypes on Kaplan-Meier analysis (P = .001) (Figure 1, C). The 5-year RFS for invasive disease was most favorable among those with hormone receptor–positive/HER2-positive disease, at 90.3%, followed by hormone receptor–negative/HER2-positive disease (85.3%) and hormone receptor–negative/HER2-negative disease (81.8%). RFS was not significantly different between hormone receptor–positive/HER2-negative disease and hormone receptor–negative/HER2-negative disease (P = .4) or between hormone receptor–positive/HER2-negative disease and hormone receptor–negative/HER2-positive disease (P = .05). A significant difference, however, in RFS was observed between hormone receptor–positive/HER2-negative disease and hormone receptor–positive/HER2-positive disease, even after adjustment of multiplicity (P = .001). Among patients with invasive disease, tumor grade, nodal status, and type of surgery were not associated with RFS (Supplementary Table 5, available online). Eight patients died from other cancer types, and 11 patients died from progressive breast cancer, of whom 6 had hormone receptor–positive/HER2-negative invasive disease (Supplementary Table 6, available online).

Table 4.

Staging, treatment regimens, and recurrence-free survival among patients with invasive tumors, by cancer subtype

Variable	All invasive	Hormone receptor positive/HER2 negative	Hormone receptor positive/HER2 positive	Hormone receptor negative/HER2 positive	Hormone receptor negative/HER2 negative
	N = 186	n = 76	n = 64	n = 23	n = 12
Stage, No. (%)
I	68 (36.6)	33 (43.4)	21 (32.8)	7 (30.4)	3 (25)
II	66 (35.5)	22 (28.9)	31 (48.4)	6 (26.1)	4 (33.3)
III	41 (22.0)	16 (21.1)	10 (15.6)	10 (43.5)	4 (33.3)
Nonmetastatic, unspecified	11 (5.9)	5 (6.6)	2 (3.1)	0 (0)	1 (8.3)
Tumor size, No. (%)
T1	88 (47.3)	39 (51.3)	31 (48.4)	9 (39.1)	3 (25.0)
T2	66 (35.5)	24 (31.6)	28 (43.8)	9 (39.1)	4 (33.3)
T3	14 (7.5)	8 (10.5)	1 (1.6)	3 (13.0)	2 (16.7)
T4	5 (2.7)	0 (0)	2 (3.1)	1 (4.3)	2 (16.7)
Unknown	13 (7.0)	5 (6.6)	2 (3.1)	1 (4.3)	1 (8.3)
Nodal status, No. (%)
N0	91 (48.9)	39 (51.3)	31 (48.4)	10 (43.5)	5 (41.7)
N1	56 (30.1)	19 (25.0)	23 (35.9)	6 (26.1)	5 (41.7)
N2	19 (10.2)	10 (13.2)	3 (4.7)	5 (21.7)	0 (0)
N3	9 (4.8)	3 (3.9)	4 (6.3)	1 (4.3)	1 (8.3)
Unknown	11 (5.9)	5 (6.6)	3 (4.7)	1 (4.3)	1 (8.3)
Type of surgery, No. (%)
Breast-conserving surgery	42 (22.6)	20 (26.3)	13 (20.3)	3 (13.0)	2 (16.7)
Mastectomy	136 (73.1)	51 (67.1)	50 (78.1)	19 (82.6)	9 (75)
Unknown	8 (4.3)	5 (6.6)	1 (1.6)	1 (4.3)	1 (8.3)
Radiation therapy, No. (%)
Yes	72 (38.7)	33 (43.4)	24 (37.5)	6 (26.1)	4 (33.3)
No	97 (52.2)	35 (46.1)	36 (56.3)	15 (65.2)	6 (50.0)
Unknown	17 (9.1)	8 (10.5)	4 (6.3)	2 (8.7)	2 (16.7)
Neoadjuvant or adjuvant chemotherapy, No. (%)
Yes	146 (78.5)	49 (64.5)	61 (95.3)	20 (87.0)	11 (91.7)
No	34 (18.3)	22 (28.9)	2 (3.1)	3 (13.0)	1 (8.3)
Unknown	6 (3.2)	5 (6.6)	1 (1.6)	0 (0)	0 (0)
Adjuvant endocrine therapy, No. (%)
Yes	127 (68.3)	62 (81.6)	55 (85.9)	2 (8.7)	3 (25.0)
No	43 (23.1)	7 (9.2)	4 (6.3)	18 (78.3)	8 (66.7)
Unknown	16 (8.6)	7 (9.2)	5 (7.8)	3 (13.0)	1 (8.3)
Follow-up, median (IQR), mo	72.7 (32.3-133.1)	47.1 (21.0-106.6)	81.4 (35.0-148.1)	76.7 (32.0-115.6)	138.9 (72.7-153.3)
5-year recurrence-free survival, % (95% confidence interval)	77.9 (70.1 to 83.9)	61.1 (46.4 to 72.9)	90.3 (78.2 to 95.9)	85.3 (61.0 to 95.1)	81.8 (44.7 to 95.1)

Discussion

The understanding of breast cancer in Li-Fraumeni syndrome has been evolving, but detailed data regarding response to therapy and outcomes have been limited because of the rarity of the condition. Understanding Li-Fraumeni syndrome and breast cancer has been especially challenging because treatment pathways have evolved rapidly in breast oncology with the introduction of targeted therapeutics (25-27). This international collaborative effort allowed for a comprehensive description of a recent large, well-annotated cohort of Li-Fraumeni syndrome patients with early breast cancer.

TP53 is the most frequently somatically mutated gene across all breast cancer subtypes (28), and the presence of mutated TP53 has been associated with worse prognosis (29). Although triple-negative/basal-like breast cancer is the subtype with the highest rate of TP53 somatic mutations, for unclear reasons, carriers of TP53 germline pathogenic variants rarely develop triple-negative breast cancer (11,15-19). Recently, Sheng et al. (30) suggested that TP53 germline pathogenic variants could be an adverse independent factor for breast cancer RFS (hazard ratio = 2.24, P = .02), but the TP53 study cohort was small (n = 50) and did not permit analysis by breast cancer subtype. Further, TP53 variants more recently reclassified as benign or likely benign and variants of uncertain significance per ClinVar annotations were included in this previous work.

In our study, tumor subtype was significantly associated with RFS (P = .001), with a high recurrence rate observed for hormone receptor–positive/HER2-negative tumors (5-year RFS = 61.1%). This finding is notable because hormone receptor–positive/HER2-negative early breast cancer has the best prognosis among subtypes in the general population. Considering that 42.1% of our hormone receptor–positive/HER2-negative cohort had nodal involvement, recent data from the monarchE (31) and related trials (26,27,32,33) open the possibility that TP53-related hormone receptor–positive/HER2-negative locally advanced breast cancer in Li-Fraumeni syndrome could benefit from adjuvant use of CDK4/CDK6 inhibitors. Given the less favorable outcomes for Li-Fraumeni syndrome hormone receptor–positive/HER2-negative tumors, mechanisms of primary endocrine resistance attributable to TP53 interaction require further investigation in clinical studies (34,35). Moreover, TP53 loss of heterozygosity is an early mutational event in tumors from patients with a germline TP53 pathogenic variant (36) that could interfere with endocrine treatment responsiveness from initiation of treatment.

Changes in guidelines and migration to multigene panel testing has been responsible for the identification of increasing numbers of TP53 pathogenic variation carriers outside the classic Li-Fraumeni syndrome phenotype (14,37-40). A new Li-Fraumeni syndrome classification proposed in 2021 included the category of “attenuated Li-Fraumeni syndrome” for TP53 carriers without a personal history of cancer who were younger than 18 years of age and did not meet classic or revised Chompret criteria (23). In fact, 37.0% of the present cohort met the “attenuated Li-Fraumeni syndrome” criterion; eligibility for this cohort included a personal history of breast cancer only, which is different from most other Li-Fraumeni syndrome cohorts (2,10,41). In this way, our cohort is similar to other modern reports: 22% of participants from the German Li-Fraumeni syndrome registry had attenuated Li-Fraumeni syndrome, and most had a personal diagnosis of breast cancer >30 years (42).

Screening recommendations for women with Li-Fraumeni syndrome include annual breast magnetic resonance imaging starting at age 20 years and the controversial addition of annual mammogram at age 30 years (14). We were unable to comment on the role of screening because only 8% of patients knew their TP53 status before their first breast cancer diagnosis, although 47.6% had at least 1 relative diagnosed with breast cancer before they were 50 years of age. In addition, we observed a relatively high prevalence of women with breast cancer diagnosed during pregnancy or lactation (n = 14 [6.2%]). Breast awareness and clinical breast exam every 6 months during pregnancy and lactation is recommended for carriers of pathogenic variants in high-penetrance breast cancer genes (43,44). Identifying women with Li-Fraumeni syndrome and engaging them in high-risk breast cancer screening would avoid advanced disease at breast cancer diagnosis.

Knowledge of TP53 germline status may also influence treatment recommendations and guidelines (14), given the increased risk of a future breast cancer and the increased risk of radiation-induced sarcomas, which may prompt mastectomies instead of breast-conserving surgery and allow some to avoid radiation (8,12-14). We observed a high uptake of mastectomies to treat in situ and invasive nonmetastatic disease (73.2% and 73.1%, respectively). In our cohort, approximately 60% of patients had undergone genetic testing within 1 year of breast cancer diagnosis. Information about the TP53 germline status could have affected surgical decisions. A higher uptake of contralateral risk-reducing mastectomy was observed in patients tested before or within 1 year of breast cancer diagnosis than among those whose Li-Fraumeni syndrome was recognized more than 1 year after their breast cancer diagnosis. These findings are similar to those described for women with BRCA1/2 pathogenic variants, where knowledge of the germline result may inform surgical decisions at the time of breast cancer diagnosis (45-47). For patients who did not undergo contralateral risk-reducing mastectomy, we found a substantial rate of metachronous contralateral breast cancer (5-year contralateral breast cancer rate = 25.3%), in accordance with prior studies (10,11). The annual rate of contralateral breast cancer in TP53 pathogenic variant carriers who are diagnosed with their first breast cancer before age 36 years has been estimated at 7% (48).

RT in TP53 carriers is usually recommended only when there is a significant risk of locoregional recurrence (12,49). In our cohort, among the 34.8% of patients who received RT, the 5-year radiation-induced sarcoma risk was almost 5%. This rate is lower than prior reports for Li-Fraumeni syndrome (6%-33%) (7,50-52), but it is still much higher than those rates observed in unselected breast cancer cohorts (0.1%-0.2% at 10 years after radiation) (53). Different definitions for radiation-induced tumors and differences in cohort sizes, median follow-up, type of radiation delivery, and total dose of radiation could explain the different rates observed among published Li-Fraumeni syndrome cohorts. Yet the risk of radiation-induced sarcomas in Li-Fraumeni syndrome is a reality, and balancing risks and benefits is important. Novel RT approaches may play a role in reducing treatment toxicities in patients with Li-Fraumeni syndrome and others (54). Moreover, with improvements in adjuvant systemic therapies for nonmetastatic HER2-positive disease (27) and an established role for CDK4/CDK6 inhibitors in the adjuvant setting for hormone receptor–positive/HER2-negative breast cancer (33,55), the benefits of adjuvant radiation to reduce the locoregional recurrence risk may not outweigh the risks for patients with Li-Fraumeni syndrome.

Among the strengths of this study are the large cohort size and the recent period of diagnosis and treatment, which enable more pertinent assessment of outcomes. Further, the cohort is limited to true germline TP53 carriers now that patients with aberrant clonal expansion/clonal hematopoiesis of indeterminate potential can be recognized and excluded and classification of missense TP53 variants is more accurate. We were able to confirm prior findings regarding young age of breast cancer onset, a significant rate of contralateral breast cancer, and radiation-induced sarcomas. Nearly two-thirds of patients received a Li-Fraumeni syndrome diagnosis through germline testing within 1 year of breast cancer diagnosis, and these results may have influenced mastectomy uptake and RT omission.

This study has several limitations, as well, including its retrospective and descriptive nature, incomplete data available from databases and medical records, possible ascertainment bias, lack of molecular profiling data, and the short median follow-up.

In this study of Li-Fraumeni syndrome–associated breast cancer, a predominance of hormone receptor–positive/HER2-negative and HER2-positive tumors was confirmed. The high recurrence rate for hormone receptor–positive/HER2-negative subtype was an unexpected observation that requires confirmation in future datasets ideally supported with correlative biologic findings. This novel finding may affect recommendations for breast cancer management in patients with Li-Fraumeni syndrome.

Data availability

Deidentified data will be shared on request from the corresponding author with approval by the lead collaborators. Because the data include clinical and germline mutation data, some data may not be covered by the several material transfer agreements.

Author contributions

Renata L. Sandoval, MD, PhD (Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Project administration; Supervision; Validation; Visualization; Writing — original draft; Writing — review & editing); Nabihah Tayob, PhD (Formal analysis; Supervision; Writing — review & editing); Amanda B. Spurdle, PhD (Writing — review & editing); Cristina Fortuno, PhD (Data curation; Writing — review & editing); Kara N. Mawell, MD, PhD (Writing — review & editing); Catherine Noguès, MD (Data curation; Writing — review & editing); Marion Imbert-Bouteille, MD, MSc (Data curation; Writing — review & editing); Olivier Caron, MD (Data curation; Writing — review & editing); Maria Isabel Achatz, MD, PhD (Data curation; Resources; Writing — review & editing); Sophie Cahill, MS (Data curation; Project administration; Supervision; Writing — review & editing); Alessandra Gennari, MD, PhD (Methodology; Writing — review & editing); Benjamin Verret, MD (Methodology; Writing — review & editing); Brittany L. Bychkovsky, MD, MSc (Methodology; Supervision; Writing — original draft; Writing — review & editing); Natalia Polidorio, MD, PhD (Data curation; Writing — review & editing); Li Tianyu, MS (Formal analysis; Methodology; Writing — review & editing); Michele Bottosso, MD (Conceptualization; Data curation; Formal analysis; Methodology; Supervision; Validation; Visualization; Writing — original draft; Writing — review & editing); Fabrice Andre, MD, PhD (Resources; Supervision; Writing — review & editing); Judy Garber, MD, MPH (Conceptualization; Funding acquisition; Investigation; Methodology; Resources; Supervision; Visualization; Writing — original draft; Writing — review & editing).

Conflicts of interest

Dr Achatz reports speaker fees with AstraZeneca; Merck, Sharp & Dohme; and Pfizer. Dr Fabrice reports research grants from AstraZeneca, Daiichi Sankyo, Roche, Lilly, Pfizer, Owkin, and Novartis and serving as a compensated speaker/advisory board (to the hospital) for AstraZeneca, Daiichi Sankyo, Roche, Lilly, Pfizer, and Relay Therapeutics; and advisory board compensation to the author for Lilly. Dr Garber reports research collaboration with Ambry Genetics and Invitae (no compensation).

Funding

This work was supported by the National Cancer Institute at the National Institutes of Health (grant No. 5R01CA242218). A.B.S. was supported by an National Health and Medical Research Council Investigator Fellowship (No. APP177524). C.F. was supported by funding from the National Breast Cancer Foundation, Australia (No. IIRS-21-102).

Acknowledgements

The funder did not play a role in the design of the study; the collection, analysis, and interpretation of the data; the writing of the manuscript; and the decision to submit the manuscript for publication.

Ethics Declaration: This study was performed in line with the principles of the Declaration of Helsinki. Institutional review boards from each academic center approved the study and provided a waiver of consent for the deidentified data.

Author notes

Renata L. Sandoval and Michele Bottosso contributed equally to this work.