https://orcid.org/0000-0002-2146-3448
Abstract
Biliary tract cancer (BTC) is a highly malignant tumor, with limited therapy regimens and short response duration. In this study, we aim to assess the efficacy and safety of the combination of camrelizumab, apatinib, and capecitabine as the first- or second-line treatment in patients with advanced BTC.
In this phase 2, nonrandomized, prospective study, eligible patients received camrelizumab (200 mg, d1, Q3W), apatinib (250 mg, qd, d1-d21, Q3W), and capecitabine (1000 mg/m², bid, d1-d14, Q3W) until trial discontinued. The primary endpoint was the objective response rate (ORR). The secondary endpoints were disease control rate, progression-free survival (PFS), overall survival (OS), and safety.
From July 2019 to April 2023, we enrolled a total of 28 patients, of whom 14 patients were in the first-line treatment setting and 14 patients were in the second-line setting. At the data cutoff (April 30, 2023), the median follow-up duration was 18.03 months. Eight of 28 patients reached objective response (ORR: 28.57%), with an ORR of 50% and 7.1% for first-line and second-line treatment patients (P = .033). The median PFS was 6.30 months and the median OS was 12.80 months. Grade 3 or 4 adverse events (AEs) occurred in 9 (32.14%) patients, including elevated transaminase, thrombocytopenia, etc. No serious treatment-related AEs or treatment-related deaths occurred.
In this trial, the combination of camrelizumab, apatinib, and capecitabine showed promising antitumor activity and manageable toxicity in patients with advanced BTC, especially in the first-line setting.
In this study, camrelizumab combined with apatinib and capecitabine was used for the first time as a first- or second-line treatment regimen for advanced biliary tract cancer (BTC) and showed significant antitumor activity and manageable toxicities. Although no direct comparisons can be made, the clinical outcomes of the combination were commendable when compared with those reported in trials of traditional chemotherapy and immune monotherapy. This study adds to the evidence supporting the combination of anti-PD-1 therapy with anti-angiogenesis therapy and chemotherapy for patients with advanced BTC.
Background
Biliary tract cancer (BTC) refers to a group of invasive adenocarcinomas originating from the biliary tract, including gallbladder cancer, intrahepatic cholangiocarcinoma (ICC), and extrahepatic cholangiocarcinoma.1 Biliary tract cancer is generally uncommon, but its incidence has shown a consistent and steady rise in recent years.2 Most patients are diagnosed with advanced-stage disease and fail to receive radical surgery, exhibiting poor prognoses.3,4
Chemotherapy is the most common treatment for advanced BTC, including gemcitabine combined with cisplatin (CisGem) as the first-line treatment,5-7 and 5-fluorouracil-leucovorin-oxaliplatin (FOLFOX) as the second-line treatment.8 However, the current treatment regimen is limited, and the exploration of chemotherapy regimens over the past decade failed to show a significant improvement in patient survival and prognosis.8
In recent years, immune checkpoint inhibitor (ICI) therapy has shown significant antitumor activity in several solid tumors and hematological tumors.9 And ICIs have initially led to breakthrough progress in overall survival (OS) and response rate for patients with BTC.10 In the Keynote-966 study, pembrolizumab plus gemcitabine and cisplatin showed a clinically meaningful improvement in OS in advanced patients with BTC. Importantly, based on the TOPAZ study,11 durvalumab combined with CisGem has been listed as the preferred first-line treatment in the National Comprehensive Cancer Network (NCCN) guidelines. Camrelizumab (SHR-1210), a fully humanized programmed death protein 1 monoclonal antibody,12 in combination with gemcitabine and oxaliplatin, has demonstrated promising antitumor activity for BTC.10
Furthermore, many studies13,14 have found that anti-angiogenesis therapy can enhance the response of ICIs by modulating immunosuppression. Apatinib, a small-molecule VEGFR tyrosine kinase inhibitor,15 showed an antitumor activity with acceptable safety for patients with advanced BTC.16,17 In addition, capecitabine, a commonly used chemotherapy drug for BTC, can effectively inhibit disease progression and improve patient survival.18,19
However, given the rare treatment options and unsatisfactory efficacy, research on novel combination therapy is warranted, and new predictive biomarkers for BTC are urgently needed.20-22 On this basis, this trial was designed to evaluate the efficacy and safety of camrelizumab combined with apatinib and capecitabine as first- or second-line treatment and explore predictive biomarkers in patients with advanced BTC.
Method
Patient characteristics
This phase 2, nonrandomized, prospective cohort study assessed the safety and efficacy of camrelizumab combined with apatinib and capecitabine for patients with BTC at the Beijing Friendship Hospital affiliated to Capital Medical University. The major inclusion criteria were age of 18-80 years old; histologically confirmed metastatic or recurrent BTC, with at most one previous treatment regimen (if disease progressed within 6 months after the last adjuvant chemotherapy, the study regimen is defined as second-line treatment); an Eastern Cooperative Oncology Group (ECOG) performance status of 0 or 1 and a life expectancy of at least 3 months; at least 1 measurable lesion following the Response Evaluation Criteria in Solid Tumors (RECIST, version 1.1)23; an adequate bone marrow hematopoietic function (hemoglobin ≥ 90 g/L, neutrophil count ≥ 1.5 × 109/L, and platelet count ≥ 75 × 109/L), hepatic function (total bilirubin ≤ 1.5ULN, aspartate aminotransferase and alanine aminotransferase ≤ 2.5ULN [for patients with ICC or liver metastases ≤ 5ULN]), and renal function (blood creatinine ≤ 177umol/L and urinary protein ≤2+). Major exclusion criteria included history of anti-PD-1 or anti-PD-L1 therapy; previous treatment with VEGFR-TKIs or capecitabine; active or history of autoimmune disease; history of organ transplantation; serious cardiovascular and cerebrovascular diseases; neurological and mental illness; or other serious comorbidities.
Ethical committee clearance
The trial was approved by the ethics board of the Beijing Friendship Hospital affiliated to Capital Medical University and was done in accordance with the Declaration of Helsinki. All patients provided written informed consent. This trial was registered with ClinicalTrials.gov, number NCT 04720131.
Procedures
Eligible patients received 200 mg intravenous camrelizumab in the first day of a 21-day cycle, 250 mg oral apatinib once daily continuously, and 1000 mg/m2 oral capecitabine twice daily in the first 14 days until confirmed disease progression, death, intolerable adverse events (AEs), withdrawal of consent, or investigator’s decision.
Treatment response was assessed at baseline and every 6 weeks by the investigator, using enhanced CT or MRI based on the RECIST criteria. Adverse events were monitored by physical examination and laboratory tests (including complete blood count, serum chemistry, ECG, urine test, thyroid hormone test, etc.) and graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events (NCI-CTCAE, version 4.03) during the course of treatment. Immune-related AEs (irAEs) were identified by the investigator and managed according to the NCCN Clinical Practice Guidelines in Oncology: Management of Immunotherapy-Related Toxicities (Version 1.2019).24
PD-L1 expression was assessed in formalin-fixed paraffin-embedded tissue specimens by immunohistochemistry (VENTANA PD-L1 [SP263] assay) before the first dosage. PD-L1 CPS was defined as the proportion of PD-L1-positive cells (tumor cells, lymphocytes, and macrophages) in the total number of tumor cells multiplied by 100. Whole exome sequencing was done in tumor samples by next-generation sequencing (NGS; Illumina sequencing). Microsatellite stability assay was assessed in tumor samples by NGS, and the tumor mutational burden (TMB) was assessed based on more than 5% non-synonymous variants of allele frequency. TMB-high cutoff was defined on the upper quartile of all patients’ TMB with NGS information. The patient disposition of the trial is shown in Figure 1.
Patient disposition.
Outcomes
The primary endpoint was objective response rate (ORR, the proportion of patients with measurable disease who reached complete response [CR] or partial response [PR]). The secondary endpoints included the disease control rate (DCR, the proportion of patients with CR, PR, or stable disease [SD]), progression-free survival (PFS, time from the first dosage to the disease progression or death from any cause), OS (time from the first dosage to death from any cause) and safety. Exploratory endpoints were to assess the association between treatment response and PD-L1 expression, genomic mutations status, AEs, and other prognostic factors.
Statistical analysis
On the basis of the response rates of gemcitabine and cisplatin as first-line treatment and oxaliplatin as second-line treatment in BTC,5,8 we set a null hypothesis of 20% ORR, and we hypothesized an expected ORR of 60% for first-line treatment patients; we set a null hypothesis of 5% ORR, and hypothesized an expected ORR of 35% for second-line treatment patients. With a power of 90% and one-sided α of 5%, using the phase 2 Simon’s 2-stage design, we calculated that 14 patients were needed in the first-line treatment cohort and 14 patients were needed in the second-line treatment cohort. If 5 or more patients reached an objective response in the first-line treatment, and 2 or more patients reached an objective response in the second-line treatment, the primary endpoint was reached.
All patients enrolled with post-baseline image evaluation were assessed for the treatment response and calculated the 95% CIs of ORR and DCR using an exact binomial test. The PFS and OS and their 95% CIs were analyzed using the Kaplan-Meier survival curse. Safety data were analyzed in patients with at least one dose of the regimen. In the post hoc exploratory analysis, we assessed treatment efficacy between subgroups, including first- or second-line treatment, biomarkers (PD-L1 expression and TMB), history of radical surgery, corticosteroid treatment, etc. All statistical analyses were 2-sided and significance was set at P < .05. The software used for statistical analyses included R software (version 4.3.2), GraphPad Prism 8, and SPSS Statistics (version 25.0).
Results
Patient characteristics
Between July 2019 and March 2023, 33 patients were screened and 28 patients were enrolled from Beijing Friendship Hospital, Capital Medical University. Among them, 14 patients were included in the first-line treatment cohort and 14 were included in the second-line treatment cohort. In addition, in all enrolled patients, there were 20 patients with information of PD-L1 expression and NGS data. The patient’s baseline characteristics are shown in Table 1.
Patients characteristics.
| Age (years) | |
| Median | 65 |
| Range | 42-75 (IQR 59.75-67) |
| Sex, no. (%) | |
| Male | 20 (71%) |
| Female | 8 (29%) |
| ECOG PSa, no. (%) | |
| 0 | 9 (32%) |
| 1 | 19 (68%) |
| Primary tumor site, no. (%) | |
| Gallbladder | 8 (29%) |
| Intrahepatic | 3 (11%) |
| Perihilar | 4 (14%) |
| Distal bile duct | 13 (46%) |
| Previous curative surgery, no. (%) | |
| Yes | 18 (64%) |
| No | 10 (36%) |
| The site of metastases, no. (%) | |
| Liver | 22 (79%) |
| Peritoneum | 7 (25%) |
| Distant lymph node | 5 (18%) |
| Lung | 1 (4%) |
| Bone | 1 (4%) |
| Treatment setting, no. (%) | |
| First line | 14 (50%) |
| Second line | 14 (50%) |
| PD-L1a, no. (%) | |
| PD-L1 CPS ≥ 1 | 7 (35.0%) |
| PD-L1 CPS < 1 | 13 (65.0%) |
| Microsatellite status, no. (%) | |
| MSS | 18 (90.0%) |
| MSI | 2 (10.0%) |
| TMB, mediana | 2.54 (IQR 1.41–4.50) |
| Age (years) | |
| Median | 65 |
| Range | 42-75 (IQR 59.75-67) |
| Sex, no. (%) | |
| Male | 20 (71%) |
| Female | 8 (29%) |
| ECOG PSa, no. (%) | |
| 0 | 9 (32%) |
| 1 | 19 (68%) |
| Primary tumor site, no. (%) | |
| Gallbladder | 8 (29%) |
| Intrahepatic | 3 (11%) |
| Perihilar | 4 (14%) |
| Distal bile duct | 13 (46%) |
| Previous curative surgery, no. (%) | |
| Yes | 18 (64%) |
| No | 10 (36%) |
| The site of metastases, no. (%) | |
| Liver | 22 (79%) |
| Peritoneum | 7 (25%) |
| Distant lymph node | 5 (18%) |
| Lung | 1 (4%) |
| Bone | 1 (4%) |
| Treatment setting, no. (%) | |
| First line | 14 (50%) |
| Second line | 14 (50%) |
| PD-L1a, no. (%) | |
| PD-L1 CPS ≥ 1 | 7 (35.0%) |
| PD-L1 CPS < 1 | 13 (65.0%) |
| Microsatellite status, no. (%) | |
| MSS | 18 (90.0%) |
| MSI | 2 (10.0%) |
| TMB, mediana | 2.54 (IQR 1.41–4.50) |
aThere were 20 patients with PD-L1 status and NGS information.
Abbreviations: ECOG PS, Eastern Cooperative Oncology Group Performance status; MSS, mismatch repair; MSI, microsatellite instability; CPS, combined positive score; TMB, tumor mutational burden.
Patients characteristics.
| Age (years) | |
| Median | 65 |
| Range | 42-75 (IQR 59.75-67) |
| Sex, no. (%) | |
| Male | 20 (71%) |
| Female | 8 (29%) |
| ECOG PSa, no. (%) | |
| 0 | 9 (32%) |
| 1 | 19 (68%) |
| Primary tumor site, no. (%) | |
| Gallbladder | 8 (29%) |
| Intrahepatic | 3 (11%) |
| Perihilar | 4 (14%) |
| Distal bile duct | 13 (46%) |
| Previous curative surgery, no. (%) | |
| Yes | 18 (64%) |
| No | 10 (36%) |
| The site of metastases, no. (%) | |
| Liver | 22 (79%) |
| Peritoneum | 7 (25%) |
| Distant lymph node | 5 (18%) |
| Lung | 1 (4%) |
| Bone | 1 (4%) |
| Treatment setting, no. (%) | |
| First line | 14 (50%) |
| Second line | 14 (50%) |
| PD-L1a, no. (%) | |
| PD-L1 CPS ≥ 1 | 7 (35.0%) |
| PD-L1 CPS < 1 | 13 (65.0%) |
| Microsatellite status, no. (%) | |
| MSS | 18 (90.0%) |
| MSI | 2 (10.0%) |
| TMB, mediana | 2.54 (IQR 1.41–4.50) |
| Age (years) | |
| Median | 65 |
| Range | 42-75 (IQR 59.75-67) |
| Sex, no. (%) | |
| Male | 20 (71%) |
| Female | 8 (29%) |
| ECOG PSa, no. (%) | |
| 0 | 9 (32%) |
| 1 | 19 (68%) |
| Primary tumor site, no. (%) | |
| Gallbladder | 8 (29%) |
| Intrahepatic | 3 (11%) |
| Perihilar | 4 (14%) |
| Distal bile duct | 13 (46%) |
| Previous curative surgery, no. (%) | |
| Yes | 18 (64%) |
| No | 10 (36%) |
| The site of metastases, no. (%) | |
| Liver | 22 (79%) |
| Peritoneum | 7 (25%) |
| Distant lymph node | 5 (18%) |
| Lung | 1 (4%) |
| Bone | 1 (4%) |
| Treatment setting, no. (%) | |
| First line | 14 (50%) |
| Second line | 14 (50%) |
| PD-L1a, no. (%) | |
| PD-L1 CPS ≥ 1 | 7 (35.0%) |
| PD-L1 CPS < 1 | 13 (65.0%) |
| Microsatellite status, no. (%) | |
| MSS | 18 (90.0%) |
| MSI | 2 (10.0%) |
| TMB, mediana | 2.54 (IQR 1.41–4.50) |
aThere were 20 patients with PD-L1 status and NGS information.
Abbreviations: ECOG PS, Eastern Cooperative Oncology Group Performance status; MSS, mismatch repair; MSI, microsatellite instability; CPS, combined positive score; TMB, tumor mutational burden.
Efficacy
At the data cutoff time (April 30, 2023), the median follow-up duration was 18.03 months (95% CI, 12.9-23.2). With 5 patients still on the study treatment and 23 patients discontinuing this trial (19 patients discontinued due to disease progression, 2 patients due to intolerable AEs, 2 patients due to consent withdrawal), the median PFS (mPFS) was 6.30 months (95% CI, 2.85-9.75), and the mPFS of patients in first- and second-line treatment groups were 6.30 and 5.70 months (HR 0.85, 95% CI, 0.34-2.11, P [logrank] = .72). Nine (32.1%) of 28 patients were still alive and 19 (67.9%) patients had died, the median OS (mOS) was 12.80 months (95% CI, 7.26-18.41), and mOS were 13.77 and 7.87 months, respectively, in the first- and second-line treatment groups (HR 0.89, 95% CI, 0.35-2.24, P [logrank] = .79; Figure 2).
Kaplan-Meier curves for PFS and OS in all patients. (A) The PFS in all patients. (B) The OS in all patients. (C) The PFS for treatment setting subgroups. (D) The OS for treatment setting subgroups.
Of all enrolled patients, 1 (3.6%) patient achieved CR, 7 (25.0%) patients achieved PR, 18 (64.3%) patients reached SD, and 2 (7.1%) patients were evaluated with progressive disease (PD; Figure 3). Objective response was achieved in 8 of 28 patients (ORR: 28.6%, 95% CI, 15.3%-47.1%), and the ORR was 50% (95% CI, 26.8%-73.2%) in the first-line treatment group and was 7.14% (95% CI, 1.27%-31.47%) in the second-line treatment group (HR = 7.003, P = .033). The disease control was achieved in 26 of 28 patients (DCR: 92.86%, 95% CI, 77.4%-98.0%; Table 2).
The maximum tumor shrinkage from baseline in evaluable lesions (N = 28).
The best treatment responses of camrelizumab combined with apatinib and capecitabine in patients with BTC (N = 28).
| Best response | All patients enrolled (N = 28) | First-line treatment (N = 14) | Second-line treatment (N = 14) |
|---|---|---|---|
| Complete response | 1 (3.57%) | 1 (7.14%) | 0 (0%) |
| Partial response | 7 (25.00%) | 6 (42.86%) | 1 (7.14%) |
| Stable disease | 18 (64.29%) | 6 (42.86%) | 12 (85.71%) |
| Progressive disease | 2 (7.14%) | 1 (7.14%) | 1 (7.14%) |
| Objective response | 8 (28.57%) | 7 (50.00%) | 1 (7.14%) |
| Disease control | 26 (92.86%) | 13 (92.86%) | 13 (92.86%) |
| Best response | All patients enrolled (N = 28) | First-line treatment (N = 14) | Second-line treatment (N = 14) |
|---|---|---|---|
| Complete response | 1 (3.57%) | 1 (7.14%) | 0 (0%) |
| Partial response | 7 (25.00%) | 6 (42.86%) | 1 (7.14%) |
| Stable disease | 18 (64.29%) | 6 (42.86%) | 12 (85.71%) |
| Progressive disease | 2 (7.14%) | 1 (7.14%) | 1 (7.14%) |
| Objective response | 8 (28.57%) | 7 (50.00%) | 1 (7.14%) |
| Disease control | 26 (92.86%) | 13 (92.86%) | 13 (92.86%) |
Bold font emphasizes key points. All data are presented as n (%).
The best treatment responses of camrelizumab combined with apatinib and capecitabine in patients with BTC (N = 28).
| Best response | All patients enrolled (N = 28) | First-line treatment (N = 14) | Second-line treatment (N = 14) |
|---|---|---|---|
| Complete response | 1 (3.57%) | 1 (7.14%) | 0 (0%) |
| Partial response | 7 (25.00%) | 6 (42.86%) | 1 (7.14%) |
| Stable disease | 18 (64.29%) | 6 (42.86%) | 12 (85.71%) |
| Progressive disease | 2 (7.14%) | 1 (7.14%) | 1 (7.14%) |
| Objective response | 8 (28.57%) | 7 (50.00%) | 1 (7.14%) |
| Disease control | 26 (92.86%) | 13 (92.86%) | 13 (92.86%) |
| Best response | All patients enrolled (N = 28) | First-line treatment (N = 14) | Second-line treatment (N = 14) |
|---|---|---|---|
| Complete response | 1 (3.57%) | 1 (7.14%) | 0 (0%) |
| Partial response | 7 (25.00%) | 6 (42.86%) | 1 (7.14%) |
| Stable disease | 18 (64.29%) | 6 (42.86%) | 12 (85.71%) |
| Progressive disease | 2 (7.14%) | 1 (7.14%) | 1 (7.14%) |
| Objective response | 8 (28.57%) | 7 (50.00%) | 1 (7.14%) |
| Disease control | 26 (92.86%) | 13 (92.86%) | 13 (92.86%) |
Bold font emphasizes key points. All data are presented as n (%).
Biomarkers
In the exploratory analysis, 20 patients had PD-L1 expression and NGS information. Among them, 7 (35.0%) patients were PD-L1 CPS ≥ 1, 13 (65.0%) patients were PD-L1 CPS < 1; 18 patients were MSS, 2 patients were microsatellite instability (MSI); the median-TMB was 2.54 mut/Mb (IQR 1.41-5.03; Table 1). Five of 20 patients (25%) with TMB in the upper quartile cutoff (≥5.03 mut/Mb) were classified as having high TMB. Two (28.6%) of 7 patients with PD-L1 CPS ≥ 1 achieved objective response, and 5 (38.5%) of 13 patients with PD-L1 CPS < 1 achieved objective response (HR = 0.74, P = .658). Moreover, the ORR was 60% (3/5) and 26.7% (4/15) in TMB-high and TMB-low patients, respectively (HR = 2.25, P = .290). Two patients with MSI only achieved SD. In addition, the common DNA mutations included TP53 (75%), KMT2C (20%), CDKN2A (35%), ATM (30%), TET2 (30%), and STED2 (30%), 1 patient had a mutation in FGFR2 (5%) and no patients had mutations in KRAS (0%) and IDH1 (0%). The details of genetic variations are shown in Figure 4.
Heatmap of exome mutation.
In the post hoc prognosis analysis, the mPFS was 6.30 months (95% CI, 3.14-9.47) in patients with PD-L1 CPS ≥ 1 and 10.6 months (95% CI, 3.89-17.31) in patients with PD-L1 CPS < 1 (HR: 1.98, 95% CI, 0.55-7.14, P [logrank] = .363), and the mOS was 10.27 months (95% CI, 4.37-16.17) in patients with PD-L1 CPS ≥ 1 and 16.97 months (95% CI, 4.86-29.08) in patients with PD-L1 CPS < 1 (HR: 2.23, 95% CI, 0.60-8.37, P [logrank] = .143; Figure 5). Patients with PD-1 CPS ≥ 1 exhibit a poorer prognosis trend, but no significant statistical difference in ORR, PFS, and OS was shown. The mPFS was 10.59 months (95% CI, 1.205-14.729) in TMB-high subgroup and 5.57 months (95% CI, 0.75-10.38) in TMB-low subgroup (HR: 0.63, 95% CI, 0.20-1.98, P [logrank] = .243), and the mOS was 13.62 (95% CI) and 13.77 months (95% CI), respectively (HR: 1.11, 95% CI, 0.33-3.77, P [logrank] = .864; Figure 5. In addition, patients with TET2 or STED2 mutation showed better PFS compared with wild type (TET2: mPFS not reached vs 5.17 months, P = .038; STED2: mPFS 9.68 vs 5.17 months, P = .023). More exploratory analyses of associations between survival and genetic mutations in TET2, STED2, and TP53 are shown in Supplementary Material.
Kaplan-Meier curves for PFS and OS of PD-L1 and TMB subgroup analysis (N = 28). (A) The PFS for PD-L1 status subgroups. (B) The OS for PD-L1 status subgroups. (C) The PFS for TMB status subgroups. (D) The OS for TMB status subgroups.
Safety
All 28 patients were included in the safety analysis. The overall incidence of any grade AEs was 100%. The most common AEs were anemia in 18 (64.3%) patients, neutrophil count decreased in 11 patients (39.3%), platelet count decreased in 12 patients (42.9%), AST increased in 21 patients (75.0%), ALT increased in 11 patients (39.3%), and hypoalbuminemia in 14 patients (50.0%). Grade 3 or 4 AEs occurred in 9 (32.14%) patients, including elevated transaminase, thrombocytopenia, proteinuria, abdominal pain, and rash. With symptomatic treatment, dose reduction, and dosage interruptions, AEs in this study were manageable and tolerable. Eventually, there were 2 (7.1%) patients who discontinued this trial for AEs, and no serious treatment-related AEs or treatment-related deaths occurred. Details of AEs are shown in Table 3.
The incidence of AEs in enrolled patients (N = 28).
| Adverse events | Any grade | Grade 3-4 (n) |
|---|---|---|
| Hematological | ||
| Anemia | 18 (64.29%) | |
| Neutrophil count decreased | 11 (39.29%) | 2 (7.14%) |
| Platelet count decreased | 12 (42.86%) | 3 (10.71%) |
| Non-hematological | ||
| ALT increased | 11 (39.29%) | 1 (3.57%) |
| AST increased | 21 (75.00%) | 4 (14.29%) |
| Hypoalbuminemia | 14 (50.00%) | 2 (7.14%) |
| Hypothyroidism | 8 (28.57%) | |
| Hyperthyroidism | 12 (42.86%) | |
| Abdominal pain | 10 (35.71%) | 1 (3.57%) |
| Diarrhea | 6 (21.43%) | |
| Anorexia | 8 (28.57%) | |
| Nausea | 7 (25.00%) | |
| Rash | 11 (39.29%) | 1 (3.57%) |
| Mucositis | 2 (7.14%) | |
| Arthritis | 1 (3.57%) | |
| Impaired fasting glucose | 8 (28.57%) | |
| Adrenocortical hypofunction | 2 (7.14%) | |
| Reactive cutaneous capillary endothelial proliferation | 5 (17.86%) | |
| Hypertension | 2 (7.14%) | |
| Proteinuria | 5 (17.86%) | |
| Fatigue | 6 (21.43%) | |
| Dizziness | 2 (7.14%) | |
| Upper gastrointestinal bleeding | 1 (3.57%) | |
| Intestinal obstruction | 1 (3.57%) | |
| Adverse events | Any grade | Grade 3-4 (n) |
|---|---|---|
| Hematological | ||
| Anemia | 18 (64.29%) | |
| Neutrophil count decreased | 11 (39.29%) | 2 (7.14%) |
| Platelet count decreased | 12 (42.86%) | 3 (10.71%) |
| Non-hematological | ||
| ALT increased | 11 (39.29%) | 1 (3.57%) |
| AST increased | 21 (75.00%) | 4 (14.29%) |
| Hypoalbuminemia | 14 (50.00%) | 2 (7.14%) |
| Hypothyroidism | 8 (28.57%) | |
| Hyperthyroidism | 12 (42.86%) | |
| Abdominal pain | 10 (35.71%) | 1 (3.57%) |
| Diarrhea | 6 (21.43%) | |
| Anorexia | 8 (28.57%) | |
| Nausea | 7 (25.00%) | |
| Rash | 11 (39.29%) | 1 (3.57%) |
| Mucositis | 2 (7.14%) | |
| Arthritis | 1 (3.57%) | |
| Impaired fasting glucose | 8 (28.57%) | |
| Adrenocortical hypofunction | 2 (7.14%) | |
| Reactive cutaneous capillary endothelial proliferation | 5 (17.86%) | |
| Hypertension | 2 (7.14%) | |
| Proteinuria | 5 (17.86%) | |
| Fatigue | 6 (21.43%) | |
| Dizziness | 2 (7.14%) | |
| Upper gastrointestinal bleeding | 1 (3.57%) | |
| Intestinal obstruction | 1 (3.57%) | |
Data are presented as n (%)a.
aPatients are counted for each applicable specific AE and could have more than one treatment-related event.
Abbreviations: AE, adverse events; ALT, alanine aminotransferase; AST, aspartate aminotransferase.
The incidence of AEs in enrolled patients (N = 28).
| Adverse events | Any grade | Grade 3-4 (n) |
|---|---|---|
| Hematological | ||
| Anemia | 18 (64.29%) | |
| Neutrophil count decreased | 11 (39.29%) | 2 (7.14%) |
| Platelet count decreased | 12 (42.86%) | 3 (10.71%) |
| Non-hematological | ||
| ALT increased | 11 (39.29%) | 1 (3.57%) |
| AST increased | 21 (75.00%) | 4 (14.29%) |
| Hypoalbuminemia | 14 (50.00%) | 2 (7.14%) |
| Hypothyroidism | 8 (28.57%) | |
| Hyperthyroidism | 12 (42.86%) | |
| Abdominal pain | 10 (35.71%) | 1 (3.57%) |
| Diarrhea | 6 (21.43%) | |
| Anorexia | 8 (28.57%) | |
| Nausea | 7 (25.00%) | |
| Rash | 11 (39.29%) | 1 (3.57%) |
| Mucositis | 2 (7.14%) | |
| Arthritis | 1 (3.57%) | |
| Impaired fasting glucose | 8 (28.57%) | |
| Adrenocortical hypofunction | 2 (7.14%) | |
| Reactive cutaneous capillary endothelial proliferation | 5 (17.86%) | |
| Hypertension | 2 (7.14%) | |
| Proteinuria | 5 (17.86%) | |
| Fatigue | 6 (21.43%) | |
| Dizziness | 2 (7.14%) | |
| Upper gastrointestinal bleeding | 1 (3.57%) | |
| Intestinal obstruction | 1 (3.57%) | |
| Adverse events | Any grade | Grade 3-4 (n) |
|---|---|---|
| Hematological | ||
| Anemia | 18 (64.29%) | |
| Neutrophil count decreased | 11 (39.29%) | 2 (7.14%) |
| Platelet count decreased | 12 (42.86%) | 3 (10.71%) |
| Non-hematological | ||
| ALT increased | 11 (39.29%) | 1 (3.57%) |
| AST increased | 21 (75.00%) | 4 (14.29%) |
| Hypoalbuminemia | 14 (50.00%) | 2 (7.14%) |
| Hypothyroidism | 8 (28.57%) | |
| Hyperthyroidism | 12 (42.86%) | |
| Abdominal pain | 10 (35.71%) | 1 (3.57%) |
| Diarrhea | 6 (21.43%) | |
| Anorexia | 8 (28.57%) | |
| Nausea | 7 (25.00%) | |
| Rash | 11 (39.29%) | 1 (3.57%) |
| Mucositis | 2 (7.14%) | |
| Arthritis | 1 (3.57%) | |
| Impaired fasting glucose | 8 (28.57%) | |
| Adrenocortical hypofunction | 2 (7.14%) | |
| Reactive cutaneous capillary endothelial proliferation | 5 (17.86%) | |
| Hypertension | 2 (7.14%) | |
| Proteinuria | 5 (17.86%) | |
| Fatigue | 6 (21.43%) | |
| Dizziness | 2 (7.14%) | |
| Upper gastrointestinal bleeding | 1 (3.57%) | |
| Intestinal obstruction | 1 (3.57%) | |
Data are presented as n (%)a.
aPatients are counted for each applicable specific AE and could have more than one treatment-related event.
Abbreviations: AE, adverse events; ALT, alanine aminotransferase; AST, aspartate aminotransferase.
In post hoc analysis, 18 patients (64.3%) with irAEs were evaluated for survival outcome. The mOS was 10.27 and 12.83 months in patients with and without irAEs (HR 0.56, 95% CI, 0.21-1.51, P = .20), and the mPFS was 7.12 and 4.67 months, respectively (HR 0.73, 95% CI, 0.24-2.21, P = .52). Eleven patients (39.3%) received at least once corticosteroid for treatment or diagnosis of irAEs. Patients who received corticosteroid seem to have a longer mPFS (7.97 vs 5.07 months, HR 0.52, 95% CI, 0.21-1.27, P = .14). The median OS of patients with and without corticosteroid treatment were 13.77 and 9.77 months, respectively (HR 0.92, 95% CI, 0.36-2.30, P = .849; Figure 6).
Kaplan-Meier curves for PFS and OS of irAEs subgroup analysis (N = 28). (A) The PFS for irAEs subgroups. (B) The OS for irAEs subgroups. (C) The PFS for corticosteroid treatment subgroups. (D) The OS for corticosteroid treatment subgroups.
Discussion
As early as 2019, designed to explore new immunotherapy regimens, this phase 2, nonrandomized, prospective study recruited patients with advanced BTC for the first-line and second-line treatment. In this trial, camrelizumab plus apatinib and capecitabine significantly improved the OS and ORR, and showed a manageable safety profile as first- or second-line therapy for patients with unresectable locally advanced or metastatic BTC.
The recently reported clinical trials of ICIs are changing the standard first-line treatment for advanced BTC of the last 10 years.11,25,26 In the preplanned interim analysis of TOPAZ-1 trial,11 which is the first global estimation of the efficacy of durvalumab (PD-L1 inhibitor) combined with CisGem as the first-line treatment for patients with BTC, the ORR reached 26.7%, mPFS reached 7.2 months and mOS reached 12.8 months, making durvalumab a recommended treatment in ESMO and NCCN guidelines. The KEYNOTE-966 trial,25 the first global phase 3 study of a PD-1 inhibitor for advanced BTC, reported pembrolizumab plus CisGem significantly improved the OS (median OS 12.7 vs 10.9 months). In IMMUCHEC trial,27 a proof-of-concept study, CisGem plus durvalumab with or without tremelimumab showed similar efficacy for BTC. In addition, ICIs combined with anti-angiogenic targeted therapy have shown promising prospects for second-line treatment of advanced BTC. Anlotinib plus TQB2450 (an anti-PD-L1 antibody) demonstrated an ORR of 21.21%, a mPFS of 6.24 months, and a mOS of 15.77 months in pretreated advanced patients with BTC.28 In the analysis of the BTC population from LEAP-005 trial,17 lenvatinib plus pembrolizumab as second-line treatment demonstrated that ORR was 10% and mPFS 6.1 months, but mOS was only 8.6 months. Moreover, a previous study16 reported that camrelizumab plus apatinib as a second-line treatment showed an ORR of 19% and a mOS of 13.1 months.
In our study, with the ORR reaching 50%, the first-line treatment cohort reached primary endpoint. With an mPFS of 6.30 months, and mOS of 13.77 months in the first-line treatment group, camrelizumab combined apatinib and capecitabine showed significant clinical antitumor activity as a first-line treatment for advanced BTC. However, with the ORR of 7.1%, the second-line treatment cohort failed to achieve the primary endpoint. Nevertheless, the median PFS and median OS of the second-line treatment group were 5.70 and 7.87 months, and the second-line treatment still brought improvements in OS and PFS for patients. Although no direct comparisons can be made, the clinical outcomes of the combination of camrelizumab, apatinib, and capecitabine were commendable when compared with those reported in trials of traditional chemotherapy and immune monotherapy. Our study adds to the evidence supporting the combination of anti-PD-1 therapy with anti-angiogenesis therapy and chemotherapy for patients with advanced BTC.
Although the incidence of AEs was 100% and grade 3-4 AEs occurred in 9 patients (32.14%), these events were manageable with appropriate supportive therapy. The most common AEs included ALT/AST elevated and anemia, followed by neutrophil count decreased, platelet count decreased, and thyroid dysfunction. Based on the known profiles reported in the previous study,29 the combining of ICIs and TKIs did not result in many intolerable AEs in our trial. Otherwise, there were fewer hematological AEs in this study, when compared with conventional intravenous chemotherapy.5,8 It is worth noting that this treatment regimen is well tolerated in elderly patients and provides a choice for elderly patients with BTC. In addition, irAEs occurred in 18 patients (64.3%), while corticosteroids were used in 11 patients (39.3% of all patients) to manage irAEs. The use of corticosteroids was generally short-term. In the subgroup analysis, patients with irAEs failed to show a higher response to treatment,30 but patients who received corticosteroid seem to have a longer PFS. Although there were no significant statistical differences in survival between patients with or without corticosteroids, the judicious use of corticosteroids is crucial for prolonging PFS in immunotherapy patients.
Due to the high heterogeneity and low incidence rate of BTCs, the selection of a suitable population for immunotherapy is more challenging. The post hoc analysis included PD-L1 status, TMB, gene mutation from NGS data, and other efficacy-related biomarkers (Table 1). Similar to previous research results,31 the positive rate of PD-L1 in our patients was 35%. PD-L1 status might be a useful biomarker for ICI monotherapy, but not for ICIs combined with chemotherapy, where other factors might affect the efficacy of the treatment regimen.31 However, in this study, the baseline PD-L1 expression seems to indicate a negative correlation with treatment response and long-term survival. Dynamic evaluation of PD-L1 expression during treatment may be more reliable, but it still cannot serve as a predictive biomarker in the pretreatment stage. The tumor mutation burden of malignant tumors in the biliary system is generally low,32 so we set the cutoff line to distinguish TMB status as upper quartile for each group. In our trials, the median PFS of patients with TMB-high was 10.59, which was better than 5.57 months in patients with TMB-low (HR: 0.63, 95% CI, 0.20-1.98, P [logrank] = .243). However, there was no statistical significance in TMB subgroup analysis. In addition, there were only 2 patients with instable microsatellite status, who were evaluated as SD. The common mutated genes in this study include TP53, KMT2C, CDKN2A, ATM, etc., which is consistent with the previous reports.33,34 However, only one patient (5%) was identified with FGFR2 mutation and no patient was identified with IDH1 or KRAS mutation.34 Moreover, this study provided interesting information on the correlation of some gene mutations and treatment efficacy. Exploratory subgroup analyses found patients with TET2 (P = .038) mutation or STED2 (P = .023) mutation had an increased PFS (more details are shown in the Supplementary Material). TET2 and STED2 mutation are associated with DNA methylation and tumor maintenance.35 The inactivation of tumor cell TET2 might be involved in immune escape in intrahepatic cholangiocarcinoma,36 which may also be a potential target for ICIs. However, there is currently no evidence to suggest TET2 was predictive of response to immunotherapy. Further larger sample studies are warranted to clarify the important role of this gene mutation in immunotherapy.
There are several limitations in our trial. Firstly, with a nonrandomized design and a small sample size from a single center, sampling error and selection bias could not be well controlled. Additionally, due to the COVID-19 pandemic and other factors, patient enrollment was delayed in this study, resulting in a longer study interval. Furthermore, tumor paraffin specimens were only available in a (20/28) proportion of patients resulting in insufficiency of PD‐L1 expression and NGS analysis. Finally, as the investigator-initiated trial, objective response was assessed by investigators and radiologists, rather than an independent central review.
In conclusion, the camrelizumab combined with apatinib and capecitabine in this study served as a first- or second-line treatment regimen for advanced BTC for the first time. With the ORR of 28.57% (95% CI, 15.25%-47.06%) and median PFS of 6.30 months (95% CI, 2.85-9.75) and median OS of 12.80 months (95% CI, 7.26-18.41), especially with an ORR of 50% in the first-line setting, camrelizumab combined with apatinib and capecitabine showed significant antitumor activity and manageable toxicities in patients with advanced BTC. Larger randomized studies are warranted to compare the efficacy of this regimen and chemotherapy in patients with advanced BTC.
Supplementary material
Supplementary material is available at The Oncologist online.
Acknowledgments
We thank all the patients and their families and all the participating staff.
Additional Information: First authors Chao Jing and Zhigang Bai contributed equally. Corresponding authors Wei Deng, Zhongtao Zhang, and Wei Guo contributed equally.
Author Contributions
Conception and design: Wei Deng, Zhigang Bai. Administrative support: Zhongtao Zhang, Wei Guo, Zhigang Bai. Provision of study materials or patients: Wei Deng, Zhigang Bai, Wei Guo, Kun Liu, Hongwei Wu. Collection and assembly of data: Chao Jing, Xiaobao Yang. Data analysis and interpretation: Chao Jing, Wei Deng. Manuscript writing: All authors. Final approval of manuscript: All authors.
Funding
This study was supported by the National Key Technologies R&D Program (No. 2015BAI13B09), Beijing Hospitals Authority Clinical Medicine Development of special funding support (XMLX202102).
Conflict of interest
The authors declare no potential conflicts of interest.
Ethics approval
The trial was approved by the ethics board of the Beijing Friendship Hospital affiliated to Capital Medical University and was done in accordance with the Declaration of Helsinki. All patients provided written informed consent.
Data availability
The data that support the findings of this study are available from the corresponding author upon reasonable request.
