Abstract
Objective
To compare camrelizumab plus capecitabine and oxaliplatin followed by camrelizumab plus apatinib (camre+CAPOX followed by camre+apa), CAPOX alone, and camrelizumab plus CAPOX followed by camrelizumab (camre+CAPOX followed by camre) as initial treatment for gastric or gastro-oesophageal junction adenocarcinoma.
Design
Randomised, open label, phase 3 study.
Setting
75 hospitals in China, 13 March 2019 to 16 August 2021.
Participants
885 adults (≥18 years) with previously untreated, human epidermal growth factor receptor 2 (HER2) negative, unresectable, locally advanced or metastatic gastric or gastro-oesophageal junction adenocarcinoma.
Interventions
Patients were randomised (2:2:1) to receive camre+CAPOX followed by camre+apa, CAPOX only, or camre+CAPOX followed by camre, stratified by Eastern Cooperative Oncology Group performance status, peritoneal metastasis, and programmed death ligand 1 (PD-L1) combined positive score. Assignment to camre+CAPOX followed by camre was introduced midway through enrolment.
Main outcome measures
The primary endpoint was overall survival for camre+CAPOX followed by camre+apa versus CAPOX alone in the PD-L1 positive population (combined positive score >1) and the overall population who received at least one dose of study drug. Comparisons of camre+CAPOX followed by camre versus CAPOX alone and of camre+CAPOX followed by camre+apa versus camre+CAPOX–camre were descriptive. Safety was assessed in all patients who received at least one dose of study drug.
Results
352 patients received camre+CAPOX followed by camre+apa, 349 received CAPOX alone, and 177 received camre+CAPOX followed by camre. At the time of data cut off, 454 of 592 (76.7%) deaths had occurred in the PD-L1 positive population and 709 of 878 (80.8%) in the overall population. Overall survival was longer with camre+CAPOX followed by camre+apa than with CAPOX alone in the PD-L1 positive population (median 15.0 v 12.5 months; hazard ratio 0.80 (95% CI 0.65 to 0.98); one sided P=0.02) and in the overall population (median 13.5 v 12.1 months; hazard ratio 0.80 (0.68 to 0.94); one sided P=0.004). Use of camre+CAPOX followed by camre also showed longer overall survival versus CAPOX in the PD-L1 positive population (median 15.3 v 12.5 months; hazard ratio 0.76 (0.58 to 0.97); one sided nominal P=0.01) and overall population (median 14.2 v 12.1 months; hazard ratio 0.80 (0.65 to 0.98); one sided nominal P=0.02). No overall survival benefit was observed with camre+CAPOX followed by camre+apa versus camre+CAPOX followed by camre. Treatment related adverse events of grade ≥3 occurred in 239 of 352 (67.9%) patients in the camre+CAPOX followed by camre+apa group, 158 of 349 (45.3%) in the CAPOX alone group, and 83 of 177 (46.9%) in the camre+CAPOX followed by camre group.
Conclusions
Initial treatment with camrelizumab plus CAPOX followed by camrelizumab based maintenance was associated with longer overall survival than CAPOX alone in human epidermal growth factor receptor 2 (HER2) negative, unresectable, locally advanced or metastatic gastric or gastro-oesophageal junction adenocarcinoma. Exploratory comparisons between the two camrelizumab based regimens showed no additional survival benefit, with higher rates of treatment related adverse events of grade ≥3 and treatment discontinuations when apatinib was added during maintenance.
Trial registration
ClinicalTrials.gov NCT03813784
Introduction
Gastric cancer is the fifth most commonly diagnosed malignancy and the fifth leading cause of death related to cancer worldwide,1 with an incidence that varies geographically, and particularly high rates in Asian countries.1 2 3 Gastric cancer poses a major global health challenge, largely because it is often diagnosed at an advanced stage owing to non-specific early symptoms that can be mistaken for symptoms of other gastrointestinal disorders. The five year overall survival rate for advanced stage disease is lower than 5%.4
About 80% of patients with gastric cancer have human epidermal growth factor receptor 2 (HER2) negative disease.5 6 7 Before the advent of immunotherapy, the standard initial treatment for such patients was a fluoropyrimidine plus platinum doublet chemotherapy regimen, which achieved a median overall survival of about one year and a median progression-free survival of six months.8 More recently, immune checkpoint inhibitors have transformed the treatment of HER2 negative gastric cancer. Adding an anti-programmed death 1 antibody to standard chemotherapy improves survival, particularly in patients with higher expression of programmed death ligand 1 (PD-L1), and immune checkpoint inhibitors plus chemotherapy is now a standard initial treatment option in this setting.9 10 11 12 13 In these regimens, doublet chemotherapy is typically discontinued after four to six cycles to mitigate cumulative toxicity, with patients switched to maintenance treatment with either an immune checkpoint inhibitor alone or an immune checkpoint inhibitor plus a single cytotoxic drug. However, the optimal maintenance strategy and the potential role of treatment intensification during maintenance remain uncertain, and prospective phase 3 data directly comparing immune checkpoint inhibitors based maintenance approaches are lacking. At the time our trial was designed, phase 3 evidence supporting immune checkpoint inhibitors plus chemotherapy over chemotherapy alone had not yet been established, and capecitabine plus oxaliplatin (CAPOX) was the recommended initial regimen for HER2 negative advanced disease in this setting.
Apatinib is a small molecule, highly selective tyrosine kinase inhibitor of vascular endothelial growth factor receptor 2 and is approved in China for third line treatment and beyond of advanced gastric or gastro-oesophageal junction adenocarcinoma.14 15 Its antitumour activity arises not only from the inhibition of tumour angiogenesis and tumour cell proliferation, but also from the modulation of the immunosuppressive tumour microenvironment. Preclinical studies have shown that concurrent blockade of programmed death-1 and vascular endothelial growth factor receptor 2 may produce synergistic antitumour effects.16 17 In a phase 1 study, camrelizumab (a humanised IgG4 monoclonal antibody with high affinity for programmed death 1) combined with apatinib showed encouraging antitumour activity and manageable toxicity in patients with previously treated gastric or gastro-oesophageal junction cancer and hepatocellular carcinoma.18 These findings raised the hypothesis that introducing an antiangiogenic agent during maintenance might enhance or prolong the benefits of immune checkpoint inhibitor based initial treatment, but whether such intensification provides meaningful clinical benefit beyond immune checkpoint inhibitor maintenance alone, given the potential for additional toxicity, is unknown. An earlier phase 2 study showed that camrelizumab plus chemotherapy for four to six cycles followed by camrelizumab plus apatinib achieved an objective response rate of 58.3%, a median progression-free survival of 6.8 months, and a median overall survival of 14.9 months in patients with previously untreated, HER2 negative, advanced or metastatic gastric or gastro-oesophageal junction adenocarcinoma.19
We conducted a multicentre, randomised, phase 3 trial to build on these findings and to investigate the evolving role of programmed death 1 inhibitors as initial treatment. In our study, patients received camrelizumab plus CAPOX followed by maintenance with camrelizumab plus apatinib, camrelizumab plus CAPOX followed by camrelizumab, or CAPOX alone. We evaluated whether camrelizumab plus CAPOX followed by camrelizumab based maintenance improves survival compared with CAPOX alone and explored the relative benefits and harms of adding apatinib to camrelizumab maintenance in patients with previously untreated, HER2 negative, advanced or metastatic gastric or gastro-oesophageal junction adenocarcinoma.
Methods
Study design and patients
We conducted this multicentre, randomised, open label, phase 3 trial at 75 sites across China (see supplementary table S1). Written informed consent was obtained from all participants before enrolment.
Patients were eligible if they were aged 18 years or older and had received a histologically confirmed diagnosis of unresectable, HER2 negative, locally advanced or metastatic gastric or gastro-oesophageal junction adenocarcinoma, irrespective of PD-L1 expression. Patients who had received earlier systemic anticancer therapy were ineligible for our study, except those who had completed neoadjuvant or adjuvant treatment (chemotherapy, radiotherapy, or chemoradiotherapy) for six months or longer before recurrence or progression. Additional inclusion criteria were individuals with an Eastern Cooperative Oncology Group performance status of 0 or 1, who had at least one measurable lesion according to Response Evaluation Criteria in Solid Tumours version 1.1, and adequate organ function. Patients were required to provide either a fresh tumour sample or archival tissue obtained in the six month period before randomisation. Key exclusion criteria included known HER2 positive status, uncontrolled or symptomatic active central nervous system metastases, active or suspected autoimmune disease, and earlier treatment with inhibitors for programmed death 1, PD-L1, programmed death-ligand 2, CD137, or cytotoxic T lymphocyte antigen 4, or other agents targeting T cell costimulatory or immune checkpoint pathways.
Randomisation and masking
At the time of study design, no phase 3 evidence had yet established the superiority of immune checkpoint inhibitor therapy plus chemotherapy over chemotherapy alone in patients with previously untreated, HER2 negative, advanced or metastatic gastric or gastro-oesophageal junction adenocarcinoma. The trial was therefore initially designed to compare induction therapy with camrelizumab plus capecitabine and oxaliplatin (CAPOX), followed by maintenance camrelizumab plus apatinib (camre+CAPOX followed by camre+apa), compared with CAPOX alone, with patients randomised 1:1. After the publication of results from the KEYNOTE-062 phase 3 study (pembrolizumab ±chemotherapy v chemotherapy alone)20 and regulatory feedback, a protocol amendment (6 March 2020) added a third group—camrelizumab plus CAPOX followed by camrelizumab (camre+CAPOX followed by camre)—and adjusted the randomisation to 1:1:1 once two fifths of the amended sample size had been enrolled. Consequently, the final randomisation ratio was 2:2:1 to camre+CAPOX followed by camre+apa, CAPOX alone, or camre+CAPOX followed by camre (see supplementary table S2).
Patients were randomly assigned through a centralised interactive web response system using block randomisation (block size of six), stratified by Eastern Cooperative Oncology Group performance status (0 v 1), presence of peritoneal metastasis (yes v no), and PD-L1 combined positive score (>1 v ≤1). This was an open label study: neither investigators nor patients were masked to treatment allocation.
Procedures
Patients assigned to the camre+CAPOX followed by camre+apa group received camrelizumab (200 mg intravenously every three weeks) plus CAPOX (capecitabine 1000 mg/m2 orally twice daily on days 1-14 of each 21 day treatment cycle, and oxaliplatin 130 mg/m2 intravenously on day 1 of each cycle) for four to six cycles, followed by camrelizumab (200 mg intravenously every three weeks) in combination with apatinib (250 mg orally once daily). Patients in the CAPOX group received capecitabine and oxaliplatin alone. Those in the camre+CAPOX followed by camre group received camrelizumab plus CAPOX for four to six cycles, followed by camrelizumab monotherapy (see supplementary figure S1). Dose reductions were allowed for CAPOX and apatinib but not for camrelizumab. Treatment was continued until completion of two years of camrelizumab therapy, disease progression, unacceptable toxicity, patient withdrawal, investigator decision, or study termination. In the camre+CAPOX followed by camre+apa and camre+CAPOX followed by camre groups, treatment beyond initial disease progression (according to Response Evaluation Criteria in Solid Tumours version 1.1) was permitted at the investigator’s discretion.
Investigators assessed tumours according to Response Evaluation Criteria in Solid Tumours version 1.1 using computed tomography or magnetic resonance imaging at baseline, every six weeks during the first 42 weeks, and every nine weeks thereafter. Complete or partial responses required confirmation by a subsequent scan at least four weeks after initial documentation. An independent data monitoring committee oversaw safety throughout the study. Adverse events related to treatment were collected until 90 days after the last study dose and graded according to the Common Terminology Criteria for Adverse Events, version 5.0. During survival follow-up, we obtained data on a monthly basis.
A central laboratory (Shanghai Xiawei Biotechnology, Amoy Diagnostics, Shanghai, China; certified by the College of American Pathologists) evaluated PD-L1 combined positive score using an immunohistochemistry assay with the E1L3N antibody clone.
Outcomes
The primary endpoint was overall survival with camre+CAPOX followed by camre+apa versus CAPOX alone, assessed in both the PD-L1 positive population (combined positive score >1) and the overall population. We defined overall survival as the time from randomisation to death from any cause, with patients who were alive at the end of follow-up or who were lost to follow-up censored at their last known date of contact. Secondary endpoints included overall survival rates at 12, 18, and 24 months, progression-free survival, objective response rate, disease control rate, and duration of response, all comparing camre+CAPOX followed by camre+apa with CAPOX alone, as well as safety across all three groups (camre+CAPOX followed by camre+apa, CAPOX alone, and camre+CAPOX followed by camre). Exploratory analyses evaluated efficacy outcomes between camre+CAPOX followed by camre and CAPOX, and between camre+CAPOX followed by camre+apa and camre+CAPOX followed by camre. Prespecified subgroup analyses of overall survival and progression-free survival by PD-L1 combined positive score (>1 v ≤1) and other baseline characteristics were also exploratory and were not adjusted for multiplicity.
Statistical analysis
The sample size for the camre+CAPOX followed by camre+apa and CAPOX alone groups was calculated to ensure adequate statistical power to show superiority of camre+CAPOX followed by camre+apa over CAPOX alone in overall survival in both the PD-L1 positive and overall populations, with overall type 1 error controlled at a one sided α of 0.025. Based on findings from our earlier phase 2 study,19 we increased the planned enrolment for these two groups from 568 to 708 in a protocol amendment dated 6 March 2020. In parallel, following release of the KEYNOTE-062 results,20 a third group (camre+CAPOX followed by camre) was added, increasing the total planned sample size to 885 participants (see supplementary table S2).
In the original study design, 259 deaths were expected to provide 80% power to detect a hazard ratio of 0.68 for overall survival in the PD-L1 positive population (median 16.9 months with camre+CAPOX followed by camre+apa v 11.5 months with CAPOX alone), and 438 deaths to provide 88% power to detect a hazard ratio of 0.72 in the overall population (median 16.0 v 11.5 months). We tested each comparison at a one sided α of 0.0125. Assuming a PD-L1 positive prevalence of 60%, the planned enrolment was 568 patients in total, with 284 assigned to each group.
In the final design, we applied a fixed sequence testing strategy: overall survival was first tested between camre+CAPOX followed by camre+apa and CAPOX in the PD-L1 positive population, and if statistically significant, subsequently in the overall population, both at a one sided α of 0.025. We estimated that 340 deaths would provide 88% power to detect a hazard ratio of 0.71 (median overall survival 16.2 v 11.5 months) in the PD-L1 positive population, and 576 deaths to provide 91% power to detect a hazard ratio of 0.76 (median overall survival 15.1 v 11.5 months) in the overall population. Assuming a 60% prevalence of PD-L1 positivity, the planned enrolment for the camre+CAPOX followed by camre+apa and CAPOX alone groups was 708 patients (354 in each group). As the comparison of overall survival between the camre+CAPOX followed by camre+apa and camre+CAPOX followed by camre groups was descriptive, we determined the sample size for the camre+CAPOX followed by camre group on estimation strategy rather than formal hypothesis testing. Assuming a hazard ratio of 0.89, enrolment of 354 patients in the camre+CAPOX followed by camre+apa group and 177 in the camre+CAPOX followed by camre group, with 380 deaths, would provide a 78% probability of restricting the half width of the 95% confidence interval (CI) to ≤0.31, a 65% probability to ≤0.26, and a 50% probability to ≤0.21.
We assessed efficacy in patients with PD-L1 positivity who received at least one dose of study drug, and in all randomised patients who received at least one dose of study drug. Safety analyses included all patients who received at least one dose of study drug. The Kaplan-Meier method was used to estimate median overall survival, progression-free survival, and duration of response, and the Brookmeyer-Crowley method was used to calculate 95% CIs. Survival rates were estimated with the Kaplan-Meier method, and the corresponding 95% CIs were derived using the normal approximation. We compared overall survival and progression-free survival with the log rank test, stratified by Eastern Cooperative Oncology Group performance status (0 v 1), presence of peritoneal metastasis (yes v no), and PD-L1 expression status (positive v negative; overall population only). We used stratified Cox proportional hazards models to estimate hazard ratios with 95% CIs, and Cox proportional hazards models for subgroup analyses.
For PD-L1 combined positive score, treatment-by-subgroup interactions for overall survival were explored by including an interaction term between treatment group and PD-L1 status (combined positive score >1 v ≤1) in the stratified Cox models. We calculated two sided P values for interaction without adjustment for multiple testing and thus should be interpreted as exploratory. Confirmed objective response rate and disease control rate were calculated, with 95% CIs estimated by the Clopper-Pearson method, and 95% CIs for between group differences estimated by the Newcombe method. The comparisons of camre+CAPOX followed by camre versus CAPOX and camre+CAPOX followed by camre+apa versus camre+CAPOX followed by camre were exploratory, with no formal statistical hypotheses; we report nominal one sided P values. SAS software (version 9.4) was used for all statistical analyses.
Patient and public involvement
Patients were informed of the trial objectives and procedures at the time of recruitment but were not involved in the study design. In view of the confidentiality of clinical data, patients did not participate in data analysis or preparation of the manuscript. Trial results were, however, communicated to patients who expressed interest during follow-up clinic visits.
Results
Patients
Between 13 March 2019 and 16 August 2021, 885 eligible patients were randomised. Of these patients, 878 (99.2%) received at least one dose of the assigned treatment: 352 in the camre+CAPOX followed by camre+apa group, 349 in the CAPOX group, and 177 in the camre+CAPOX followed by camre group (fig 1). Baseline characteristics were generally well balanced across groups (table 1). We observed a PD-L1 combined positive score >1 in 227 (64.5%) of 352 patients receiving camre+CAPOX followed by camre+apa, 238 (68.2%) of 349 patients receiving CAPOX, and 127 (71.8%) of 177 patients receiving camre+CAPOX followed by camre. Metastatic disease at baseline was present in all patients receiving camre+CAPOX followed by camre+apa (n=352, 100.0%), 347 (99.4%) of 349 patients receiving CAPOX, and all patients receiving camre+CAPOX followed by camre (n=177, 100.0%). The most common sites of metastasis were the liver (153 (43.5%) of 352 patients receiving camre+CAPOX followed by camre+apa, 171 (49.0%) of 349 patients receiving CAPOX, and 84 (47.5%) of 177 patients receiving camre+CAPOX followed by camre) and peritoneum (68 of (19.3%) 352 patients receiving camre+CAPOX followed by camre+apa, 71 (20.3%) of 349 patients receiving CAPOX, and 39 (22.0%) of 177 patients receiving camre+CAPOX followed by camre).
Fig 1.
Flowchart of inclusion for study on camrelizumab plus CAPOX with camrelizumab based maintenance versus CAPOX alone as preferred treatment for gastric or gastro-oesophageal junction adenocarcinoma .*43 patients were screened twice. †Patients who did not experience disease progression after four cycles of chemotherapy but were considered by the investigator to be at risk from continuing the current regimen were deemed to have completed an adequate course of chemotherapy and their treatment was discontinued. CAPOX=capecitabine and oxaliplatin; camre+CAPOX followed by camre+apa=camrelizumab plus CAPOX followed by camrelizumab plus apatinib; camre+CAPOX followed by camre=camrelizumab plus CAPOX followed by camrelizumab alone; PD-L1=programmed death ligand 1
Table 1.
Baseline personal and disease characteristics of participants assigned to receive camrelizumab plus CAPOX with camrelizumab based maintenance versus CAPOX alone in the initial treatment of gastric or gastro-oesophageal junction adenocarcinoma. Data are number (percentage) unless otherwise specified
| Characteristics | Camre+CAPOX-camre+apa* (n=352) | CAPOX (n=349) | Camre+CAPOX-camre† (n=177) |
|---|---|---|---|
| Median (IQR) age (years) | 62 (55-67) | 62 (54-68) | 60 (54-68) |
| Sex: | |||
| Male | 250 (71.0) | 272 (77.9) | 128 (72.3) |
| Female | 102 (29.0) | 77 (22.1) | 49 (27.7) |
| ECOG performance status: | |||
| 0 | 105 (29.8) | 104 (29.8) | 51 (28.8) |
| 1 | 247 (70.2) | 245 (70.2) | 126 (71.2) |
| Primary tumour location: | |||
| Gastric | 250 (71.0) | 249 (71.3) | 123 (69.5) |
| Gastro-oesophageal junction | 102 (29.0) | 100 (28.7) | 54 (30.5) |
| Metastatic disease | 352 (100.0) | 347 (99.4) | 177 (100.0) |
| No of distant metastatic sites: | |||
| ≤2 | 269 (76.4) | 273 (78.2) | 142 (80.2) |
| >2 | 83 (23.6) | 76 (21.8) | 35 (19.8) |
| Liver metastasis | 153 (43.5) | 171 (49.0) | 84 (47.5) |
| Peritoneal metastasis | 68 (19.3) | 71 (20.3) | 39 (22.0) |
| Histological subtype (Lauren classification)‡: | |||
| Diffuse | 62 (17.6) | 49 (14.0) | 28 (15.8) |
| Intestinal | 219 (62.2) | 243 (69.6) | 116 (65.5) |
| Mixed | 52 (14.8) | 45 (12.9) | 28 (15.8) |
| Unknown | 1 (0.3) | 4 (1.1) | 1 (0.6) |
| Not detected | 18 (5.1) | 8 (2.3) | 4 (2.3) |
| PD-L1 combined positive score: | |||
| ≤1 | 121 (34.4) | 108 (30.9) | 50 (28.2) |
| >1 | 227 (64.5) | 238 (68.2) | 127 (71.8) |
| <5 | 206 (58.5) | 214 (61.3) | 87 (49.2) |
| ≥5 | 142 (40.3) | 132 (37.8) | 90 (50.8) |
| <10 | 265 (75.3) | 260 (74.5) | 126 (71.2) |
| ≥10 | 83 (23.6) | 86 (24.6) | 51 (28.8) |
| Undetectable/unknown | 4 (1.1) | 3 (0.9) | 0 |
Camre=camrelizumab; CAPOX=capecitabine plus oxaliplatin; apa=apatinib; ECOG=Eastern Cooperative Oncology Group; IQR=interquartile range; PD-L1=programmed death ligand 1.
Camre+CAPOX followed by camre+apatinib.
Camre+CAPOX followed by camre.
Lauren classification is a histological classification of gastric adenocarcinoma into intestinal, diffuse, and mixed types. Unknown/not detected indicate missing or unevaluable data.
As of 7 June 2023, the median follow-up duration was 13.5 months (interquartile range (IQR) 7.5-24.4 months) in the camre+CAPOX followed by camre+apa group, 12.1 months (IQR 6.4-21.0 months) in the CAPOX group, and 13.7 (IQR 7.9-23.7) months in the camre+CAPOX followed by camre group. All 878 patients had discontinued the assigned treatment (fig 1). The most common reason for discontinuation was radiographically confirmed disease progression in both the camre+CAPOX followed by camre+apa group and the camre+CAPOX followed by camre group. However, discontinuation in the CAPOX group most often occurred after completion of at least four to six cycles of chemotherapy or because of radiographically confirmed disease progression. Per protocol, 224 (63.6%) of 352 patients in the camre+CAPOX followed by camre+apa group and 113 (63.8%) of 177 patients in the camre+CAPOX followed by camre group who remained progression-free after four to six cycles of initial camre+CAPOX subsequently received maintenance therapy with camrelizumab plus apatinib or camrelizumab alone, respectively.
Efficacy in PD-L1 positive patients
Among the 592 patients with PD-L1 positive disease, 454 (76.7%) had died by the end of data collection. Median overall survival was 15.0 months (95% CI 12.1 to 16.8) with camre+CAPOX followed by camre+apa, 12.5 months (11.0 to 14.0) with CAPOX, and 15.3 months (11.9 to 19.1) with camre+CAPOX followed by camre. The corresponding overall survival rates at 12, 18, and 24 months were 56.8% (95% CI 50.1% to 63.0%), 41.9% (35.4% to 48.2%), and 33.8% (27.7% to 40.0%) with camre+CAPOX followed by camre+apa. The corresponding overall survival rates were 52.9% (46.4% to 59.0%), 33.2% (27.3% to 39.2%), and 22.8% (17.6% to 28.3%) with CAPOX, and 58.9% (49.8% to 66.9%), 44.5% (35.7% to 53.0%), and 33.1% (25.0% to 41.4%) with camre+CAPOX followed by camre, respectively (table 2, table 3).
Table 2.
Summary of efficacy outcomes in PD-L1 positive population (combined positive score >1) for camre+CAPOX followed by camre+apa, CAPOX alone, and camre+CAPOX followed by camre. Data are percentage (95% CI) unless otherwise specified
| Outcomes | Camre+CAPOX-camre+apa*(n=227) | CAPOX (n=238) | Camre+CAPOX-camre†(n=127) |
|---|---|---|---|
| Overall survival (months) | |||
| Median (95% CI): | 15.0 (12.1 to 16.8) | 12.5 (11.0 to 14.0) | 15.3 (11.9 to 19.1) |
| Treatment v CAPOX (hazard ratio (95% CI), P value) | 0.80 (0.65 to 0.98), 0.02 | — | 0.76 (0.58 to 0.97), 0.01 |
| Treatment v camre+CAPOX-camre† (hazard ratio (95% CI), P value) | 1.08 (0.83 to 1.40), 0.29 | — | — |
| 12 months | 56.8 (50.1 to 63.0) | 52.9 (46.4 to 59.0) | 58.9 (49.8 to 66.9) |
| 18 months | 41.9 (35.4 to 48.2) | 33.2 (27.3 to 39.2) | 44.5 (35.7 to 53.0) |
| 24 months | 33.8 (27.7 to 40.0) | 22.8 (17.6 to 28.3) | 33.1 (25.0 to 41.4) |
| Progression-free survival (months) | |||
| Median (95% CI): | 7.9 (6.9 to 8.6) | 5.7 (5.6 to 6.8) | 6.9 (5.8 to 8.4) |
| Treatment v CAPOX (hazard ratio (95% CI), P value) | 0.67 (0.53 to 0.85), <0.001 | — | 0.71 (0.54 to 0.93), 0.007 |
| Treatment v camre+CAPOX-camre† (hazard ratio (95% CI), P value) | 1.02 (0.79 to 1.33), 0.44 | — | — |
| Best overall response (No (%)): | |||
| Complete response | 4 (1.8) | 1 (0.4) | 1 (0.8) |
| Partial response | 138 (60.8) | 114 (47.9) | 74 (58.3) |
| Stable disease | 55 (24.2) | 83 (34.9) | 35 (27.6) |
| Progression disease | 22 (9.7) | 21 (8.8) | 11 (8.7) |
| Not evaluable | 8 (3.5) | 19 (8.0) | 6 (4.7) |
| Confirmed objective response: | 62.6 (55.9 to 68.9) | 48.3 (41.8 to 54.9) | 59.1 (50.0 to 67.7) |
| Difference v CAPOX | 14.2 (5.2 to 22.9) | — | — |
| Disease control: | 86.8 (81.7 to 90.9) | 83.2 (77.8 to 87.7) | 86.6 (79.4 to 92.0) |
| Difference v CAPOX | 3.6 (−3.0 to 10.1) | — | — |
| Median (95% CI) duration of response (months) | 9.0 (7.6 to 12.6) | 5.4 (4.5 to 6.8) | 9.2 (5.7 to 14.5) |
The study was powered to detect the superiority of camre+CAPOX followed by camre+apa v CAPOX in overall survival (one sided stratified log rank P values are reported). camre+CAPOX followed by camre v CAPOX and camre+CAPOX followed by camre+apa v camre+CAPOX followed by camre were exploratory, without formal statistical hypotheses (nominal one sided stratified log rank P values are reported).
Camre=camrelizumab; CAPOX=capecitabine plus oxaliplatin; CI=confidence interval; PD-L1=programmed death ligand 1.
Camre+CAPOX followed by camre+apatinib.
Camre+CAPOX followed by camre.
Table 3.
Summary of efficacy outcomes in overall population for camre+CAPOX followed by camre+apa, CAPOX alone, and camre+CAPOX followed by camre. Data are percentage (95% CI) unless otherwise specified
| Outcomes | Camre+CAPOX-camre+apa* (n=352) | CAPOX (n=349) | Camre+CAPOX-camre† (n=177) |
|---|---|---|---|
| Overall survival (months) | |||
| Median (95% CI): | 13.5 (11.9 to 15.6) | 12.1 (10.7 to 13.5) | 14.2 (11.4 to 15.7) |
| Treatment v CAPOX (hazard ratio (95% CI), P value) | 0.80 (0.68 to 0.94), 0.004 | — | 0.80 (0.65 to 0.98), 0.02 |
| Treatment v camre+CAPOX-camre† (hazard ratio (95% CI), P value) | 1.02 (0.82 to 1.26), 0.45 | — | — |
| 12 months | 54.5 (49.2 to 59.6) | 50.7 (45.4 to 55.8) | 56.3 (48.7 to 63.3) |
| 18 months | 38.9 (33.8 to 44.0) | 31.2 (26.4 to 36.1) | 39.3 (32.0 to 46.4) |
| 24 months | 29.5 (24.8 to 34.3) | 19.8 (15.8 to 24.1) | 28.2 (21.7 to 35.0) |
| Progression-free survival (months) | |||
| Median (95% CI): | 7.5 (6.8 to 8.4) | 5.6 (5.4 to 5.9) | 6.7 (5.7 to 7.7) |
| Treatment v CAPOX (hazard ratio (95% CI), P value) | 0.63 (0.52 to 0.77), <0.001 | — | 0.74 (0.59 to 0.93), 0.004 |
| Treatment v camre+CAPOX-camre† (hazard ratio (95% CI), P value) | 0.92 (0.74 to 1.15), 0.23 | — | — |
| Best overall response (No (%)): | |||
| Complete response | 6 (1.7) | 2 (0.6) | 3 (1.7) |
| Partial response | 195 (55.4) | 156 (44.7) | 96 (54.2) |
| Stable disease | 104 (29.5) | 135 (38.7) | 57 (32.2) |
| Progression disease | 28 (8.0) | 31 (8.9) | 15 (8.5) |
| Not evaluable | 19 (5.4) | 25 (7.2) | 6 (3.4) |
| Confirmed objective response: | 57.1 (51.7 to 62.3) | 45.3 (40.0 to 50.7) | 55.9 (48.3 to 63.4) |
| Difference v CAPOX | 11.8 (4.4 to 19.0) | — | — |
| Disease control: | 86.6 (82.6 to 90.0) | 84.0 (79.7 to 87.6) | 88.1 (82.4 to 92.5) |
| Difference v CAPOX | 2.7 (−2.6 to 8.0) | — | — |
| Median (95% CI) duration of response (months) | 9.0 (7.9 to 11.2) | 5.5 (4.6 to 6.8) | 7.0 (5.7 to 10.3) |
The study was powered to detect the superiority of camre+CAPOX followed by camre+apa v CAPOX in overall survival (one sided stratified log rank P values are reported). camre+CAPOX followed by camre v CAPOX and camre+CAPOX followed by camre+apa v camre+CAPOX followed by camre were exploratory, without formal statistical hypotheses (nominal one sided stratified log rank P values are reported).
Camre=camrelizumab; CAPOX=capecitabine plus oxaliplatin; CI=confidence interval; PD-L1=programmed death ligand 1.
Camre+CAPOX followed by camre+apatinib.
Camre+CAPOX followed by camre.
Overall survival was significantly prolonged with camre+CAPOX followed by camre+apa compared with CAPOX (hazard ratio 0.80, 95% CI 0.65 to 0.98; one sided P=0.02; fig 2), and the survival benefit was generally consistent across prespecified subgroups (see supplementary figure S2A). Camre+CAPOX followed by camre also improved overall survival versus CAPOX (hazard ratio 0.76, 0.58 to 0.97; one sided nominal P=0.01; fig 3). By contrast, we observed no survival advantage for camre+CAPOX followed by camre+apa over camre+CAPOX followed by camre (hazard ratio 1.08, 0.83 to 1.40; one sided nominal P=0.29), indicating broadly similar overall survival between the two camrelizumab based maintenance regimens (see supplementary figure S3A).
Fig 2.
Kaplan-Meier estimates of overall survival and progression-free survival for patients using camre+CAPOX followed by camre+apa versus CAPOX as preferred treatment for gastric or gastro-oesophageal junction adenocarcinoma. Top panel shows PD-L1 positive population, bottom panel shows overall population. Log rank tests and Cox regression models were stratified by Eastern Cooperative Oncology Group performance status (0 v 1), presence versus absence of peritoneal metastasis, and PD-L1 expression status (positive v negative, overall population only). P values are one sided and based on stratified log rank tests. P values for progression-free survival comparisons are nominal. CAPOX=capecitabine and oxaliplatin; camre+CAPOX-camre+apa=camrelizumab plus CAPOX followed by camrelizumab plus apatinib; CI=confidence interval; PD-L1=programmed death ligand 1
Fig 3.
Kaplan-Meier estimates of overall survival and progression-free survival for patients using camre+CAPOX followed by camre versus CAPOX. Top panel shows PD-L1 positive population, bottom panel shows overall population. Log rank tests and Cox regression models were stratified by Eastern Cooperative Oncology Group performance status (0 v 1), presence versus absence of peritoneal metastasis, and PD-L1 expression status (positive v negative, overall population only). P values are one sided, derived from stratified log rank tests, and nominal. CAPOX=capecitabine and oxaliplatin; camre+CAPOX followed by camre=camrelizumab plus CAPOX followed by camrelizumab alone; CI=confidence interval; PD-L1=programmed death ligand 1
At the data cut off, disease progression or death had occurred in 387 (65.4%) of 592 patients with PD-L1 positive tumours. Median progression-free survival was 7.9 months (95% CI 6.9 to 8.6 months) with camre+CAPOX followed by camre+apa, 5.7 months (5.6 to 6.8) with CAPOX, and 6.9 months (5.8 to 8.4) with camre+CAPOX followed by camre (table 2). Progression-free survival was also longer with camre+CAPOX followed by camre+apa versus CAPOX (hazard ratio 0.67, 95% CI 0.53 to 0.85; one sided nominal P<0.001; fig 2) and with camre+CAPOX followed by camre versus CAPOX (0.71, 0.54 to 0.93; P=0.007; fig 3). No meaningful difference in progression-free survival was observed between camre+CAPOX followed by camre+apa and camre+CAPOX followed by camre (hazard ratio 1.02, 0.79 to 1.33; one sided nominal P=0.44; see supplementary figure S4A). Among patients randomised concurrently, the Kaplan-Meier curves for progression-free survival began to diverge at about six months, with rates of 48.4% (95% CI 38.9% to 57.2%) versus 39.4% (30.4% to 48.2%) at nine months for camre+CAPOX followed by camre+apa and 26.4% (18.1% to 35.5%) versus 21.9% (14.6% to 30.1%) at 18 months for camre+CAPOX followed by camre (supplementary file, supplementary figure S4B).
Confirmed objective response rate was 62.6% (95% CI 55.9% to 68.9%) in the camre+CAPOX followed by camre+apa group, compared with 48.3% (41.8% to 54.9%) in the CAPOX group, yielding a treatment difference of 14.2 percentage points (95% CI 5.2% to 22.9%; table 2). Disease control rate was comparable between the two groups. Among responders, the median duration of response was longer with camre+CAPOX followed by camre+apa than with CAPOX (9.0 months (95% CI 7.6 to 12.6) v 5.4 months (4.5 to 6.8)).
Efficacy in the overall population regardless of PD-L1 expression
In the overall population, 709 (80.8%) of 878 patients had died at the end of data collection. Median overall survival was 13.5 months (95% CI 11.9 to 15.6) with camre+CAPOX followed by camre+apa, 12.1 months (10.7 to 13.5) with CAPOX, and 14.2 months (11.4 to 15.7) with camre+CAPOX followed by camre. The corresponding overall survival rates at 12 months were 54.5% (95% CI 49.2% to 59.6%), 50.7% (45.4% to 55.8%), and 56.3% (48.7% to 63.3%); at 18 months, 38.9% (95% CI 33.8% to 44.0%), 31.2% (95% CI 26.4% to 36.1%), and 39.3% (95% CI 32.0% to 46.4%); and at 24 months, 29.5% (95% CI 24.8% to 34.3%), 19.8% (95% CI 15.8% to 24.1%), and 28.2% (95% CI 21.7% to 35.0%), for the camre+CAPOX followed by camre+apa, CAPOX, and camre+CAPOX followed by camre groups, respectively (table 2, table 3).
We observed a significant overall survival benefit with camre+CAPOX followed by camre+apa versus CAPOX (hazard ratio 0.80, 95% CI 0.68 to 0.94; one sided P=0.004; fig 2). This advantage was broadly consistent across prespecified subgroups (supplementary file, supplementary figure S2B). Overall survival also favoured camre+CAPOX followed by camre over CAPOX (hazard ratio 0.80, 95% CI 0.65 to 0.98; one sided nominal P=0.02; fig 3). In contrast, camre+CAPOX followed by camre+apa did not show an overall survival benefit compared with camre+CAPOX followed by camre (hazard ratio 1.02, 95% CI 0.82 to 1.26; one sided nominal P=0.45; supplementary file, supplementary figure S3B), indicating similar overall survival with the two camrelizumab based maintenance strategies. Histograms depicting monthly changes in overall survival probability are presented in the supplementary file, supplementary figure S5.
As of the data cut off, disease progression or death had occurred in 609 (69.4%) of 878 patients. Median progression-free survival was 7.5 months (95% CI 6.8 to 8.4 months) with camre+CAPOX followed by camre+apa, 5.6 months (5.4 to 5.9 months) with CAPOX, and 6.7 months (5.7 to 7.7 months) with camre+CAPOX followed by camre. Progression-free survival was markedly prolonged with camre+CAPOX followed by camre+apa versus CAPOX (hazard ratio 0.63, 95% CI 0.52 to 0.77; one sided nominal P<0.001; fig 2) and with camre+CAPOX followed by camre versus CAPOX (0.74, 0.59 to 0.93; one sided nominal P=0.004; fig 3). No meaningful difference was observed between camre+CAPOX followed by camre+apa and camre+CAPOX followed by camre (hazard ratio 0.92, 95% CI 0.74 to 1.15; one sided nominal P=0.23; supplementary file, supplementary figure S4C), suggesting comparable disease control with the two camrelizumab based maintenance approaches. Histograms illustrating the monthly change in progression-free survival probability are provided in the supplementary file, supplementary figure S6. Among patients randomised concurrently, the Kaplan-Meier curves separated at around six months, with progression-free survival rates at nine months of 44.7% (95% CI 36.8% to 52.3%) versus 36.5% (29.1% to 43.9%) and 18 month rates of 21.8% (15.2% to 29.1%) versus 17.1% (11.6% to 23.5%) for camre+CAPOX followed by camre+apa and camre+CAPOX followed by camre, respectively (supplementary file, supplementary figure S4D).
Following discontinuation of study treatment, 152 (43.2%) of 352 patients in the camre+CAPOX followed by camre+apa group, 209 (59.9%) of 349 patients in the CAPOX group, and 82 (46.3%) of 177 patients in the camre+CAPOX followed by camre group received at least one subsequent antitumour therapy (supplementary file, supplementary table S3). To further explore the potential confounding effect of antiangiogenic therapy after observing disease progression, we conducted a post hoc analysis in which overall survival was censored at the initiation of any subsequent antiangiogenic treatment. Consistent with the primary analysis, no clear separation in overall survival was observed between the camre+CAPOX followed by camre+apa and camre+CAPOX followed by camre groups (supplementary file, supplementary figure S7).
We achieved confirmed objective responses in 57.1% (95% CI 51.7% to 62.3%) of patients in the camre+CAPOX followed by camre+apa group versus 45.3% (40.0% to 50.7%) in the CAPOX group, corresponding to an absolute difference of 11.8 percentage points (4.4% to 19.0%; table 2, table 3). Disease control rates were similar between the groups. Among responders, the median duration of response was longer with camre+CAPOX followed by camre+apa than with CAPOX (9.0 months (95% CI 7.9 to 11.2 months) v 5.5 months (4.6 to 6.8 months)).
Exploratory analyses by PD-L1 status
Across PD-L1 subgroups, hazard ratios for overall survival versus CAPOX were generally below 1.0 for camre+CAPOX followed by camre+apa and camre+CAPOX followed by camre (supplementary file, supplementary table S4), with more favourable point estimates and narrower CIs in patients with combined positive score >1 than in those with combined positive score ≤1. Treatment-by-PD-L1 interactions were explored by including an interaction term between treatment group and PD-L1 combined positive score category (>1 v ≤1) in stratified Cox models. The interaction tests (P=0.12 and P=0.08, respectively) did not provide evidence that treatment effects on overall survival differed between combined positive score >1 and combined positive score ≤1 subgroups; interaction hazard ratios with 95% confidence intervals are shown in the supplementary file, supplementary table S4. All subgroup and interaction analyses were exploratory, two sided, and not adjusted for multiple comparisons.
Safety
In the camre+CAPOX followed by camre+apa group, the median exposure durations were 5.9 months (IQR 3.4-9.9 months) for camrelizumab, 3.7 (2.6-4.0) months for capecitabine, 4.0 (2.8-4.3) months for oxaliplatin, and 3.7 (2.0-7.3) months for apatinib. In the CAPOX group, the median treatment durations were 3.9 months (IQR 2.3-4.8 months) for capecitabine and 4.1 (2.5-4.9) months for oxaliplatin. In the camre+CAPOX followed by camre group, the median exposure durations were 5.8 months (IQR 3.5-10.4 months) for camrelizumab, 3.9 (2.7-4.0) months for capecitabine, and 4.1 (2.9-4.3) months for oxaliplatin.
We reported adverse events related to treatment in 351 (99.7%) of 352 patients in the camre+CAPOX followed by camre+apa group, 341 (97.7%) of 349 patients in the CAPOX group, and 175 (98.9%) of 177 patients in the camre+CAPOX followed by camre group (see supplementary table S5; table 4). Grade 3 or higher adverse events related to treatment occurred in 239 (67.9%) of 352 participants, 158 (45.3%) of 349 participants, and 83 (46.9%) of 177 participants in the three groups, respectively. Grade 3 or higher treatment related adverse events were more frequent in the camre+CAPOX followed by camre+apa group than in the CAPOX or camre+CAPOX followed by camre groups. Compared with the CAPOX and camre+CAPOX followed by camre groups, patients in the camre+CAPOX followed by camre+apa group experienced a higher incidence of any grade and grade ≥3 hypertension and raised liver enzyme levels related to treatment (table 4; see supplementary figure S8).
Table 4.
Common treatment related adverse events with camrelizumab plus CAPOX followed by camrelizumab based maintenance versus CAPOX alone as initial treatment for gastric or gastro-oesophageal junction adenocarcinoma. Data are number (percentage) unless otherwise stated
| Treatment related adverse effects | Camre+CAPOX-camre+apa (n=352) | CAPOX (n=349) | Camre+CAPOX-camre (n=177) | |||||
|---|---|---|---|---|---|---|---|---|
| Any grade | Grade ≥3 | Any grade | Grade ≥3 | Any grade | Grade ≥3 | |||
| Any | 351 (99.7) | 239 (67.9) | 341 (97.7) | 158 (45.3) | 175 (98.9) | 83 (46.9) | ||
| RCEP | 246 (69.9) | 6 (1.7) | 0 | 0 | 133 (75.1) | 3 (1.7) | ||
| Platelet count decreased | 213 (60.5) | 77 (21.9) | 214 (61.3) | 61 (17.5) | 84 (47.5) | 17 (9.6) | ||
| Neutrophil count decreased | 211 (59.9) | 76 (21.6) | 192 (55.0) | 58 (16.6) | 105 (59.3) | 29 (16.4) | ||
| WBC decreased | 211 (59.9) | 26 (7.4) | 181 (51.9) | 21 (6.0) | 102 (57.6) | 9 (5.1) | ||
| Anaemia | 201 (57.1) | 33 (9.4) | 165 (47.3) | 26 (7.4) | 97 (54.8) | 16 (9.0) | ||
| Nausea | 185 (52.6) | 2 (0.6) | 183 (52.4) | 4 (1.1) | 75 (42.4) | 1 (0.6) | ||
| Vomiting | 180 (51.1) | 8 (2.3) | 164 (47.0) | 8 (2.3) | 79 (44.6) | 1 (0.6) | ||
| AST increased | 167 (47.4) | 32 (9.1) | 119 (34.1) | 3 (0.9) | 55 (31.1) | 2 (1.1) | ||
| Asthenia | 144 (40.9) | 6 (1.7) | 97 (27.8) | 4 (1.1) | 59 (33.3) | 1 (0.6) | ||
| ALT increased | 143 (40.6) | 24 (6.8) | 71 (20.3) | 1 (0.3) | 40 (22.6) | 1 (0.6) | ||
| Decreased appetite | 136 (38.6) | 10 (2.8) | 109 (31.2) | 6 (1.7) | 48 (27.1) | 0 | ||
| Hypertension | 97 (27.6) | 50 (14.2) | 2 (0.6) | 2 (0.6) | 7 (4.0) | 2 (1.1) | ||
| Diarrhoea | 94 (26.7) | 9 (2.6) | 62 (17.8) | 3 (0.9) | 37 (20.9) | 1 (0.6) | ||
| Blood bilirubin increased | 85 (24.1) | 19 (5.4) | 52 (14.9) | 4 (1.1) | 23 (13.0) | 1 (0.6) | ||
| Proteinuria | 82 (23.3) | 4 (1.1) | 17 (4.9) | 0 | 16 (9.0) | 0 | ||
| Hypoalbuminaemia | 74 (21.0) | 0 | 60 (17.2) | 0 | 38 (21.5) | 1 (0.6) | ||
| GGT increased | 73 (20.7) | 30 (8.5) | 26 (7.4) | 4 (1.1) | 23 (13.0) | 4 (2.3) | ||
| Blood ALP increased | 67 (19.0) | 19 (5.4) | 24 (6.9) | 0 | 21 (11.9) | 3 (1.7) | ||
| Hypothyroidism | 65 (18.5) | 0 | 4 (1.1) | 0 | 33 (18.6) | 0 | ||
| Hypoaesthesia | 62 (17.6) | 0 | 69 (19.8) | 0 | 28 (15.8) | 0 | ||
| Weight decreased | 62 (17.6) | 3 (0.9) | 40 (11.5) | 0 | 24 (13.6) | 1 (0.6) | ||
| PPE syndrome | 62 (17.6) | 13 (3.7) | 25 (7.2) | 4 (1.1) | 14 (7.9) | 4 (2.3) | ||
Treatment related adverse events reported in at least 15% of patients in any treatment group are shown.
ALP=alkaline phosphatase; ALT=alanine aminotransferase; apa=apatinib; AST=aspartate aminotransferase; camre=camrelizumab; CAPOX=capecitabine plus oxaliplatin; GGT=γ-glutamyltransferase; PPE=palmar plantar erythrodysaesthesia; RCEP=reactive capillary endothelial proliferation; WBC=white blood cell count.
Treatment related adverse events leading to discontinuation of any study drug occurred in 81 (23.0%) of 352 patients in the camre+CAPOX followed by camre+apa group, 21 (6.0%) of 349 patients in the CAPOX group, and 18 (10.2%) of 177 patients in the camre+CAPOX followed by camre group (see supplementary table S6). We reported serious treatment related adverse events in 126 (35.8%) of 352 patients, 53 (15.2%) of 349 patients, and 37 (20.9%) of 177 patients in the camre+CAPOX followed by camre+apa group, the CAPOX group, and the camre+CAPOX followed by camre group, respectively (see supplementary table S7). Deaths related to treatment occurred in 10 (2.8%) of 352 patients in the camre+CAPOX followed by camre+apa group: three (0.9%) from upper gastrointestinal haemorrhage, two (0.6%) from an unknown cause, and one (0.3%) each from gastrointestinal haemorrhage, pneumonitis, interstitial lung disease, cachexia, hepatic failure, tumour progression, and shock. Treatment related deaths occurred in three (0.9%) of 349 patients in the CAPOX group (two (0.6%) from respiratory failure, one (0.3%) from cachexia). Three (1.7%) of 177 patients in the camre+CAPOX followed by camre group died from causes related to treatment (two (1.1%) from an unknown cause and one (0.6%) from immune mediated lung disease).
Immune mediated adverse events of any grade, as assessed by investigators, occurred in 276 (78.4%) of 352 patients in the camre+CAPOX followed by camre+apa group and 145 (81.9%) of 177 patients in the camre+CAPOX followed by camre group (see supplementary table S8). Grade 3 or higher immune mediated adverse events were reported in 49 (13.9%) of 352 patients in the camre+CAPOX followed by camre+apa group and 16 (9.0%) of 177 patients in the camre+CAPOX followed by camre group. The most frequent immune mediated adverse events were reactive capillary endothelial proliferation (246 (69.9%) of 352 patients in the camre+CAPOX followed by camre+apa group and 133 (75.1%) of 177 patients in the camre+CAPOX followed by camre group) and hypothyroidism (48 (13.6%) of 352 patients and 26 (14.7%) of 177 patients, respectively).
Discussion
Principal findings
This phase 3 trial evaluated a sequential regimen consisting of an immune checkpoint inhibitor plus chemotherapy followed by an immune checkpoint inhibitor combined with a tyrosine kinase inhibitor in patients with previously untreated, HER2 negative, locally advanced or metastatic gastric or gastro-oesophageal junction adenocarcinoma. Camre+CAPOX followed by camre+apa significantly improved overall survival compared with CAPOX, both in the PD-L1 positive subgroup and in the overall population, irrespective of PD-L1 expression. In descriptive analyses, camre+CAPOX followed by camre also improved overall survival versus CAPOX. However, we observed no additional survival benefit when apatinib was added to camrelizumab maintenance, whereas adverse events related to treatment of grade ≥3 and treatment discontinuations were more frequent with camre+CAPOX followed by camre+apa. Exploratory subgroup analyses by PD-L1 status suggested that the relative benefit of camre+CAPOX followed by camre+apa and camre+CAPOX followed by camre versus CAPOX was not confined to PD-L1 positive tumours, and treatment-by-PD-L1 interaction tests did not show significant interactions.
Comparison with other studies
In the overall population, both camre+CAPOX followed by camre+apa and camre+CAPOX followed by camre reduced the risk of death by about 20% compared with CAPOX alone (median overall survival, 13.5 and 14.2 months v 12.1 months). These results are consistent with findings from the CheckMate 649, KEYNOTE-859, ORIENT-16, and RATIONALE-305 trials, in which adding nivolumab, pembrolizumab, sintilimab, or tislelizumab to chemotherapy reduced the risk of death by 20-23% (median overall survival, 13.0-15.2 months v 11.4-12.9 months).9 10 11 12 In the PD-L1 positive population, camre+CAPOX followed by camre+apa and camre+CAPOX followed by camre reduced the risk of death by 20% and 24%, respectively (median overall survival, 15.0 and 15.3 months v 12.5 months with CAPOX alone), in line with the survival benefits reported in CheckMate 649 (nivolumab plus chemotherapy: 23%, 14.0 v 11.3 months) and KEYNOTE-859 (pembrolizumab plus chemotherapy: 26%, 12.9 v 11.5 months).9 10 Together with these trials, our data reinforce the role of immune checkpoint inhibitors plus chemotherapy as standard initial treatment for HER2 negative gastric cancer and support camrelizumab as a viable anti-programmed death 1 option in this setting. As in other pivotal trials,9 10 11 12 we observed numerically greater survival benefits in patients with higher PD-L1 expression. In the camre+CAPOX followed by camre+apa group, median overall survival was 15.0, 15.8, and 16.5 months in the combined positive score >1, ≥5, and ≥10 subgroups, respectively.
Exploratory interaction analyses did not, however, identify a significant modification of treatment effect by PD-L1 status, and hazard ratios were generally <1.0 in both PD-L1 positive and PD-L1 negative subgroups. These findings are broadly consistent with previous trials showing that PD-L1 enrichment increases the magnitude and certainty of benefit but does not necessarily identify the only group that can benefit from first line programmed death 1 based chemoimmunotherapy.9 10 11 12 13 Given the smaller sample size and wider CIs in the combined positive score ≤1 subgroup in our study, we cannot exclude a smaller or absent benefit in PD-L1 negative disease, and this remains an important unresolved question.
Our study also contributes to the limited evidence on maintenance strategies after induction chemoimmunotherapy for gastric cancer. In concurrently randomised patients, overall survival and progression-free survival were similar between the camre+CAPOX followed by camre+apa and camre+CAPOX followed by camre groups, whereas the addition of apatinib increased grade ≥3 toxicity and treatment discontinuations. Most objective responses occurred during the induction phase with camre+CAPOX, and response rates were comparable between the camre+CAPOX followed by camre+apa and camre+CAPOX followed by camre groups, suggesting that intensified antiangiogenic therapy during maintenance did not translate into a clinically meaningful survival advantage. In this context, and in view of the increased toxicity burden, our findings do not support routine use of camre+CAPOX followed by camre+apa as maintenance after camre+CAPOX in unselected patients.
In terms of safety, camre+CAPOX followed by camre+apa exhibited a toxicity profile consistent with the known safety profiles of chemotherapy, camrelizumab, and apatinib. The overall incidence of adverse events was higher in the two groups that received camre than those that received CAPOX alone, likely reflecting the effect of additional drug usage. Immune mediated adverse events occurred at similar rates in the camre+CAPOX followed by camre+apa and camre+CAPOX followed by camre groups (78.4% v 81.9%), most commonly reactive capillary endothelial proliferation and thyroid dysfunction, consistent with previous reports of camrelizumab use.21 22 23 24 25 The predominant toxicities across all treatment arms were haematological, hepatic, and gastrointestinal, patterns that align with other chemotherapy based regimens. Haematological and hepatic events were readily detected through routine laboratory monitoring, enabling timely intervention. Gastrointestinal adverse events were mainly nausea and vomiting and were generally manageable, although diarrhoea warranted careful evaluation to exclude immune related colitis. However, we found that intensified maintenance with camre+CAPOX followed by camre+apa was associated with a higher frequency of clinically important toxicities. Grade 3 or higher adverse events related to treatment and serious adverse events related to treatment were more common in the camre+CAPOX followed by camre+apa group than in the CAPOX or camre+CAPOX followed by camre groups, as were treatment related adverse events that led to discontinuation of any study drug. Hypertension, proteinuria, and hepatic laboratory abnormalities occurred more frequently and at higher grades with apatinib, consistent with vascular endothelial growth factor receptor 2 inhibition and previous experience in later lines of therapy.14 15 16 17 18 Deaths resulting from adverse effects related to treatment were also numerically more frequent with camre+CAPOX followed by camre+apa than with the other regimens. Although only two of these deaths occurred during camre+apa maintenance and chance variation cannot be excluded, these findings underscore the need for careful patient selection, close monitoring, and proactive management of toxicity when regimens with multiple drugs are used. Taken together, the similar efficacy but higher toxicity observed with camre+CAPOX followed by camre+apa suggest that, for many patients, camre+CAPOX followed by camre may offer a more favourable balance of benefit and risk.
Strengths and limitations of this study
The main strengths of this study were the large sample size, the multicentre design in 75 hospitals, and the use of overall survival as the primary endpoint, which is less prone to bias than surrogate outcomes and is directly relevant to patients and clinicians. We conducted our trial in a population with a high burden of gastric cancer and our results suggest (within an East Asian cohort) that adding a programmed death 1 inhibitor to standard doublet chemotherapy improves survival compared with chemotherapy alone. In addition, the protocol specified maintenance phase with two different camre based strategies allowed us to generate comparative data on the role of adding a vascular endothelial growth factor receptor 2 inhibitor to maintenance immunotherapy, an area for which phase 3 evidence has been lacking.
Several limitations should be acknowledged. Firstly, this open label trial used multiple therapeutic agents, and tumour responses were not assessed by a blinded independent central review committee. Nevertheless, overall survival—the primary outcome—is objective and less susceptible to assessment bias. Secondly, although the primary comparison of camre+CAPOX followed by camre+apa versus CAPOX was formally powered and controlled for type 1 error, the comparison between the two camre based regimens was descriptive and not based on a prespecified hypothesis test. The study therefore lacks the statistical power and control of multiplicity needed to draw definitive conclusions about the superiority or non-inferiority of one maintenance strategy over another.
Thirdly, exploratory analyses of PD-L1 subgroups and treatment-by-PD-L1 interaction were not prespecified and were underpowered, particularly in the PD-L1 negative (combined positive score ≤1) subgroup. Although we detected no significant interaction, the CIs were wide, and we cannot reliably determine whether the magnitude of benefit differs by PD-L1 status. Next, because approval of immune checkpoint inhibitors plus chemotherapy was not yet granted at the time of study design, the trial was primarily intended to establish the superiority of immune checkpoint inhibitors plus chemotherapy over chemotherapy alone (the standard treatment at the time). To sustain treatment intensity while mitigating toxicities related to chemotherapy, we adopted a chemotherapy-free maintenance strategy with an immune checkpoint inhibitor and an antiangiogenic agent. In contemporary practice, where chemoimmunotherapy is already established as standard initial treatment, our findings are most relevant to optimising subsequent maintenance strategies rather than to re-establishing the benefit of chemoimmunotherapy over chemotherapy alone. Furthermore, we did not collect sufficient microsatellite instability/mismatch repair data to explore its association with treatment efficacy. We also did not assess quality of life outcomes; these should be incorporated in future studies to better characterise the net clinical benefit of prolonged maintenance therapy. Finally, with the emergence of targeted therapies for claudin 18.2 (CLDN18.2) the outlook for initial treatment of CLDN18.2 positive, HER2 negative disease has changed.26 The regimens investigated in this study should therefore be re-evaluated in the context of new treatment options.
Conclusions
Camre+CAPOX followed by camre based maintenance improved overall survival compared with CAPOX alone in patients with HER2 negative, unresectable, locally advanced or metastatic gastric or gastro-oesophageal junction adenocarcinoma, in both the PD-L1 positive population and the overall population. Within the context of this trial, adding apatinib to camre based maintenance did not further prolong survival compared with camre alone, but was associated with higher rates of grade ≥3 toxicity, treatment discontinuation, and treatment related death. These findings indicate that the survival gain over CAPOX is driven mainly by camrelizumab, and that routine intensification of maintenance therapy with an antiangiogenic agent may not be justified in unselected patients. Future studies should clarify the optimal timing, patient selection, and sequence of immune checkpoint inhibitor and antiangiogenic agents combinations and incorporate robust quality of life and biomarker data to refine individualised treatment strategies.
What is already known on this topic
Immunotherapy combined with chemotherapy is standard initial treatment for previously untreated, human epidermal growth factor receptor 2 (HER2) negative gastric or gastro-oesophageal junction adenocarcinoma
In immune based combination therapy, the optimal maintenance regimen, as well as the timing and composition of treatment after discontinuation of chemotherapy remain undefined
Preclinical and early clinical studies suggest potential synergy between immune checkpoint inhibition and antiangiogenic agents, but optimal integration into initial treatment is unclear
What this study adds
Camrelizumab plus capecitabine and oxaliplatin followed by camrelizumab based maintenance improved overall survival versus capecitabine and oxaliplatin alone in HER2 negative, unresectable, locally advanced or metastatic gastric or gastro-oesophageal junction adenocarcinoma
Exploratory analyses suggested that adding apatinib to camrelizumab maintenance did not further prolong survival and increased grade ≥3 treatment related adverse events and treatment discontinuations
Camrelizumab maintenance alone might improve the balance between efficacy and tolerability for many patients
Acknowledgments
We thank all patients and their families, as well as the investigators and site staff, for their participation and commitment. We thank Lin Dong and Hui Dong (PhD, medical writers at Hengrui) for medical writing support, in accordance with Good Publication Practice guidelines. We also gratefully acknowledge Chi Xu and Tao Yuan for their support with post hoc analyses, data programming, and preparation of the statistical analysis code.
Web extra.
Extra material supplied by authors
Web appendix: Supplementary file
Contributors: ZP and YZ are joint first authors. LS (linshenpku@163.com; shenlin@bjmu.edu.cn) and ZP (zhipeng@bjmu.edu.cn) are joint corresponding authors. LS is the guarantor. LS conceived and designed the study. All co-authors contributed to data acquisition, analysis, or interpretation. YW performed the statistical analyses. All authors participated in drafting or critically revising the manuscript. The corresponding author attests that all listed authors meet authorship criteria and that no others meeting the criteria have been omitted.
Funding: This study was funded by Jiangsu Hengrui Pharmaceuticals. LS designed the trial in collaboration with Jiangsu Hengrui Pharmaceuticals. The authors and the funder were involved in data collection, analysis, and interpretation, as well as in ensuring the accuracy and completeness of the data, preparation of the report, and the decision to submit the manuscript for publication. The clinical investigators were independent of the funder. All authors had access to the aggregated study data, approved the final manuscript, and take responsibility for the integrity of the data and the accuracy of the analyses.
Competing interests: All authors have completed the ICMJE uniform disclosure form at www.icmje.org/disclosure-of-interest/ and declare: support from Jiangsu Hengrui Pharmaceuticals; LS reports receiving research grants from Beigene and serving on data and safety monitoring boards or advisory boards for MSD, Boehringer Ingelheim, Servier, AstraZeneca, and Transcenta Holding, outside the submitted work; BB and YW are employees of Jiangsu Hengrui Pharmaceuticals; no financial relationships with any organisations that may have a financial interest in the submitted work in the previous three years; and no relationships or activities that could have influenced the submitted work.
Transparency: The lead author (LS) affirms that this manuscript is an honest, accurate, and transparent account of the study being reported; that no important aspects have been omitted; and that any discrepancies from the study as planned (and, if relevant, registered) have been explained.
Dissemination to participants and related patient and public communities: Participating sites were informed of the results. The results can be communicated to study participants who express an interest during clinic visits. Dissemination to the public will be carried out through media outreach.
Provenance and peer review: Not commissioned; externally peer reviewed.
Ethics statements
Ethical approval
The study protocol and all amendments were approved by the ethics committee at each participating centre (see supplementary table S1) and conducted in accordance with the Declaration of Helsinki and Good Clinical Practice guidelines.
Data availability statement
The code used to analyse the data in the paper can be found in the supplemental file. The data underlying the findings in this paper are openly and publicly available and can be accessed here: https://data.mendeley.com/datasets/z22zyc2b2b/1. If you encounter problems accessing the data, please contact the corresponding author (LS and ZP).
References
- 1. Bray F, Laversanne M, Sung H, et al. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 2024;74:229-63. 10.3322/caac.21834. [DOI] [PubMed] [Google Scholar]
- 2. Torre LA, Siegel RL, Ward EM, Jemal A. Global Cancer Incidence and Mortality Rates and Trends--An Update. Cancer Epidemiol Biomarkers Prev 2016;25:16-27. 10.1158/1055-9965.EPI-15-0578. [DOI] [PubMed] [Google Scholar]
- 3. GBD 2017 Stomach Cancer Collaborators . The global, regional, and national burden of stomach cancer in 195 countries, 1990-2017: a systematic analysis for the Global Burden of Disease study 2017. Lancet Gastroenterol Hepatol 2020;5:42-54. 10.1016/S2468-1253(19)30328-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Asplund J, Kauppila JH, Mattsson F, Lagergren J. Survival Trends in Gastric Adenocarcinoma: A Population-Based Study in Sweden. Ann Surg Oncol 2018;25:2693-702. 10.1245/s10434-018-6627-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Kim WH, Gomez-Izquierdo L, Vilardell F, et al. HER2 Status in Gastric and Gastroesophageal Junction Cancer: Results of the Large, Multinational HER-EAGLE Study. Appl Immunohistochem Mol Morphol 2018;26:239-45. 10.1097/PAI.0000000000000423. [DOI] [PubMed] [Google Scholar]
- 6. Tanner M, Hollmén M, Junttila TT, et al. Amplification of HER-2 in gastric carcinoma: association with Topoisomerase IIalpha gene amplification, intestinal type, poor prognosis and sensitivity to trastuzumab. Ann Oncol 2005;16:273-8. 10.1093/annonc/mdi064. [DOI] [PubMed] [Google Scholar]
- 7. Yan B, Yau EX, Bte Omar SS, et al. A study of HER2 gene amplification and protein expression in gastric cancer. J Clin Pathol 2010;63:839-42. 10.1136/jcp.2010.076570. [DOI] [PubMed] [Google Scholar]
- 8. Lordick F, Carneiro F, Cascinu S, et al. ESMO Guidelines Committee. Electronic address: clinicalguidelines@esmo.org . Gastric cancer: ESMO Clinical Practice Guideline for diagnosis, treatment and follow-up. Ann Oncol 2022;33:1005-20. 10.1016/j.annonc.2022.07.004. [DOI] [PubMed] [Google Scholar]
- 9. Janjigian YY, Shitara K, Moehler M, et al. First-line nivolumab plus chemotherapy versus chemotherapy alone for advanced gastric, gastro-oesophageal junction, and oesophageal adenocarcinoma (CheckMate 649): a randomised, open-label, phase 3 trial. Lancet 2021;398:27-40. 10.1016/S0140-6736(21)00797-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Rha SY, Oh DY, Yañez P, et al. KEYNOTE-859 investigators . Pembrolizumab plus chemotherapy versus placebo plus chemotherapy for HER2-negative advanced gastric cancer (KEYNOTE-859): a multicentre, randomised, double-blind, phase 3 trial. Lancet Oncol 2023;24:1181-95. 10.1016/S1470-2045(23)00515-6. [DOI] [PubMed] [Google Scholar]
- 11. Xu J, Jiang H, Pan Y, et al. ORIENT-16 Investigators . Sintilimab Plus Chemotherapy for Unresectable Gastric or Gastroesophageal Junction Cancer: The ORIENT-16 Randomized Clinical Trial. JAMA 2023;330:2064-74. 10.1001/jama.2023.19918. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Qiu MZ, Oh DY, Kato K, et al. RATIONALE-305 Investigators . Tislelizumab plus chemotherapy versus placebo plus chemotherapy as first line treatment for advanced gastric or gastro-oesophageal junction adenocarcinoma: RATIONALE-305 randomised, double blind, phase 3 trial. BMJ 2024;385:e078876. 10.1136/bmj-2023-078876. [DOI] [PubMed] [Google Scholar]
- 13. Kang YK, Chen LT, Ryu MH, et al. Nivolumab plus chemotherapy versus placebo plus chemotherapy in patients with HER2-negative, untreated, unresectable advanced or recurrent gastric or gastro-oesophageal junction cancer (ATTRACTION-4): a randomised, multicentre, double-blind, placebo-controlled, phase 3 trial. Lancet Oncol 2022;23:234-47. 10.1016/S1470-2045(21)00692-6. [DOI] [PubMed] [Google Scholar]
- 14. Li J, Qin S, Xu J, et al. Randomized, Double-Blind, Placebo-Controlled Phase III Trial of apatinib in Patients With Chemotherapy-Refractory Advanced or Metastatic Adenocarcinoma of the Stomach or Gastroesophageal Junction. J Clin Oncol 2016;34:1448-54. 10.1200/JCO.2015.63.5995. [DOI] [PubMed] [Google Scholar]
- 15. Li J, Qin S, Xu J, et al. apatinib for chemotherapy-refractory advanced metastatic gastric cancer: results from a randomized, placebo-controlled, parallel-arm, phase II trial J Clin Oncol 2013;31:3219-25. 10.1200/JCO.2013.48.8585. [DOI] [PubMed] [Google Scholar]
- 16. Scott LJ. apatinib: A Review in Advanced Gastric Cancer and Other Advanced Cancers. Drugs 2018;78:747-58. 10.1007/s40265-018-0903-9. [DOI] [PubMed] [Google Scholar]
- 17. Luo Q, Dong Z, Xie W, et al. apatinib remodels the immunosuppressive tumor ecosystem of gastric cancer enhancing anti-PD 1 immunotherapy. Cell Rep 2023;42:112437. 10.1016/j.celrep.2023.112437. [DOI] [PubMed] [Google Scholar]
- 18. Xu J, Zhang Y, Jia R, et al. Anti-PD 1 Antibody SHR-1210 Combined with apatinib for Advanced Hepatocellular Carcinoma, Gastric, or Esophagogastric Junction Cancer: An Open-label, Dose Escalation and Expansion Study [published Online First: 20181022]. Clin Cancer Res 2019;25:515-23. 10.1158/1078-0432.CCR-18-2484. [DOI] [PubMed] [Google Scholar]
- 19. Peng Z, Wei J, Wang F, et al. camrelizumab Combined with Chemotherapy Followed by camrelizumab plus apatinib as First-line Therapy for Advanced Gastric or Gastroesophageal Junction Adenocarcinoma. Clin Cancer Res 2021;27:3069-78. 10.1158/1078-0432.CCR-20-4691. [DOI] [PubMed] [Google Scholar]
- 20. Shitara K, Van Cutsem E, Bang YJ, et al. Efficacy and Safety of Pembrolizumab or Pembrolizumab Plus Chemotherapy vs Chemotherapy Alone for Patients With First-line, Advanced Gastric Cancer: The KEYNOTE-062 Phase 3 Randomized Clinical Trial. JAMA Oncol 2020;6:1571-80. 10.1001/jamaoncol.2020.3370. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Qin S, Chan SL, Gu S, et al. CARES-310 Study Group . camrelizumab plus rivoceranib versus sorafenib as first-line therapy for unresectable hepatocellular carcinoma (CARES-310): a randomised, open-label, international phase 3 study Lancet 2023;402:1133-46. 10.1016/S0140-6736(23)00961-3. [DOI] [PubMed] [Google Scholar]
- 22. Luo H, Lu J, Bai Y, et al. ESCORT-1st Investigators . Effect of camrelizumab vs Placebo Added to Chemotherapy on Survival and Progression-Free Survival in Patients With Advanced or Metastatic Esophageal Squamous Cell Carcinoma: The ESCORT-1st Randomized Clinical Trial. JAMA 2021;326:916-25. 10.1001/jama.2021.12836. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Ren S, Chen J, Xu X, et al. CameL-sq Study Group . camrelizumab Plus Carboplatin and Paclitaxel as First-Line Treatment for Advanced Squamous NSCLC (CameL-Sq): A Phase 3 Trial J Thorac Oncol 2022;17:544-57. 10.1016/j.jtho.2021.11.018. [DOI] [PubMed] [Google Scholar]
- 24. Zhou C, Chen G, Huang Y, et al. CameL Study Group . camrelizumab plus carboplatin and pemetrexed versus chemotherapy alone in chemotherapy-naive patients with advanced non-squamous non-small-cell lung cancer (CameL): a randomised, open-label, multicentre, phase 3 trial Lancet Respir Med 2021;9:305-14. 10.1016/S2213-2600(20)30365-9. [DOI] [PubMed] [Google Scholar]
- 25. Huang J, Xu J, Chen Y, et al. ESCORT Study Group . camrelizumab versus investigator’s choice of chemotherapy as second-line therapy for advanced or metastatic oesophageal squamous cell carcinoma (ESCORT): a multicentre, randomised, open-label, phase 3 study Lancet Oncol 2020;21:832-42. 10.1016/S1470-2045(20)30110-8. [DOI] [PubMed] [Google Scholar]
- 26. Shah MA, Shitara K, Ajani JA, et al. Zolbetuximab plus CAPOX in CLDN18.2-positive gastric or gastroesophageal junction adenocarcinoma: the randomized, phase 3 GLOW trial. Nat Med 2023;29:2133-41. 10.1038/s41591-023-02465-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Web appendix: Supplementary file
Data Availability Statement
The code used to analyse the data in the paper can be found in the supplemental file. The data underlying the findings in this paper are openly and publicly available and can be accessed here: https://data.mendeley.com/datasets/z22zyc2b2b/1. If you encounter problems accessing the data, please contact the corresponding author (LS and ZP).
