Alpha¹ by KaimenConfidential

Kaimen Rigor review

Camrelizumab plus CAPOX with camrelizumab based maintenance versus CAPOX alone as initial treatment for gastric or gastro-oesophageal junction adenocarcinoma: randomised phase 3 trial.

Peng Z, Zhang Y, Xu H, Yang Y, Yang M, Zhang M, Cheng Y, Chen X, Pan Y, Wang F, Li Q, Xu N, Liu L, Gu K, Xiao J, Zhang R, Zhao Q, Chen J, Lin L, Chen Y, Deng Y, Cai S, Bai B, Wang Y, Shen L

DOI: 10.1136/bmj-2025-086115
Record issued: 2026-08-10
Engine: 7.29.0
Exported: 2026-09-20

Prepared by Alpha1. This document is confidential: it is intended for the recipient it was shared with and must not be redistributed. The live record at alpha1science.com/verify/0233de75-951b-4c1b-8965-d5036c81256a is authoritative.

Kaimen Rigor review

Kaimen Rigor reviewed

3.25 / 5

Adequate — real issues to address

Interventional Study

Critical

High-priority

How this rating was calculated

Started at 5★ — deductions

IntegrityIntegrity concern ×3−1.5★
ReportingData & code availability partially met−0.25★

Model-reported confidence: moderateReviewer agreement: moderateEngine: 7.29.0Source: full text

Not verified in this report

The numeric-impossibility checks (GRIM/GRIMMER/DEBIT/SPRITE) did not run on this paper: the pass that reads its reported means did not complete. No reported mean was checked for arithmetic impossibility.

Findings by category

101ArithmeticNumerical inconsistencies 102OverreachOverstated conclusions 103AuthenticityData authenticity concerns 104DisclosureReporting gaps 05AvailabilityNone found 106CopyeditCopyediting

This Kaimen Rigor review uses Kaimen Rigor reviewers trained on a curated corpus of high-fidelity and retracted papers, with expert supervision and curation. It can still make mistakes; verify each finding against the source before relying on it.

This is a well-conducted and largely well-reported phase 3 RCT with a sound premise, rigorous randomization and power planning, proper ethics and registration, exact primary p-values, and spot-check-consistent statistics. The main weaknesses are substantive reporting gaps (analysis code not in a version-controlled repository; incomplete biological-variable/demographic reporting) and several copyedit-level inconsistencies, including a treatment-related death count that does not add up and a mismatched data-repository DOI.

Three independent reviewer runs of the same model were synthesized with a copyedit pass and verification components. The statistics component recomputed only 5 reported tests (all consistent); remaining statistical claims are unverified, not confirmed. Bench-science sub-criteria were not applicable to this human drug trial. Reviewers diverged on biological variables (warn vs pass) and key resources (warn vs pass), resolved here by weighing the specific evidence.

Numerical inconsistencies

1 finding · worst low

Values that contradict each other or are impossible for the stated sample: recomputed p-values and test statistics, GRIM/GRIMMER checks on summary numbers, percentages against their own counts, totals against their parts, and estimates against their own confidence intervals.

Internal contradictions in the reported numbersAssessed

Recomputed 5 tests: 5 consistent, 0 inconsistent; 5 via agent-written checks.

Agent-written checks— recomputed from reported CIs / ratios by code the agent wrote

CONSISTENTreported p = .020 · recomputed p = .033Reviewer 1
Recompute two-tailed p from hazard ratio and 95% CI for primary analysis in PD-L1 positive population (HR 0.80, 95% CI 0.65-0.98).
“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)”
Find in source
Taken as given: The hazard ratio is the point estimate.; The 95% CI is for the hazard ratio (log-transformed).; The p-value reported is one-sided, so the two-tailed p from the CI should be divided by 2 for comparison.
Method: pCI function from the sandbox for a ratio estimate (log=1) returns a two-tailed p. The reported one-sided p is approximately half of the two-tailed p.
How we recomputed it: pCI(0.80, 0.65, 0.98, 1)
CONSISTENTreported p = .004 · recomputed p = .007Reviewer 1
Recompute two-tailed p from hazard ratio and 95% CI for primary analysis in overall population (HR 0.80, 95% CI 0.68-0.94).
“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)”
Find in source
Taken as given: The hazard ratio is the point estimate.; The 95% CI is for the hazard ratio (log-transformed).; The p-value reported is one-sided, so the two-tailed p from the CI should be divided by 2 for comparison.
Method: pCI function for a ratio estimate (log=1) returns a two-tailed p. The reported one-sided p is approximately half of the two-tailed p.
How we recomputed it: pCI(0.80, 0.68, 0.94, 1)
CONSISTENTreported p = .037 · recomputed p = .033Reviewer 2
Two-tailed p derived from reported HR 0.80 and 95% CI 0.65-0.98 for OS (camre+apa vs CAPOX, PD-L1 positive), compared with reported one-sided P=0.02.
“hazard ratio 0.80 (95% CI 0.65 to 0.98); one sided P=0.02”
Find in source
Taken as given: The 0.80, 0.65, and 0.98 are the HR and its two-sided 95% CI bounds for the same comparison; The paper reports a one-sided P; the two-sided p from the CI should be about twice the reported value; The CI is log-symmetric (log=1 for a ratio)
Method: Recompute two-tailed p from the point estimate and 95% CI via pCI on the log scale; compare to 2× the reported one-sided p.
How we recomputed it: pCI(0.80, 0.65, 0.98, 1)
CONSISTENTreported p = .007 · recomputed p = .007Reviewer 2
Two-tailed p derived from reported HR 0.80 and 95% CI 0.68-0.94 for OS (camre+apa vs CAPOX, overall population), compared with reported one-sided P=0.004.
“hazard ratio 0.80 (0.68 to 0.94); one sided P=0.004”
Find in source
Taken as given: The 0.80, 0.68, and 0.94 are the HR and its two-sided 95% CI bounds for the same comparison; The paper reports a one-sided P; the two-sided p from the CI should be about twice the reported value; The CI is log-symmetric (log=1 for a ratio)
Method: Recompute two-tailed p from the point estimate and 95% CI via pCI on the log scale; compare to 2× the reported one-sided p.
How we recomputed it: pCI(0.80, 0.68, 0.94, 1)
CONSISTENTreported p = .032 · recomputed p = .036Reviewer 2
Two-tailed p derived from reported HR 0.76 and 95% CI 0.58-0.97 for OS (camre vs CAPOX, PD-L1 positive, nominal), compared with reported one-sided P=0.01.
“hazard ratio 0.76 (0.58 to 0.97); one sided nominal P=0.01”
Find in source
Taken as given: The 0.76, 0.58, and 0.97 are the HR and its two-sided 95% CI bounds for the same comparison; The paper reports a one-sided nominal P; the two-sided p from the CI should be about twice the reported value; The CI is log-symmetric (log=1 for a ratio)
Method: Recompute two-tailed p from the point estimate and 95% CI via pCI on the log scale; compare to 2× the reported one-sided p.
How we recomputed it: pCI(0.76, 0.58, 0.97, 1)

lowinternal contradiction
The abstract reports '454 of 592 (76.7%) deaths had occurred in the PD-L1 positive population' and '709 of 878 (80.8%) in the overall population'; these proportions are internally consistent with the per-group denominators, so no error is apparent, but the PD-L1 positive n (592) versus the sum of per-group PD-L1>1 counts (227+238+127=592) is consistent.
“454 of 592 (76.7%) deaths had occurred in the PD-L1 positive population and 709 of 878 (80.8%) in the overall population”
AbstractFind in source
lowinternal contradiction
The enumerated causes of treatment-related death in the camre+apa group sum to 12 (3 + 2 + 7 individual causes), but the stated total is 10 deaths.
“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.”
ResultsFind in source

Overstated conclusions

1 finding · worst low

Conclusions that reach past what the paper's own results support — including a significance claim that no longer holds when the statistic is recomputed, and efficacy resting on an unvalidated surrogate endpoint.

Conclusions only partially backed by the presented evidenceAssessed

12 major claims checked against the paper's own evidence: all adequately supported.

partialReviewers 1, 2, 3
The survival gain over CAPOX is driven mainly by camrelizumab.
The claim follows from the lack of incremental benefit of apatinib and the camre-alone benefit, but it is an inference from descriptive comparisons rather than a directly tested hypothesis; the paper frames it as an interpretation.
Evidence: No OS difference between the two camre-based groups combined with OS benefit of camre alone versus CAPOX.
“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.”
ConclusionFind in source
supportedReviewer 1
Camrelizumab plus CAPOX followed by camrelizumab plus apatinib significantly improved overall survival compared with CAPOX alone in the PD-L1 positive population and overall population.
The primary analysis provides direct evidence with hazard ratios and p-values meeting the pre-specified alpha.
Evidence: HR 0.80 (95% CI 0.65-0.98; one-sided P=0.02) in PD-L1 positive; HR 0.80 (95% CI 0.68-0.94; one-sided P=0.004) in 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).”
AbstractFind in source
supportedReviewer 1
Camrelizumab plus CAPOX followed by camrelizumab also improved overall survival versus CAPOX.
Exploratory analyses show consistent hazard ratios with nominal p-values, supporting the claim.
Evidence: HR 0.76 (95% CI 0.58-0.97; one-sided nominal P=0.01) in PD-L1 positive; HR 0.80 (95% CI 0.65-0.98; one-sided nominal P=0.02) in overall population.
“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).”
AbstractFind in source
supportedReviewer 1
No additional survival benefit was observed when apatinib was added to camrelizumab maintenance.
Direct comparison between the two camrelizumab-based regimens shows no significant difference.
Evidence: HR 1.08 (95% CI 0.83-1.40; one-sided nominal P=0.29) in PD-L1 positive; HR 1.02 (95% CI 0.82-1.26; one-sided nominal P=0.45) in overall population.
“No overall survival benefit was observed with camre+CAPOX followed by camre+apa versus camre+CAPOX followed by camre.”
AbstractFind in source
supportedReviewer 1
Higher rates of treatment related adverse events of grade ≥3 and treatment discontinuations when apatinib was added during maintenance.
Safety data clearly show higher toxicity in the camre+CAPOX followed by camre+apa group.
Evidence: Grade ≥3 treatment-related adverse events: 67.9% in camre+CAPOX-camre+apa vs 45.3% in CAPOX and 46.9% in camre+CAPOX-camre; treatment discontinuations: 23.0% vs 6.0% and 10.2%.
“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.”
AbstractFind in source
supportedReviewer 2
Camrelizumab+CAPOX followed by camrelizumab+apatinib improves overall survival versus CAPOX alone in the PD-L1 positive and overall populations.
The primary endpoint is OS with HR 0.80 (95% CI 0.65-0.98, one-sided P=0.02) in PD-L1 positive and HR 0.80 (0.68-0.94, P=0.004) in the overall population; the evidence directly backs the claim.
Evidence: OS HR 0.80 (0.65-0.98) in PD-L1 positive and 0.80 (0.68-0.94) in overall population, with reported one-sided p-values.
“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).”
AbstractFind in source
supportedReviewer 2
Camrelizumab+CAPOX followed by camrelizumab alone also improves OS versus CAPOX.
The claim is backed by descriptive analyses with HR 0.76 (0.58-0.97) and 0.80 (0.65-0.98); the paper correctly labels these as exploratory with nominal p-values, so the claim is appropriately hedged.
Evidence: Nominal OS HRs 0.76 (0.58-0.97) and 0.80 (0.65-0.98) versus CAPOX in PD-L1 positive and overall populations.
“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)”
AbstractFind in source
supportedReviewer 2
Adding apatinib to camrelizumab maintenance provides no additional survival benefit and increases toxicity.
Descriptive comparisons show no OS benefit (HR 1.08, 0.83-1.40 in PD-L1 positive; 1.02, 0.82-1.26 overall) with higher grade ≥3 AEs (67.9% vs 46.9%); the paper explicitly states these are exploratory and not powered for superiority/non-inferiority.
Evidence: OS HR 1.08 (0.83-1.40) and 1.02 (0.82-1.26); grade ≥3 treatment-related AEs 67.9% vs 46.9%.
“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.”
DiscussionFind in source
supportedReviewer 3
Initial treatment with camrelizumab plus CAPOX followed by camrelizumab-based maintenance is associated with longer overall survival than CAPOX alone.
The presented HRs (0.80, 95% CI 0.65-0.98, P=0.02 in PD-L1+; 0.80, 0.68-0.94, P=0.004 overall) directly support the claim.
Evidence: Primary endpoint results in abstract and Results section, with HRs and CIs.
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...
Conclusionreviewer’s wording
supportedReviewer 3
Camre+CAPOX followed by camre (camrelizumab maintenance alone) also improved overall survival versus CAPOX.
The claim is supported as a descriptive analysis, appropriately labelled nominal/exploratory without formal hypothesis testing.
Evidence: HR 0.76 (0.58-0.97) in PD-L1+ and 0.80 (0.65-0.98) overall, reported as nominal one-sided p-values.
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).
Resultsreviewer’s wording
supportedReviewer 3
Adding apatinib to camrelizumab maintenance provided no additional survival benefit.
The exploratory HRs (1.08, 0.83-1.40 PD-L1+; 1.02, 0.82-1.26 overall) directly support the absence of benefit, appropriately framed as descriptive.
Evidence: Exploratory OS comparisons of camre+apa vs camre in both populations.
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...)
Resultsreviewer’s wording
supportedReviewer 3
Apatinib addition increased grade ≥3 treatment-related adverse events and treatment discontinuations.
Safety data directly support the claim: grade ≥3 events 67.9% vs 46.9% and discontinuations 23.0% vs 10.2%.
Evidence: Safety results and Table 4.
“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.”
AbstractFind in source

Efficacy claim is anchored to an adequate endpoint and a meaningful effect.

ADEQUATESurrogate endpoint
Primary endpoint is overall survival (mortality), a hard clinical outcome.
“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.”
Find in source
ADEQUATEEffect size
Hazard ratio 0.80 (95% CI 0.68 to 0.94) in overall population, median overall survival improved from 12.1 to 13.5 months; statistically significant and consistent with other trials establishing clinical benefit.
“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)”
Find in source

Data authenticity concerns

1 finding · worst low

An adversarial read for patterns associated with data that may not be genuine: results that look too clean, implausibly large effects, duplicated data or images, and methods that do not match the results reported.

Other integrity concernAssessed

3 integrity concerns flagged (0 high).

lowother
The DOI printed next to the Mendeley data URL (10.3322/caac.21834) is a CA: A Cancer Journal for Clinicians DOI and does not appear to correspond to this dataset; this is a citation/identifier mismatch rather than a data-integrity issue.
“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 (10.3322/caac.21834)”
Associated Data sectionFind in source

Reporting gaps

1 finding · worst medium

Required detail the manuscript never states — study design, biological variables, ethics approval and consent, key resources, statistical reporting, data and code availability, and overall transparency.

Data/code availability incompleteAssessed

Eight Dimensions7 pass · 1 warn

The introduction cites prior research on gastric cancer epidemiology, standard treatment, and previous trials (e.g., KEYNOTE-062, CheckMate 649). It discusses the rationale for adding apatinib based on preclinical and phase 1 data. The study builds on a phase 2 trial to address the lack of phase 3 data on maintenance strategies.

Sub-criteria

prior work citedADEQUATE

premise rationaleADEQUATE

limitations addressedADEQUATE

Per reviewer (3) — the votes behind the verdict

PASSReviewer 1· 95% conf

The scientific premise is well-established with prior work cited, a logical rationale, and limitations of previous research addressed.

Evidence

“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.”

Introduction ¶3Find in source

“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.”

Introduction ¶4Find in source

paraphrase[Introduction, paragraph 5]

“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%... 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.”

PASSReviewer 2· 90% conf

The introduction thoroughly establishes the clinical context, cites prior phase 1/phase 2 evidence, and derives a clear hypothesis that anti-PD-1 plus antiangiogenic maintenance may extend the benefit of chemoimmunotherapy.

Evidence

“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.”

Introduction ¶2Find in source

“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”

Introduction ¶4Find in source

paraphrase[Introduction, paragraph 3]

“Paraphrase: The introduction notes that antiangiogenic maintenance beyond immunotherapy alone is of unknown benefit given potential added toxicity.”

PASSReviewer 3· 90% conf

The premise is well established with extensive prior literature, a clear rationale, and explicit acknowledgement of prior knowledge gaps.

Evidence

paraphrase[Introduction, paragraph 4-5]

“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%...”

Introduction ¶3Find in source

paraphrase[Introduction, paragraph 4]

“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...”

The trial uses block randomization (block size of six) via a centralized interactive web response system. The sample size calculation is detailed with power assumptions. Inclusion and exclusion criteria are pre-specified. The analysis population is defined. The open-label design is disclosed but a rationale for the lack of blinding is not provided.

Sub-criteria

controlsN/A

power analysisADEQUATE

blinding levelsINADEQUATE

outlier handlingADEQUATE

randomization unitADEQUATE

inclusion exclusionADEQUATE

randomization methodADEQUATE

replicate distinctionN/A

independent replicationN/A

Per reviewer (3) — the votes behind the verdict

PASSReviewer 1· 90% conf

The study design is rigorous with adequate randomization, sample size planning, and inclusion/exclusion criteria, though the open-label design is stated without a rationale.

Evidence

paraphrase[Methods, Randomisation and masking]

“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, presence of peritoneal metastasis, and PD-L1 combined positive score.”

paraphrase[Methods, Statistical analysis]

“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.”

“This was an open label study: neither investigators nor patients were masked to treatment allocation.”

MethodsFind in source

PASSReviewer 2· 90% conf

As a human RCT, the applicable design elements (randomization method/unit, blinding, power, inclusion/exclusion, analysis population) are all reported adequately; open-label design is stated with rationale.

Evidence

“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).”

MethodsFind in source

“This was an open label study: neither investigators nor patients were masked to treatment allocation.”

MethodsFind in source

“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”

Statistical analysisFind in source

PASSReviewer 3· 90% conf

Core design elements for a human RCT are well reported: randomization method/unit, open-label status with rationale, detailed power calculations, and pre-specified inclusion/exclusion criteria.

Evidence

“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).”

MethodsFind in source

“This was an open label study: neither investigators nor patients were masked to treatment allocation.”

MethodsFind in source

paraphrase[Methods, Statistical analysis]

“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...”

Table 1 reports sex (male/female percentages), median age with IQR, and ECOG performance status across all three arms. Health status is captured via ECOG PS and metastatic burden. Race/ethnicity and comorbid conditions are not reported, so the demographics criterion is only partially met. Sex justification is not applicable because both sexes were enrolled.

Sub-criteria

demographicsINADEQUATE

sex reportedADEQUATE

sex justifiedN/A

age weight healthADEQUATE

housing conditionsN/A

species strain sourceN/A

Per reviewer (3) — the votes behind the verdict

WARNReviewer 1· 85% conf

Sex and age are reported, but weight, race/ethnicity, and comorbidities are not adequately reported, leading to incomplete biological variable characterization.

Evidence

paraphrase[Table 1]

“Median (IQR) age (years) 62 (55-67) | Sex: Male 250 (71.0), Female 102 (29.0)”

absence[Table 1, Methods]

Weight is not reported.

absence[Table 1, Methods]

Race/ethnicity and comorbidities are not explicitly reported.

PASSReviewer 2· 70% conf

Sex, age, and health status (ECOG) are reported in Table 1; the study enrolled both sexes so a single-sex justification is not required. Race/ethnicity and comorbidities are not reported, so demographics is rated inadequate but the other applicable items are adequate.

Evidence

“Median (IQR) age (years) | 62 (55-67) | 62 (54-68) | 60 (54-68) | | Sex: | | Male | 250 (71.0) | 272 (77.9) | 128 (72.3)”

Table 1Find in source

paraphrase[Table 1]

“ECOG performance status: | 0 | 105 (29.8) | 104 (29.8) | 51 (28.8)”

absence[Table 1]

Race/ethnicity and comorbidity data are not reported.

PASSReviewer 3· 85% conf

Sex, age, and health status are reported adequately; demographics are partially reported (age and sex present, but race/ethnicity and comorbidities not detailed).

Evidence

paraphrase[Table 1]

“Median (IQR) age (years) | 62 (55-67) | 62 (54-68) | 60 (54-68) | ... Male | 250 (71.0) | 272 (77.9) | 128 (72.3)”

paraphrase[Table 1]

“ECOG performance status: 0 | 105 (29.8) | 104 (29.8) | 51 (28.8)”

absence[Table 1]

Race/ethnicity and comorbidities are not reported.

The paper states that the protocol was approved by the ethics committee at each participating centre, written informed consent was obtained from all participants, and the study was conducted in accordance with the Declaration of Helsinki and Good Clinical Practice guidelines.

Sub-criteria

iacuc statementN/A

informed consentADEQUATE

irb ethics statementADEQUATE

regulatory complianceADEQUATE

Per reviewer (3) — the votes behind the verdict

PASSReviewer 1· 95% conf

Ethical approvals, informed consent, and regulatory compliance are adequately reported.

Evidence

“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.”

EthicsFind in source

“Written informed consent was obtained from all participants before enrolment.”

MethodsFind in source

“conducted in accordance with the Declaration of Helsinki and Good Clinical Practice guidelines.”

EthicsFind in source

PASSReviewer 2· 90% conf

Ethics approval by named committees at each centre, written informed consent, and regulatory compliance (Declaration of Helsinki, GCP) are all stated.

Evidence

EthicsFind in source

“Written informed consent was obtained from all participants before enrolment.”

MethodsFind in source

PASSReviewer 3· 90% conf

Ethics approval by named ethics committees at each centre, written informed consent, and regulatory compliance (Declaration of Helsinki and GCP) are all explicitly reported.

Evidence

EthicsFind in source

“Written informed consent was obtained from all participants before enrolment.”

MethodsFind in source

As a drug trial, reagents_identified is scored against the investigational products: all four drugs are named with doses, routes, and schedules (e.g., camrelizumab 200 mg IV q3w, capecitabine 1000 mg/m2 BID days 1-14, oxaliplatin 130 mg/m2 day 1, apatinib 250 mg daily). Manufacturer is implied by the funder (Jiangsu Hengrui Pharmaceuticals), though explicit vendor/manufacturer identifiers for each agent are not itemised. Software (SAS 9.4) is identified. Antibodies, cell lines, mycoplasma, and organisms are not applicable. 2 of 2 applicable criteria adequate.

Sub-criteria

mycoplasma testingN/A

reagents identifiedADEQUATE

organisms identifiedN/A

antibodies identifiedN/A

cell line authenticationN/A

software tools identifiedADEQUATE

Per reviewer (3) — the votes behind the verdict

WARNReviewer 1· 80% conf

The investigational drugs are identified with dose and regimen, but the manufacturer is not explicitly stated, and the code is not in a public repository, leading to a warn.

Evidence

paraphrase[Methods, Procedures]

“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).”

“SAS software (version 9.4) was used for all statistical analyses.”

Statistical analysisFind in source

absence[Methods, Procedures]

Manufacturer/source of drugs not explicitly stated.

PASSReviewer 2· 85% conf

The investigational products (camrelizumab, capecitabine, oxaliplatin, apatinib) are named with dose and regimen, and the statistical software (SAS 9.4) is identified; bench-science resource criteria are not applicable to this drug trial.

Evidence

“Patients assigned to the camre+CAPOX followed by camre+apa group received camrelizumab (200 mg intravenously every three weeks) plus CAPOX (capecitabine 1000 mg/m 2 orally twice daily on days 1-14 of each 21 day treatment cycle, and oxaliplatin 130 mg/m 2 intravenously on day 1 of each cycle)”

MethodsFind in source

“SAS software (version 9.4) was used for all statistical analyses.”

Statistical analysisFind in source

PASSReviewer 3· 85% conf

Investigational products (camrelizumab, capecitabine, oxaliplatin, apatinib) are identified with doses and regimens, and the statistical software is identified with version.

Evidence

paraphrase[Methods, 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)...”

“SAS software (version 9.4) was used for all statistical analyses.”

Statistical analysisFind in source

“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.”

MethodsFind in source

Tests include log-rank, Cox proportional hazards, Clopper-Pearson, and Newcombe methods. Exact one-sided p-values and hazard ratios with 95% CIs are provided. SAS 9.4 is used. Kaplan-Meier curves, forest plots, and tables with per-group n and CIs are presented. Assumptions are handled by design (stratified Cox models). The mathematical plausibility sub-criterion is not applicable due to large N and continuous outcomes.

Sub-criteria

tests namedADEQUATE

exact p valuesADEQUATE

effect sizes ciADEQUATE

data presentationADEQUATE

software identifiedADEQUATE

assumptions verifiedADEQUATE

mathematical plausibilityN/A

Per reviewer (3) — the votes behind the verdict

PASSReviewer 1· 90% conf

Statistical tests are named, assumptions are addressed by design, exact p-values and effect sizes with CIs are reported, software is identified, and data presentation is appropriate for a large clinical trial.

Evidence

paraphrase[Methods, Statistical analysis]

“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.”

paraphrase[Results, Efficacy in PD-L1 positive patients]

“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)”

paraphrase[Table 2]

“Treatment v CAPOX (hazard ratio (95% CI), P value) 0.80 (0.65 to 0.98), 0.02”

PASSReviewer 2· 85% conf

Statistical tests are named, effect sizes are reported with 95% CIs throughout, software is identified, data are presented per clinical-trial idiom (KM curves, forest plots, per-group n), and reported proportions are arithmetically plausible.

Evidence

paraphrase[Methods, Statistical analysis]

“We compared overall survival and progression-free survival with the log rank test, stratified by ... and used stratified Cox proportional hazards models to estimate hazard ratios with 95% CIs”

“hazard ratio 0.80 (95% CI 0.65 to 0.98); one sided P=0.02”

AbstractFind in source

“Grade 3 or higher adverse events related to treatment occurred in 239 (67.9%) of 352 participants”

ResultsFind in source

PASSReviewer 3· 85% conf

Statistical methods are named and appropriate for the clinical idiom, primary p-values are exact, effect sizes are reported with 95% CIs, and software is identified.

Evidence

“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)”

AbstractFind in source

Statistical analysisFind in source

“454 of 592 (76.7%) deaths had occurred in the PD-L1 positive population and 709 of 878 (80.8%) in the overall population.”

AbstractFind in source

The data are openly available at Mendeley with a DOI. The code is available in the supplemental file, but not in a public repository with a persistent identifier, which is the standard for reproducibility.

Sub-criteria

code sharingINADEQUATE

accession numbersN/A

repository depositADEQUATE

data availability statementADEQUATE

Per reviewer (3) — the votes behind the verdict

WARNReviewer 1· 85% conf

A data availability statement and a public repository (Mendeley) are provided, but the analysis code is only in the supplemental file, not in a version-controlled public repository.

Evidence

“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”

Data availabilityFind in source

“The code used to analyse the data in the paper can be found in the supplemental file.”

Data availabilityFind in source

absence[Data availability statement]

Code not in a public repository with persistent identifier.

WARNReviewer 2· 80% conf

A data availability statement with a concrete public repository (Mendeley URL) is present and the data are openly available, but the analysis code is only placed in the supplemental file rather than a version-controlled public repository.

Evidence

“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 .”

Data availabilityFind in source

“The code used to analyse the data in the paper can be found in the supplemental file.”

Data availabilityFind in source

WARNReviewer 3· 80% conf

Data are openly deposited in a public repository with a URL, but the analysis code is only in the supplemental file rather than a version-controlled public repository with a permanent identifier.

Evidence

“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”

Data availabilityFind in source

“The code used to analyse the data in the paper can be found in the supplemental file.”

Data availabilityFind in source

Methods are detailed enough for replication. The trial is registered at ClinicalTrials.gov (NCT03813784). All pre-specified outcomes are reported, including negative results. Limitations are discussed in a dedicated section. Conclusions are cautious and proportional. Funding and competing interests are transparent. A reporting guideline is not explicitly referenced.

Sub-criteria

funding coiADEQUATE

trial registrationADEQUATE

reporting guidelineINADEQUATE

methods completenessADEQUATE

all outcomes reportedADEQUATE

limitations discussedADEQUATE

conclusions proportionalADEQUATE

Per reviewer (3) — the votes behind the verdict

PASSReviewer 1· 90% conf

The manuscript reports methods comprehensively, includes trial registration, reports all outcomes, discusses limitations, draws proportional conclusions, and declares funding and conflicts. The only gap is the absence of an explicit reference to a reporting guideline (e.g., CONSORT).

Evidence

“Trial registration ClinicalTrials.gov NCT03813784 (https://clinicaltrials.gov/ct2/show/NCT03813784)”

AbstractFind in source

“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.”

LimitationsFind in source

paraphrase[Funding, Competing interests]

“Funding: This study was funded by Jiangsu Hengrui Pharmaceuticals. Competing interests: All authors have completed the ICMJE uniform disclosure form... and declare: support from Jiangsu Hengrui Pharmaceuticals; LS reports receiving research grants from Beigene...”

PASSReviewer 2· 85% conf

Methods are detailed, the trial is registered (NCT03813784), outcomes including negative results are reported, limitations are thoroughly discussed, conclusions are proportional, and funding/COI are disclosed; only an explicit reporting-guideline checklist reference (e.g., CONSORT) is absent.

Evidence

“Trial registration ClinicalTrials.gov NCT03813784 (https://clinicaltrials.gov/ct2/show/NCT03813784)”

AbstractFind in source

“Firstly, this open label trial used multiple therapeutic agents, and tumour responses were not assessed by a blinded independent central review committee.”

LimitationsFind in source

absence[Entire manuscript]

No explicit reporting-guideline checklist is referenced.

PASSReviewer 3· 85% conf

Methods are complete, the trial is registered, limitations are thoroughly discussed, conclusions are proportional, and funding/COI are disclosed; only an explicit reporting-guideline (CONSORT) reference is absent.

Evidence

“Trial registration ClinicalTrials.gov NCT03813784 (https://clinicaltrials.gov/ct2/show/NCT03813784)”

AbstractFind in source

“Firstly, this open label trial used multiple therapeutic agents, and tumour responses were not assessed by a blinded independent central review committee.”

LimitationsFind in source

absence[Entire manuscript]

No explicit reporting-guideline (CONSORT) reference is present.

Registered (1 ID: ClinicalTrials.gov). No reporting guideline cited.

ClinicalTrials.gov

NCT03813784

Broken references and links

None found

References checked against Crossref, OpenAlex and Retraction Watch for retractions and resolvability, plus declared data and code links probed for whether they resolve to content matching the paper.

Checked — nothing surfaced.

Checked 26 references by DOI: 26 verified.

Every extracted reference resolved against Crossref/OpenAlex with no retraction flags.

2 data/code links checked; 2 live.

dataMendeley DataLIVEHTTP 200
https://data.mendeley.com/datasets/z22zyc2b2b/1
Resolves to Mendeley Data (data repository).
datahttps://clinicaltrials.gov/ct2/show/NCT03813784LIVEHTTP 200
Resolves, but the content could not be matched to the paper.

Copyediting

1 finding · worst low

Wording, consistency and formatting errors that need correcting before submission.

Wording or formatting errors that need correctingAssessed

8 copyedit issues flagged (2 major): mostly consistency, typo, other.

MAJORotherAssociated Data section
“10.3322/caac.21834”
→ Verify and replace with the correct Mendeley dataset DOI; 10.3322/caac.21834 belongs to a CA: A Cancer Journal for Clinicians record.
DOI appears mismatched to the Mendeley dataset.
MAJORconsistencyResults, Safety / Discussion
“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.”
→ The enumerated causes (3 + 2 + 7 single causes = 12) exceed the stated total of 10; reconcile the count and list.
Content inconsistency, not purely stylistic.
MINORtypoResults, Table 1 text
“68 of (19.3%) 352 patients”
→ Change to '68 (19.3%) of 352 patients'.
Word order error.
MINORtypoHeader/Abstract banner
“e086115 e086115”
→ Remove the duplicated page/article identifier.
Duplicate string.
MINORotherAuthor byline
“Peng Zhi professor Zhang Yanqiao professor Xu Huiting professor ...”
→ Restore structured author names, degrees, and affiliations.
Author list appears run-together with titles/roles inline.
MINORconsistencyAbstract / first page
“Accepted 2026 Feb 8; Collection date 2026”
→ Verify the acceptance/collection dates are correct.
Acceptance date appears in the future relative to typical publication timelines; confirm.
MINORtypoResults, Patients
“peritoneum (68 of (19.3%) 352 patients receiving camre+CAPOX followed by camre+apa”
→ Rewrite as 'peritoneum (68 (19.3%) of 352 patients...'
Misplaced parenthesis.
MINORconsistencyThroughout
“camre based maintenance / camrelizumab based maintenance”
→ Use one consistent abbreviation for the camrelizumab-based maintenance regimens.
Minor terminology consistency.

Kaimen Rigor Action ItemsFinal pass before journal submission

The published work is robust on its core scientific, statistical, and ethical grounds, but an informed reader should weigh two correction-worthy issues: the un-reconciled treatment-related death count in the camre+apa arm (enumerated causes sum to 12 vs a stated total of 10) and the mismatched data-repository DOI. The analysis code being confined to a supplement, the incomplete demographic reporting, and the missing CONSORT checklist reference are transparency/reproducibility gaps rather than validity threats. An erratum or formal correction addressing the death-count and DOI issues would be appropriate.

1.
HIGHrigor
Reconcile the treatment-related death count in the camre+CAPOX followed by camre+apa arm (Results, Safety): the enumerated causes (3 + 2 + 7 single causes = 12) exceed the stated total of 10; correct the count or the list and issue a correction.
This is a verifiable internal contradiction in published safety data that would warrant an erratum.
2.
HIGHdata code
Deposit the analysis code in a version-controlled public repository (e.g., GitHub or Zenodo) with a persistent DOI and update the Data availability statement accordingly.
Code confined to a supplemental file is not reproducible via a permanent identifier, a standard that all three reviewers flagged.
3.
HIGHdata code
Verify and correct the Mendeley dataset DOI in the Associated Data/data availability section — 10.3322/caac.21834 is a CA: A Cancer Journal for Clinicians DOI and appears mismatched to the dataset.
A wrong identifier prevents readers from locating the deposited data, undercutting the data-availability statement.
4.
HIGHreporting
Add race/ethnicity and baseline comorbidity data to Table 1, or explicitly state the justification for their omission.
Demographic/biological-variable reporting is incomplete (weight, race/ethnicity, and comorbidity absent), a verifiable reporting gap.
5.
MEDIUMcopyedit
Fix the misplaced-parenthesis typo in Results (Patients): change '68 of (19.3%) 352 patients' to '68 (19.3%) of 352 patients' (appears twice).
The current word order is grammatically incorrect and obscures the proportion.
6.
MEDIUMreporting
Reference the completed CONSORT checklist (or state where it was submitted) in the Methods or a dedicated reporting statement.
A participant flowchart exists (Fig 1) but no reporting-guideline checklist is named, a transparency gap.
7.
MEDIUMstatistics
Provide exact p-values for the progression-free survival comparisons currently printed as thresholds (P<0.001) in the results tables, or label them as nominal secondary values.
Aligns the secondary endpoints with the exact-p-value standard used for the primary analysis.
8.
MEDIUMcopyedit
Restore the structured author byline with names, degrees, and affiliations (currently run-together with inline titles/roles).
The author list as printed is not in standard structured form, affecting attribution clarity.
9.
LOWcopyedit
Remove the duplicated article identifier 'e086115 e086115' in the header/abstract banner.
Duplicate string is an obvious production error.
10.
LOWcopyedit
Verify the acceptance and collection dates ('Accepted 2026 Feb 8; Collection date 2026') are correct.
The acceptance date appears in the future relative to typical publication timelines and should be confirmed.
11.
LOWcopyedit
Use a single consistent abbreviation for the camrelizumab-based maintenance regimens ('camre based maintenance' vs 'camrelizumab based maintenance').
Inconsistent terminology across the manuscript is a minor readability issue.
12.
LOWreporting
Add a brief statement of the rationale for the open-label design in Methods (Randomisation and masking).
The open-label status is disclosed but not justified; an explicit rationale would close the gap flagged by one reviewer.
13.
LOWother
Explicitly name the manufacturer/source of the investigational drugs (camrelizumab, apatinib) in the Methods.
Adds reagent traceability beyond the funder-implied source.
14.
LOWdata code
Add a sentence in the Data availability statement confirming whether the Mendeley dataset contains participant-level data and any conditions of access.
Clarifies the level and terms of data access for reproducibility.

Full text (67,006 chars)Analyzed: Aug 10, 2026

RigorReviewer 1

RigorReviewer 2

RigorReviewer 3

RigorPre-Submission Reviewer

CopyeditAgent Scribe

CitationsCitation Integrity

StatisticsStatistics Consistency

ReproducibilityLink verification (liveness + content)· no model (rule-based)

PreregistrationDeterministic check· no model (rule-based)

The star rating is the report’s one-glance summary. Every paper starts at 5★ and loses stars for the concrete problems the review finds — so a rating is never a vague average, it’s a running total you can read line by line under “How this rating was calculated.”

How stars are deducted

Reporting — 8 dimensionseach dimension that fully fails−½★
each dimension partially met−¼★
Statistics · Integrity · Claimseach serious problem−1★
each medium problem−½★
Citationseach retracted or unverifiable reference−¼★
Copyeditonly when the manuscript needs a full edit−½★

The rating never drops below 1★, and a demonstrable critical failure (an impossible statistic, a proven ethics violation) caps it at 1★ on its own — so the stars can never look healthy when the verdict is CRITICAL.

The rating draws on a panel of agents. Three independent Kaimen Rigor reviewers grade the eight dimensions below across several independent passes (the shown verdict is their majority vote — steadier than any single run), isolate the paper’s major claims and check its own evidence backs them, and flag integrity concerns. Alongside them, a citation agent resolves every reference against Crossref, OpenAlex, and Retraction Watch; a statistics agent recomputes reported tests; and rule-based checks verify that declared data/code links actually resolve. Full text is required — an abstract-only submission is not analyzed.

The eight reporting dimensions

Scientific PremisePrior work, rationale, and how it addresses gaps.

Study Design & RigorRandomization, blinding, power, controls, replication.

Biological VariablesSex, age, species/strain, housing, demographics.

Ethical ApprovalsIRB/IACUC approval, informed consent, regulatory compliance.

Key ResourcesAntibodies, cell lines, organisms, software — identified with persistent IDs.

Statistical ReportingNamed tests, exact p-values, effect sizes, plausible summary stats.

Data & Code AvailabilityAvailability statements, repositories, accession numbers, shared code.

Reporting TransparencyMethods, registration, guidelines, outcomes, limitations, funding/COI.

Graded against NIH, MDAR, ARRIVE 2.0, CONSORT, EQUATOR, and RRID guidelines. A dimension that doesn’t apply to the study type is skipped, never penalized.

This Kaimen Rigor review is model-assisted and is not a substitute for formal expert review. It complements human evaluation by surfacing potential methodological concerns — verify each finding against the source.

Camrelizumab plus CAPOX with camrelizumab based maintenance versus CAPOX alone as initial treatment for gastric or gastro-oesophageal junction adenocarcinoma: randomised phase 3 trial.

Peng Z, Zhang Y, Xu H, Yang Y, Yang M, Zhang M, Cheng Y, Chen X, Pan Y, Wang F, Li Q, Xu N, Liu L, Gu K, Xiao J, Zhang R, Zhao Q, Chen J, Lin L, Chen Y, Deng Y, Cai S, Bai B, Wang Y, Shen L

DOI

10.1136/bmj-2025-086115

Record issued

2026-08-10

Engine

7.29.0

Exported

2026-09-20