Peri-operative atezolizumab in early-stage triple-negative breast cancer: final results and ctDNA analyses from the randomized phase 3 IMpassion031 trial.
Mittendorf EA, Assaf ZJ, Harbeck N, Zhang H, Saji S, Jung KH, Hegg R, Koehler A, Sohn J, Iwata H, Telli ML, Ferrario C, Punie K, Qamra A, Dieterich M, Xu Y, Liste-Hermoso M, Shearer-Kang E, Molinero L, Chui SY, Barrios CH
- DOI
- 10.1038/s41591-025-03725-4
- Record issued
- 2026-08-15
- Engine
- 7.39.0
- Exported
- 2026-09-22
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/7bd9e270-1f19-44c5-ba10-9438bd023ff5 is authoritative.
How this rating was calculated
- StatisticsStatistic did not reproduce ×2−1★
- IntegrityIntegrity concern ×2−1★
- CitationsCitations & links (capped) ×5−1★
- ClaimsEfficacy rests on an unvalidated surrogate endpoint−0.5★
- ClaimsTreatment effect not shown to be clinically meaningful−0.5★
Citations & links are capped at −1★ combined, however many are flagged.
- 01Printed percentage does not match its own count
56% does not match the reported count 43/67
“43 of 67 (56%)”
- 02Printed percentage does not match its own count
56% does not match the reported count 43/67
“43 of 67 (56%)”
Placebo group, week 7 clearanceFind in source - 03Efficacy rests on an unvalidated surrogate endpoint
The primary efficacy claim is based on pathologic complete response (pCR), a surrogate endpoint for long-term clinical outcomes. Although the trial also reports secondary endpoints like event-free survival (EFS), disease-free survival (DFS), and overall survival (OS), the primary endpoint and the main claim of improved outcomes are anchored on pCR. The paper does not provide evidence of target engagement at the tested dose (e.g., PK/PD data) nor does it cite validated evidence linking pCR to the clinical outcomes in this specific context. The discussion acknowledges that the trial was not powered for secondary endpoints and that pCR is a surrogate, but it does not establish the required link.
“At the primary analysis, the pathologic complete response (pCR) rate was improved significantly with the addition of atezolizumab to standard neoadjuvant chemotherapy. In the intention-to-treat (ITT) population, pCR rates were 58% with atezolizumab plus…”
- 04Treatment effect not shown to be clinically meaningful
The primary reported effect is a 17% absolute improvement in pCR rate (from 41% to 58%). While statistically significant, the clinical meaningfulness of this surrogate endpoint is not anchored to a minimal clinically important difference or to long-term survival benefit. The secondary endpoints (EFS, DFS, OS) show hazard ratios below 1 but with confidence intervals crossing 1, and the trial was not powered for these. The effect size is presented without explicit anchoring to clinical meaningfulness beyond statistical significance.
“pCR rates were 58% with atezolizumab plus chemotherapy versus 41% with placebo plus chemotherapy, representing a 17% improvement (95% confidence interval (CI), 6–27%; 1-sided P = 0.0044, crossing the significance boundary).”
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 randomized phase 3 trial with rigorous design, clear reporting of ethics, resources, and data availability. The main weaknesses are minor reporting gaps: the randomization method is not detailed in this paper, and the ctDNA-evaluable population is only 42% of ITT with reasons for non-evaluability not fully explained. The copyedit found only minor issues, and the citation check flagged 5 references not found in registries, which warrant verification.
Both reviewers classified the study as interventional; no divergence. The evaluation covered the full text, including methods, results, and supplementary materials. Non-applicable sub-criteria (e.g., animal housing, cell line authentication) were excluded. The statistics verification covered only a subset of tests (3 reported with test statistics/CIs), so the paper's statistics are not fully verified. The citation check flagged 5 references not found in registries, which are potential fabrication signals.
Numerical inconsistencies
2 findings · worst highValues 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.
- Summary statistic impossible for the stated N (GRIM/GRIMMER)Recomputed
- Internal contradictions in the reported numbersAssessed
Recomputed 1 test: 1 consistent, 0 inconsistent; 1 via agent-written checks. 2 reported summary statistics mathematically impossible for the stated N (PERCENT).
- PERCENT56% does not match the reported count 43/67
“43 of 67 (56%)”
- PERCENT56% does not match the reported count 43/67
“43 of 67 (56%)”
Placebo group, week 7 clearanceFind in source
- CONSISTENTreported p = .004 · recomputed p = .002Reviewers 1, 2Verify the p-value for the primary pCR comparison (1-sided P=0.0044) using the reported pCR rates and sample sizes.
“In the intention-to-treat (ITT) population, pCR rates were 58% with atezolizumab plus chemotherapy versus 41% with placebo plus chemotherapy, representing a 17% improvement (95% confidence interval (CI), 6–27%; 1-sided P = 0.0044, crossing the significance boundary).”
Taken as given: The pCR rates are 58% of 165 (atezolizumab) and 41% of 168 (placebo), giving counts 96 and 69 respectively.; The test is a two-sided chi-square test; the reported one-sided p is half of the two-sided p.; The sample sizes are 165 and 168.Method: Recomputed two-sided Pearson chi-square p-value from cell counts (96,69,69,99) and halved it to approximate the one-sided p.How we recomputed it: pChi2x2(96, 69, 69, 99)
- lowinternal contradictionThe abstract states '14 of 70 (20%) atezolizumab-treated and 33 of 99 (33%) placebo-treated patients received additional adjuvant therapy', but Extended Data Table 1 lists 'Any' as 14 (20) and 33 (33) for the same groups, which is consistent. However, the text in Results says 'a higher percentage in the placebo group received adjuvant systemic therapy (33% compared with 20% of atezolizumab-treated patients)', which matches. No contradiction found.
“14 of 70 (20%) atezolizumab-treated and 33 of 99 (33%) placebo-treated patients received additional adjuvant therapy”
AbstractFind in source
Overstated conclusions
2 findings · worst highConclusions 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.
- Efficacy rests on an unvalidated surrogate endpointAssessed
- Treatment effect not shown to be clinically meaningfulAssessed
8 major claims checked against the paper's own evidence: all adequately supported.
- supportedReviewers 1, 2Adding atezolizumab to chemotherapy for stage II/III TNBC is associated with favorable long-term outcomes.The HRs for EFS, DFS, and OS all favor atezolizumab, though CIs cross 1; the claim is appropriately hedged as 'associated with favorable long-term outcomes'.Evidence: HRs: EFS 0.76 (0.47-1.21), DFS 0.76 (0.44-1.30), OS 0.56 (0.30-1.04).
“These data show that adding atezolizumab to chemotherapy for stage II/III TNBC is associated with favorable long-term outcomes, and ctDNA dynamics provide prognostic value beyond pCR.”
AbstractFind in source - supportedReviewers 1, 2ctDNA dynamics provide prognostic value beyond pCR.The paper shows that ctDNA clearance at surgery is associated with improved DFS and OS, and persistent ctDNA identifies poor prognosis, even among non-pCR patients.Evidence: ctDNA clearance at surgery associated with improved DFS and OS; non-pCR patients with persistent ctDNA had poor prognosis (DFS HR 2.15, OS HR 3.50).
“ctDNA dynamics provide prognostic value beyond pCR.”
AbstractFind in source - supportedReviewer 1The positive impact of atezolizumab on pCR was supported by HR point estimates below 1 favoring atezolizumab for EFS, DFS and OS.All HR point estimates are below 1, consistent with the pCR benefit, though not statistically significant.Evidence: HRs: EFS 0.76, DFS 0.76, OS 0.56.
“Long-term follow-up of the randomized phase 3 IMpassion031 trial indicated that the positive impact of atezolizumab on the primary outcome measure (pCR) was supported by HR point estimates below 1 favoring atezolizumab for EFS, DFS and OS.”
DiscussionFind in source - supportedReviewer 1Peri-operative atezolizumab was well tolerated with no unexpected safety findings.Safety data show no new signals and no treatment-related deaths with atezolizumab.Evidence: Extended Data Table 3 shows no grade 5 AESI; safety profiles consistent with primary analysis.
“Peri-operative atezolizumab (combined with neoadjuvant chemotherapy and continued as adjuvant therapy) was well tolerated and there were no unexpected safety findings.”
DiscussionFind in source - supportedReviewer 1Absence of ctDNA at baseline was rare but associated with excellent long-term outcomes.Only 6% were ctDNA-negative at baseline, and these patients had excellent outcomes, though the sample size is small.Evidence: Nine patients ctDNA-negative at baseline; only one recurrence and no deaths.
“Absence of ctDNA at baseline was rare (nine patients (6%): four of 63 (6%) in the atezolizumab group and five of 76 (7%) in the placebo group).”
ResultsFind in source - supportedReviewers 1, 2ctDNA-positive status at surgery identified a subset of non-pCR patients with poorest prognosis.Non-pCR patients with persistent ctDNA at surgery had worse DFS and OS compared to those with clearance.Evidence: DFS HR 2.15 (0.88-5.29), OS HR 3.50 (1.21-10.16).
Among patients without a pCR, the 16 with persistent ctDNA at surgery had a particularly poor prognosis compared with patients with cleared ctDNA at surgery, although the 95% CIs for the DFS HR point estimate crossed 1 (DFS HR, 2.15; 95% CI, 0.88–5.29; OS HR, 3.50; 95% CI, 1.21–10.16).
Resultsreviewer’s wording - supportedReviewer 2Patients with baseline ctDNA-negative status had excellent long-term outcomes.The paper reports that among the nine ctDNA-negative patients, only one had recurrence and none died, supporting the claim, though the sample size is very small.Evidence: Nine patients ctDNA-negative at baseline; only one recurrence and no deaths.
Absence of ctDNA at baseline was rare (nine patients (6%)...). These patients remained ctDNA-negative at all timepoints sampled. ... only one had disease recurrence (>3 years after surgery) and none had died by the data cutoff date.
Resultsreviewer’s wording - supportedReviewer 2Most patients had cleared ctDNA at surgery.The paper reports 106 of 122 (87%) ctDNA-positive at baseline cleared at surgery, supporting the claim.Evidence: 106 of 122 (87%) patients with positive ctDNA at baseline had negative ctDNA at surgery.
“Most (106 of 122 (87%)) of the patients with positive ctDNA at baseline had negative ctDNA status at the time of surgery”
ResultsFind in source
Premise concern: surrogate not validated for clinical benefit; effect size not shown to be clinically meaningful.
- INADEQUATESurrogate endpointThe primary efficacy claim is based on pathologic complete response (pCR), a surrogate endpoint for long-term clinical outcomes. Although the trial also reports secondary endpoints like event-free survival (EFS), disease-free survival (DFS), and overall survival (OS), the primary endpoint and the main claim of improved outcomes are anchored on pCR. The paper does not provide evidence of target engagement at the tested dose (e.g., PK/PD data) nor does it cite validated evidence linking pCR to the clinical outcomes in this specific context. The discussion acknowledges that the trial was not powered for secondary endpoints and that pCR is a surrogate, but it does not establish the required link.
“At the primary analysis, the pathologic complete response (pCR) rate was improved significantly with the addition of atezolizumab to standard neoadjuvant chemotherapy. In the intention-to-treat (ITT) population, pCR rates were 58% with atezolizumab plus chemotherapy versus 41% with placebo plus chemotherapy, representing a 17% improvement (95% confidence interval (CI), 6–27%; 1-sided P = 0.0044, crossing the significance boundary).”
- INADEQUATEEffect sizeThe primary reported effect is a 17% absolute improvement in pCR rate (from 41% to 58%). While statistically significant, the clinical meaningfulness of this surrogate endpoint is not anchored to a minimal clinically important difference or to long-term survival benefit. The secondary endpoints (EFS, DFS, OS) show hazard ratios below 1 but with confidence intervals crossing 1, and the trial was not powered for these. The effect size is presented without explicit anchoring to clinical meaningfulness beyond statistical significance.
“pCR rates were 58% with atezolizumab plus chemotherapy versus 41% with placebo plus chemotherapy, representing a 17% improvement (95% confidence interval (CI), 6–27%; 1-sided P = 0.0044, crossing the significance boundary).”
Data authenticity concerns
1 finding · worst lowAn 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
2 integrity concerns flagged (0 high).
- lowotherThe ctDNA-evaluable population (n=139) is 42% of the ITT population (333). The paper states 'Overall, 139 patients (42% of the ITT population) were evaluable for ctDNA analysis at baseline.' This is plausible given sample availability, but the reasons for non-evaluability are not detailed.
“Overall, 139 patients (42% of the ITT population) were evaluable for ctDNA analysis at baseline.”
ResultsFind in source
Reporting gaps
None foundRequired detail the manuscript never states — study design, biological variables, ethics approval and consent, key resources, statistical reporting, data and code availability, and overall transparency.
Checked — nothing surfaced.
The introduction cites multiple randomized phase 3 trials (KEYNOTE-522, NeoTRIP, GeparNuevo) and prior ctDNA studies, acknowledging strengths and limitations. The rationale linking the premise to the study objectives is logical, and the paper explicitly addresses limitations of prior ctDNA research (e.g., few patients received immunotherapy).
“In patients on first-line treatment for advanced TNBC, benefit from immune checkpoint blockade is greatest in those with tumors expressing programmed death-ligand 1 (PD-L1) – . However, the benefit appears to be broader in randomized phase 3 trials of preoperative (neoadjuvant) therapy for early-stage TNBC, where the addition of immunotherapy improves the efficacy of chemotherapy regardless of PD-L1 status – .”
“However, in both of these reports, very few patients received immunotherapy for TNBC, limiting their generalizability.”
“The IMpassion031 dataset provided the opportunity to expand our understanding of ctDNA as a potential biomarker for individualizing immunotherapy-containing treatment for early-stage TNBC.”
“However, in both of these reports, very few patients received immunotherapy for TNBC, limiting their generalizability.”
Randomization method is stated (1:1 ratio, stratified by stage and PD-L1 status). Blinding is described (double-blind for neoadjuvant phase, open-label for adjuvant atezolizumab). Power analysis is referenced from the primary publication. Inclusion/exclusion criteria are detailed. Outlier handling is addressed via ITT and per-protocol analyses. Controls are the placebo group. Independent replication is not applicable for a single pivotal trial.
“Eligible patients were randomized in a 1:1 ratio to receive either atezolizumab or placebo in combination with a neoadjuvant chemotherapy regimen comprising 12 weeks of nab-paclitaxel followed by 8 weeks of dose-dense doxorubicin plus cyclophosphamide, with filgrastim or pegfilgrastim support.”
“After surgery, patients initially randomized to atezolizumab received open-label atezolizumab for an additional 11 cycles.”
“The trial enrolled women or men aged ≥18 years with untreated stage II/III TNBC (centrally assessed according to American Society of Clinical Oncology/College of American Pathologists guidelines) and a primary tumor >2 cm (cT2–cT4, cN0–cN3, cM0).”
“Eligible patients were randomized in a 1:1 ratio to receive either atezolizumab or placebo”
“After surgery, patients initially randomized to atezolizumab received open-label atezolizumab for an additional 11 cycles.”
“Patients were excluded if they had previously received systemic therapy for breast cancer or anthracycline, taxane, anti-CTLA-4, anti-PD-1 or anti-PD-L1 therapy for any malignancy”
Sex is reported (all female). Age and ECOG PS are reported in Table 1. Demographics include race, stage, PD-L1 status, tumor grade, and LN involvement. Since all patients are female, sex_justified is not applicable (no single-sex justification needed as the trial enrolled only women, but this is a study population characteristic, not a design choice requiring justification).
“All patients were female.”
“Age, years | <65 | 148 (90) | 139 (83) | 124 (89) | | ≥65 | 17 (10) | 29 (17) | 15 (11)”
“Self-reported race | White | 102 (62) | 108 (64) | 80 (58) | | Asian | 47 (28) | 41 (24) | 51 (37)”
“All patients were female.”
“Age, years | <65 | 148 (90) | 139 (83) | 124 (89)”
“Self-reported race | White | 102 (62) | 108 (64) | 80 (58)”
The Methods state the trial was conducted in accordance with Good Clinical Practice and the Declaration of Helsinki, and that the protocol was approved by institutional review boards (IRBs) of participating institutions, including named IRBs. All patients provided written informed consent. This satisfies both irb_ethics_statement and informed_consent.
“The study protocol was approved by institutional review boards (IRBs) of participating institutions, including the Taipei General Hospital IRB and the Western IRB.”
“All patients provided written informed consent.”
“The trial was conducted in accordance with Good Clinical Practice and the Declaration of Helsinki.”
“The study protocol was approved by institutional review boards (IRBs) of participating institutions, including the Taipei General Hospital IRB and the Western IRB.”
“All patients provided written informed consent.”
“The trial was conducted in accordance with Good Clinical Practice and the Declaration of Helsinki.”
Atezolizumab is identified as the study drug with manufacturer (Genentech) and dosing. Chemotherapy agents are named with doses. The ctDNA assay (Signatera, Natera) is identified. Software (SAS v.9.4) is identified. Antibodies, cell lines, and mycoplasma testing are not applicable.
“Atezolizumab was administered intravenously at 840 mg every 2 weeks in the neoadjuvant phase (to align with the chemotherapy schedule) and at 1,200 mg every 3 weeks in the adjuvant phase (for convenience and to reduce the burden of treatment visits).”
“Nab-paclitaxel was administered at 125 mg m −2 every week; doxorubicin 60 mg m −2 and cyclophosphamide 600 mg m −2 were administered every 2 weeks.”
“In further exploratory analyses, the prognostic and predictive effects of baseline and on-treatment ctDNA were evaluated using the tumor-informed Signatera assay (Natera).”
“Nab-paclitaxel was administered at 125 mg m −2 every week; doxorubicin 60 mg m −2 and cyclophosphamide 600 mg m −2 were administered every 2 weeks.”
“the tumor-informed Signatera assay (Natera)”
Tests are named (Kaplan-Meier, Cox proportional hazards implied by HRs). Assumptions are handled by design (pre-specified analysis). Exact p-values are not reported for secondary endpoints (HRs with CIs), which is acceptable as estimation. Effect sizes with CIs are reported. Software is identified. Data presentation includes KM curves and forest plots. Mathematical plausibility checks: I attempted to verify some percentages but found no clear inconsistencies; the paper reports HRs and CIs that are plausible.
“Kaplan–Meier analysis was applied to estimate median duration of EFS, DFS and OS, with the Brookmeyer–Crowley method used to construct the 95% CIs.”
“Long-term outcomes favored the atezolizumab group (event-free survival hazard ratio (HR), 0.76; 95% confidence interval (CI), 0.47–1.21; disease-free survival HR, 0.76; 95% CI, 0.44–1.30; overall survival HR, 0.56; 95% CI, 0.30–1.04).”
“SAS v.9.4 was used for all statistical analyses.”
“Kaplan–Meier analysis was applied to estimate median duration of EFS, DFS and OS, with the Brookmeyer–Crowley method used to construct the 95% CIs.”
“EFS was more favorable with the atezolizumab-containing regimen (hazard ratio (HR), 0.76; 95% CI, 0.47–1.21)”
“SAS v.9.4 was used for all statistical analyses.”
The data availability statement provides concrete access routes: Vivli platform for patient-level data and EGA for biomarker data with accession number. Repository deposit is applicable for biomarker data (EGA). Accession numbers are provided. Code sharing is not applicable as no custom code is mentioned.
“Qualified researchers may request access to individual patient-level data through the clinical study data request platform ( https://vivli.org (https://vivli.org/) ).”
“Biomarker data will be made available to qualified researchers at the European Genome-Phenome Archive (EGA) under accession number EGAS50000000974 (https://ega-archive.org/studies/EGAS50000000974) .”
“Biomarker data will be made available to qualified researchers at the European Genome-Phenome Archive (EGA) under accession number EGAS50000000974”
Trial registration number is provided. Methods are detailed enough for replication. Reporting guideline (Nature Portfolio reporting summary) is mentioned. All outcomes are reported, including negative results. Limitations are explicitly discussed. Conclusions are proportional, acknowledging the trial was not powered for secondary endpoints. Funding and COI are disclosed.
“ClinicalTrials.gov identifier: NCT03197935 (https://clinicaltrials.gov/study/NCT03197935) .”
“Study limitations include the smaller sample size required for a pCR primary endpoint and the lack of power to detect differences in EFS, DFS and OS.”
“F. Hoffmann-La Roche Ltd sponsored and funded the trial, provided the study drug (atezolizumab) and collaborated with academic authors on the study design, data collection, data analysis, interpretation, preparation of the manuscript and decision to publish.”
“ClinicalTrials.gov identifier: NCT03197935”
“Study limitations include the smaller sample size required for a pCR primary endpoint and the lack of power to detect differences in EFS, DFS and OS.”
“F. Hoffmann-La Roche Ltd sponsored and funded the trial, provided the study drug (atezolizumab)”
Registered (1 ID: ClinicalTrials.gov). No reporting guideline cited.
Broken references and links
1 finding · worst lowReferences 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.
- References not resolvable to a published paperRecomputed
Checked 30 references by DOI: 22 verified — 5 DOI unresolved, 3 no DOI (shown, not verified).
- UNRESOLVED10.1056/nejmoa2202802Pembrolizumab plus chemotherapy in advanced triple-negative breast cancerCited DOI does not resolve to any Crossref record.
- UNRESOLVED10.1038/s41523-022-00472-8Patient-reported outcomes from a randomized trial of neoadjuvant atezolizumab-chemotherapy in early triple-negative breast cancerCited DOI does not resolve to any Crossref record.
- UNRESOLVED10.1001/jamaoncol.2020.0075Effect of pembrolizumab plus neoadjuvant chemotherapy on pathologic complete response in women with early-stage breast cancer: an analysis of the ongoing phase 2 adaptively randomized I-SPY2 trialCited DOI does not resolve to any Crossref record.
- UNRESOLVED10.1038/s41523-023-00512-xctDNA and residual cancer burden are prognostic in triple-negative breast cancer patients with residual diseaseCited DOI does not resolve to any Crossref record.
- UNRESOLVED10.1023/a:1026425100045Modification of the EORTC QLQC30 (version 2.0) based on content validity and reliability testing in large samples of patients with cancer: the Study Group on Quality of Life of the EORTC and the Symptom Control and Quality of Life Committees of the NCI of Canada Clinical Trials GroupCited DOI does not resolve to any Crossref record.
- NO DOINSABP B-59/GBG-96-GeparDouze: a randomized double-blind phase III clinical trial of neoadjuvant chemotherapy with atezolizumab or placebo followed by adjuvant atezolizumab or placebo in patients with stage II and III triple-negative breast cancerNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIEORTC QLQ-C30 Scoring ManualNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIManual of the Functional Assessment of Chronic Illness Therapy (FACIT) Measurement SystemNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
4 data/code links checked; 4 live.
- datahttps://vivli.orgLIVEHTTP 200Resolves, but the content could not be matched to the paper.
- datahttps://vivli.org/members/ourmembersLIVEHTTP 200Resolves, but the content could not be matched to the paper.
- datahttps://www.roche.com/research_and_development/who_we_are_how_we_work/clinical_trials/our_commitment_to_data_sharing.htmLIVEHTTP 200Resolves, but the content could not be matched to the paper.
- dataEGALIVEHTTP 200https://ega-archive.org/studies/EGAS50000000974Resolves to EGA (data repository).
Copyediting
5 minorWording, consistency and formatting errors that need correcting before submission.
No major wording or formatting errors. 5 minor suggestions below.
5 copyedit issues flagged: mostly typo, consistency, clarity.
- MINORtypoAbstract“circulating tumor (ct)DNA”→ Consider defining ctDNA at first use as 'circulating tumor DNA (ctDNA)' for consistency.Minor formatting inconsistency.
- MINORconsistencyData availability“https://vivli.org (10.1056/NEJMoa1809615)”→ The DOI appears misplaced; consider removing or correcting.Potential copy-paste error.
- MINORtypoAbstract“circulating tumor (ct)DNA”→ Consider defining ctDNA at first use as 'circulating tumor DNA (ctDNA)' for clarity.The abbreviation is defined but the parentheses are misplaced.
- MINORconsistencyResults, Patient disposition“All but one patient in each treatment group received neoadjuvant therapy and 308 (155 in the atezolizumab group, 153 in the placebo group) underwent surgery”→ Verify that the numbers 155 and 153 sum to 308; they do, but ensure consistency with Figure 1.The numbers are consistent, but the sentence could be clearer.
- MINORclarityMethods, Endpoints and assessments“In a subset of patients (approximately 50%), matched normal was based on whole plasma whole-exome sequencing.”→ Rephrase to 'In a subset of patients (approximately 50%), matched normal was derived from whole-exome sequencing of plasma.'The phrase 'whole plasma whole-exome sequencing' is awkward.
The published work is robust and well-reported. An informed reader should weigh the minor reporting gaps (randomization method not detailed, ctDNA evaluability reasons) and the 5 references not found in registries, which may warrant a correction or clarification. The statistics were only partially verified, so independent re-analysis of the reported HRs is advisable.
- 1.HIGHreportingIn the Methods (Patients and treatment), specify the randomization method (e.g., central web-based system) and allocation concealment mechanism, or explicitly reference the primary publication where this is detailed.Reviewer 2 flagged randomization_method as reported_but_inadequate; providing the method enhances reproducibility.
- 2.HIGHreportingVerify the 5 references not found in registries (DOIs: 10.1056/nejmoa2202802, 10.1038/s41523-022-00472-8, 10.1001/jamaoncol.2020.0075, 10.1038/s41523-023-00512-x, 10.1023/a:1026425100045) and correct or replace any that are erroneous or fabricated.References not found in any registry may be fabricated; verification is essential for integrity.
- 3.MEDIUMreportingIn the Results (ctDNA analyses), report the number of patients with missing ctDNA samples at each timepoint and the reasons for non-evaluability (e.g., insufficient tumor tissue, failed sequencing).The ctDNA-evaluable population is only 42% of ITT; transparency about missingness is important for interpreting the biomarker analyses.
- 4.MEDIUMstatisticsIn the Methods (Statistical analysis), explicitly state the statistical model used for HR estimation (e.g., Cox proportional hazards) and the test used for comparing ctDNA clearance rates (e.g., Fisher's exact test).Both reviewers noted that the specific tests are implied but not explicitly named; clarity aids reproducibility.
- 5.MEDIUMreportingIn the Results (Secondary outcomes), consider reporting exact p-values for EFS, DFS, and OS alongside the HRs and CIs, even though the trial was not powered for these endpoints.Exact p-values facilitate meta-analyses and allow readers to assess statistical significance more precisely.
- 6.MEDIUMreportingIn the Discussion, explicitly state that the ctDNA analyses are exploratory and not adjusted for multiplicity, and that the HR CIs for EFS and DFS include 1, so the results are not definitive.This tempers over-claiming and clarifies the exploratory nature of the biomarker findings.
- 7.MEDIUMdata codeIn the Data Availability section, correct the misplaced DOI in the Vivli link (currently shows '10.1056/NEJMoa1809615') and provide a direct link to the EGA data access request process.The copyedit flagged a potential copy-paste error; fixing it ensures accurate data access information.
- 8.MEDIUMreportingIn the Methods (Endpoints and assessments), rephrase 'whole plasma whole-exome sequencing' to 'whole-exome sequencing of plasma' for clarity.The copyedit flagged this as awkward phrasing; clarity improves readability.
- 9.LOWcopyeditIn the Abstract, define ctDNA at first use as 'circulating tumor DNA (ctDNA)' and fix the misplaced parentheses.The copyedit flagged a minor formatting inconsistency; fixing it improves professionalism.
- 10.LOWreportingIn the Methods (Endpoints and assessments), describe the ctDNA assay validation and limit of detection, and specify the version of the Signatera assay and sequencing platform used.This strengthens the key_resources reporting and helps readers assess assay reliability.
- 11.LOWreportingIn the Results (Subgroup analyses), report the number of events and person-years for each subgroup to allow readers to assess the precision of the HRs.This improves transparency and interpretability of subgroup findings.
- 12.LOWdata codeConsider depositing the statistical analysis code (e.g., SAS programs) in a public repository to enhance reproducibility.While not required, sharing code would allow independent verification of the analyses.
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.”
- 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.
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.