Gotistobart or docetaxel in metastatic squamous non-small cell lung cancer: stage 1 of the randomized phase 3 PRESERVE-003 trial.
Cho BC, Balaraman R, Chen HJ, Yu X, Fawole A, Liu ZG, Zhang J, Wu L, Yang B, Leddon JL, Hamm J, Huang Y, Wu L, Pan P, Singh P, Beardsley A, Kayali F, Davarifar A, Lee KH, Park KU, Lee Y, Li L, Wang X, Sun M, Yu Y, Jain V, Shpyro S, Wang Q, Wenger M, Şahin U, Efuni S, Song S, He K, Zheng P, Liu Y, He K, Li T, Socinski MA, Wu YL
- DOI
- 10.1038/s41591-026-04323-8
- Record issued
- 2026-08-10
- Engine
- 7.29.0
- Exported
- 2026-09-21
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/86573dda-4c72-4851-9bca-34ae3ffbe9b6 is authoritative.
How this rating was calculated
- IntegrityIntegrity concern ×2−1★
- ClaimsOverstated claim−0.5★
- ReportingData & code availability partially met−0.25★
- 01Conclusion reaches beyond the evidence
Stage 1 results suggest that gotistobart may offer a chemotherapy-free treatment option to transform the treatment paradigm of sqNSCLC.
“the results from the nonpivotal stage 1 of this phase 3 randomized study suggest that gotistobart may offer a chemotherapy-free treatment option to transform the treatment paradigm of sqNSCLC with better outcomes for patients.”
DiscussionFind in source
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 methodologically strong and unusually transparent report of an open-label, randomized phase 3 (stage 1, exploratory) trial: ethics, registration, randomization, limitations, negative results, and funding/COI are all thoroughly reported, and the 3 recomputable statistics are consistent. The main weaknesses are that the proportional-hazards assumption is not addressed, the data-availability statement names no access platform/committee (and no code-sharing), and the headline 'transform the treatment paradigm' conclusion is overstated relative to a non-powered 87-patient exploratory stage.
Statistics verification was partial: only 3 tests were machine-recomputable and all were consistent; exact and threshold-only p-values, the log-rank p, and resampling-based tests were not machine-verifiable and should not be treated as confirmed. Citations: 35 checked, 0 retracted, 0 not-found (no integrity flags). The reviewers agreed on 7 of 8 dimensions and split on data code availability; both positions are preserved in that dimension's evidence.
Numerical inconsistencies
1 finding · worst lowValues 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 3 tests: 3 consistent, 0 inconsistent; 3 via agent-written checks.
- CONSISTENTreported p = .010 · recomputed p = .012Reviewer 1Recompute the two-sided p-value for the primary OS hazard ratio from the reported HR and 95% CI using a normal (z) approximation.
“hazard ratio (HR) 0.46, 95% CI 0.25 to 0.84, two-sided P value = 0.0102”
Taken as given: the 95% CI is two-sided; the HR is a ratio so the log-scale z is appropriate (log=1); the reported p was computed from a log-rank test, which should approximate the Wald/z p derived from the CI; the CI and the p-value come from the same analysis population (87 patients with sqNSCLC)Method: z = ln(0.46)/[(ln(0.84)-ln(0.25))/(2*1.96)] ≈ 2.51; two-tailed p ≈ 0.012, consistent with the reported log-rank p of 0.0102 within approximation error.How we recomputed it: pCI(0.46,0.25,0.84,1) - CONSISTENTreported p = .010 · recomputed p = .012Reviewer 2Recompute two-sided p-value from HR and 95% CI for OS in sqNSCLC
“HR 0.46, 95% CI 0.25 to 0.84, two-sided P value = 0.0102”
Taken as given: The HR 0.46 is the point estimate from a Cox model; The 95% CI 0.25-0.84 is two-sided; The log HR is normally distributedMethod: Compute p from log HR and its 95% CI using the formula: z = log(HR) / (log(upper CI) - log(HR)) / qnorm(0.975), then two-tailed p = 2*(1 - pnorm(abs(z)))How we recomputed it: pCI(0.46, 0.25, 0.84, 1) - CONSISTENTreported p = .031 · recomputed p = .051Reviewer 2Recompute p-value from ORR comparison using Fisher's exact test
“The confirmed ORR was 20.0% (95% CI 9.6% to 34.6%) with gotistobart compared to 4.8% (95% CI 0.6% to 16.2%) with docetaxel”
Taken as given: The 9/45 and 2/42 are the response counts and totals for each arm; Fisher's exact test is appropriate for small counts; No continuity correction (mid-p = 0)Method: Two-tailed Fisher's exact test from the 2x2 table of responders vs non-respondersHow we recomputed it: pFisher2x2(9, 36, 2, 40, 0)
- lowinternal contradictionThe CI-derived z approximation for the OS HR yields p ≈ 0.012 versus the reported log-rank p = 0.0102; the difference is within normal error between a Wald/z and a log-rank test and does not indicate an error.
“hazard ratio (HR) 0.46, 95% CI 0.25 to 0.84, two-sided P value = 0.0102”
ResultsFind in source - lowinternal contradictionThe safety analysis population for docetaxel is reported as n=41, while 42 patients were randomized to docetaxel, implying one patient did not receive treatment.
Table 3: Docetaxel (n = 41) | Figure 1: 42 patients randomized to docetaxel
Table 3reviewer’s wording
Overstated conclusions
2 findings · worst mediumConclusions 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 overstated beyond the evidenceAssessed
- Conclusions only partially backed by the presented evidenceAssessed
4 major claims checked against the paper's own evidence: 1 not fully backed by the presented evidence (unsupported or overstated).
- overstatedReviewers 1, 2Stage 1 results suggest that gotistobart may offer a chemotherapy-free treatment option to transform the treatment paradigm of sqNSCLC.The evidence supports a promising signal in an exploratory, nonpivotal stage, but 'transform the treatment paradigm' reaches beyond what a single non-powered 87-patient stage can establish; the claim is speculative despite the 'may offer' hedge.Evidence: Same stage 1 efficacy/safety data (HR 0.46, ORR 20.0%, manageable safety) in 87 patients.
“the results from the nonpivotal stage 1 of this phase 3 randomized study suggest that gotistobart may offer a chemotherapy-free treatment option to transform the treatment paradigm of sqNSCLC with better outcomes for patients.”
DiscussionFind in source - partialReviewer 1By selectively depleting the Treg population, gotistobart offers a more profound restoration of antitumor immunity than CTLA-4 receptor blockade alone.The mechanistic claim is grounded in cited preclinical work, but its extension to human efficacy rests on an indirect inference from the histology differential in a small exploratory cohort, so the clinical reach is only partially supported.Evidence: Cited preclinical studies showing superior Treg depletion of gotistobart versus other anti-CTLA-4 antibodies, plus the observed differential clinical effect by histology.
“By selectively depleting the T reg population, the cellular driver of multiple suppression pathways within the tumor microenvironment, gotistobart offers a more profound restoration of antitumor immunity than CTLA-4 receptor blockade alone.”
DiscussionFind in source - partialReviewers 1, 2Gotistobart demonstrated a clinically meaningful survival benefit, with a 54% reduction in the risk of mortality compared to docetaxel.The HR of 0.46 indicates a 54% risk reduction, but the wide confidence interval (0.25-0.84) and exploratory nature limit the definitiveness of the claim.Evidence: HR 0.46, 95% CI 0.25-0.84, p=0.0102.
“gotistobart demonstrated a clinically meaningful survival benefit (primary endpoint), with a 54% reduction in the risk of mortality compared to docetaxel in patients with sqNSCLC.”
AbstractFind in source - partialReviewers 1, 2Gotistobart monotherapy can provide clinically meaningful benefit for patients with PD-(L)1-resistant and chemotherapy-resistant metastatic sqNSCLC.The OS benefit observed (HR 0.46) is promising, but the stage 1 is exploratory and not powered, so the claim is supported with caveats.Evidence: OS HR 0.46, 95% CI 0.25-0.84, p=0.0102; median OS not reached vs 10.0 months.
“Stage 1 results suggest that gotistobart monotherapy can provide clinically meaningful benefit for patients with programmed cell death protein/programmed death ligand 1-resistant and chemotherapy-resistant metastatic sqNSCLC.”
DiscussionFind in source
Efficacy claim is anchored to an adequate endpoint and a meaningful effect.
- ADEQUATESurrogate endpointThe primary endpoint is overall survival (OS), a hard clinical outcome, not a surrogate.
“The primary endpoint for the PRESERVE-003 study is OS.”
- ADEQUATEEffect sizeThe effect is large and statistically supported: HR 0.46, median OS not reached vs 10.0 months, 12-month OS rate 63.1% vs 30.3%, and the paper explicitly anchors it as clinically meaningful.
“gotistobart demonstrated a clinically meaningful survival benefit (primary endpoint), with a 54% reduction in the risk of mortality compared to docetaxel”
Data authenticity concerns
None foundAn 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.
Checked — nothing surfaced.
Reporting gaps
1 finding · worst mediumRequired 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
Prior work is cited extensively, including the modest efficacy of docetaxel and the unmet need. The rationale for gotistobart's mechanism of action is explained. However, the paper does not explicitly describe how limitations of prior research (e.g., failed trials) are addressed, though the study design implicitly addresses them.
“the median OS reported with docetaxel in our study is relatively higher than the upper end of the range reported by other studies in a second-line sqNSCLC population (8.0 to 9.4 months)”
Randomization used an Interactive Web Response System. The study is open-label (stated but no rationale). Stage 1 was exploratory and not powered, which is acknowledged. Inclusion/exclusion criteria are detailed. Outlier handling is implicit via ITT and safety populations. Blinding, randomization unit, and other criteria are adequately reported.
“Eligible patients were randomized using the Interactive Web Response System to receive gotistobart or docetaxel treatment stratified by factors that could impact prognosis: histology (squamous versus nonsquamous (stage 1 only)), presence of brain metastases (yes or no), ECOG PS score (0 versus 1) and region (USA and ex-USA).”
“randomized, open-label, active-controlled phase 3 trial”
“Stage 1 is exploratory and was not powered to demonstrate efficacy.”
“Stage 1 is exploratory and was not powered to demonstrate efficacy.”
Sex is reported (Table 1: 80% male in gotistobart, 90.5% in docetaxel). Age is reported (median 64 vs 68.5 years). Health status (ECOG PS) is reported. Demographics include race, region, smoking status. Weight is not reported, but this is not a standard requirement for clinical trials.
“Median age (range) in years | 64.0 (39–86) | 68.5 (43–84) | | Sex, no. (%) | Male | 36 (80.0) | 38 (90.5)”
“A limitation of the study is that the Asian patient population was overrepresented, and Hispanic and Black/African American patients were underrepresented in stage 1.”
“Sex, no. (%) | Male | 36 (80.0) | 38 (90.5) | | Female | 9 (20.0) | 4 (9.5)”
“ECOG PS score, no. (%) | 0 | 9 (20.0) | 7 (16.7) | | 1 | 36 (80.0) | 35 (83.3)”
“Race, no. (%) | Asian | 32 (71.1) | 30 (71.4) | | White | 11 (24.4) | 11 (26.2)”
The paper includes a dedicated ethics section stating approval by the Western Institutional Review Board and local ethics committees. Written informed consent was obtained. Compliance with regulations is stated.
“The central Western Institutional Review Board provided approval in the USA.”
“All patients provided written informed consent before enrollment.”
“The study was conducted in accordance with local and national regulations, as well as consensus ethical principles derived from the Declaration of Helsinki, the Council for International Organizations of Medical Sciences International Ethical Guidelines and the International Council for Harmonisation Good Clinical Practice Guidelines.”
“The central Western Institutional Review Board provided approval in the USA.”
“All patients provided written informed consent before enrollment.”
Gotistobart is identified by name, source (OncoC4), and dose regimen. Docetaxel is identified by name and dose. SAS version 9.4 is specified. No other resources (antibodies, cell lines, etc.) are applicable.
“Statistical analyses were conducted using SAS (version 9.4 or later).”
“Statistical analyses were conducted using SAS (version 9.4 or later).”
Ext and effect sizes with CIs are reported. The paper does not mention checking proportional hazards or other assumptions, which is a minor omission. Data presentation includes KM curves and tables with n per group.
“hazard ratio (HR) 0.46, 95% CI 0.25 to 0.84, two-sided P value = 0.0102”
“HRs and associated 95% CIs were calculated from a Cox proportional-hazards model.”
The statement says data will be made available to qualified researchers upon request within 12 months, but no concrete platform or committee is specified. This does not meet the standard for adequate reporting.
“Upon completion of this clinical trial, the data that support the findings of this study will be made available to qualified researchers. Proposals should be directed to pzheng@oncoc4.com.”
“To gain access, data requestors will need to sign a data access agreement. All data provided will be anonymized to respect the privacy of patients who have participated in the trial”
“Data will be made available; to qualified researchers upon request, within 12 months of the study completion data and for a period of 2 years thereafter.”
All applicable sub-criteria are adequately addressed. The paper is written in a transparent manner appropriate for a clinical trial report.
“PRESERVE-003 (ClinicalTrials.gov registration: NCT05671510 (https://clinicaltrials.gov/ct2/show/NCT05671510) ; date of registration: 4 January 2023)”
“For patients with non‑sqNSCLC, the median OS, 12-month survival rates and median PFS were numerically better in the docetaxel arm than in the gotistobart arm”
“The use of investigator assessment rather than blinded, independent central reviewer assessment of PFS and ORR may be limited by the open-label design.”
“ClinicalTrials.gov identifier: NCT05671510”
“A limitation of the study is that the Asian patient population was overrepresented, and Hispanic and Black/African American patients were underrepresented in stage 1.”
Registered (1 ID: ClinicalTrials.gov). No reporting guideline cited.
Broken references and links
None foundReferences 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 35 references by DOI: 31 verified — 4 no DOI (shown, not verified).
- NO DOINCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines): Non-Small Cell Lung Cancer Version 8No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIFirst-in-human study of the first acid pH-sensitive and recycling CTLA-4 antibody that preserves the immune tolerance checkpoint to avoid immunotherapy-related adverse events in cancer patientsNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIPharmacokinetics of first and repeated dosing of non-irAE-inducing anti-CTLA-4 monoclonal antibody ONC-392 in advanced cancer patientsNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOISingle-agent safety and activities of target-preserving anti-CTLA-4 antibody gotistobart (ONC-392/BNT316) in PD-(L)1 resistant metastatic NSCLC and population PK analysis in patients with solid tumorsNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
3 data/code links checked; 3 live.
- datahttps://clinicaltrials.gov/search?term=NCT05671510LIVEHTTP 200Resolves, but the content could not be matched to the paper.
- datahttps://clinicaltrials.gov/ct2/show/NCT05671510LIVEHTTP 200Resolves, but the content could not be matched to the paper.
- datahttp://clinicaltrials.gov/search?term=NCT05671510LIVEHTTP 200Resolves, but the content could not be matched to the paper.
Copyediting
3 minorWording, consistency and formatting errors that need correcting before submission.
No major wording or formatting errors. 3 minor suggestions below.
3 copyedit issues flagged: mostly consistency, clarity.
- MINORconsistencyAbstract versus Results, Safety“grade ≥3 treatment-related adverse events in 42% and 49% of patients receiving gotistobart and docetaxel”→ Align the abstract percentages with the body (42.2% and 48.8%) or state they are rounded.The body reports 42.2% versus 48.8%; the abstract rounds to 42% and 49%. Acceptable rounding, but alignment would remove ambiguity.
- MINORclarityDiscussion, CTLA-4 dose context“ipilimumab—another CTLA‑4-directed antibody—is approved at substantially lower doses (typically 3 mg kg−1)”→ Consider noting that the comparison is to a different agent/indication to avoid confusion with the 3 mg/kg gotistobart arm terminated by the DMC.The 3 mg/kg gotistobart arm was terminated for poor outcomes, while ipilimumab is approved at 3 mg/kg; the sentence is clear but could be tightened.
- MINORconsistencySafety results, Table 3 footnote“AEs, no. (%) | Gotistobart (n = 45) | Docetaxel (n = 41)”→ Clarify that one patient in the docetaxel arm did not receive treatment, explaining the n=41 in safety analysis.The discrepancy between randomized n=42 and safety n=41 is explained in the text but could be more explicit.
As a published record, the study is robust and meets clinical-trial reporting standards, but an informed reader should weigh the exploratory, non-powered nature of stage 1 and the open-label, investigator-assessed endpoints when interpreting the OS/HR signal, and should treat the 'transform the treatment paradigm' conclusion as stronger than the evidence supports. A correction or clarification addressing the proportional-hazards assumption, the safety-population n discrepancy (n=41 vs randomized n=42), and tempering the headline claim would strengthen the record; the data-access statement should name a platform/committee to be fully actionable.
- 1.HIGHrigorPublish a correction or clarification in the Discussion replacing the 'transform the treatment paradigm of sqNSCLC' claim with language limited to a promising, exploratory signal, since a non-powered 87-patient stage 1 cannot support a paradigm-shift conclusion.The claim-audit component rated this headline conclusion as overstated relative to the evidence; over-claiming is the most common and most consequential reviewer/reader objection.
- 2.HIGHstatisticsAdd a statement to the Statistical analysis section (or a correction note) on whether the proportional-hazards assumption of the Cox model was assessed for OS and PFS; if not assessed, say so and flag it as a limitation.Both reviewers noted the assumption check is absent, and readers cannot assess the validity of the primary HR without knowing whether PH held.
- 3.MEDIUMdata codeClarify in the Data availability section whether the SAS analysis code and analysis datasets will be shared alongside the patient data, and name a specific managed-access platform (e.g., Vivli, YODA) or data-access committee rather than only an email address.Reviewer 2 judged the statement vague without a named platform/committee, and Reviewer 1 noted code-sharing is unaddressed; making the route concrete and including code improves reproducibility.
- 4.MEDIUMreportingAdd a brief sample-size/power rationale for stage 1 in the Statistical analysis section, even if only to quantify the number of patients needed to characterize the dose and frame the exploratory intent.Reviewer 1 flagged the absence of any a priori sample-size rationale; the current text only states the stage 'was not powered to demonstrate efficacy'.
- 5.MEDIUMreportingState the rationale for the open-label design in the Methods (Study design) section, and note explicitly in the Discussion's limitations that open-label, investigator-assessed PFS/ORR may introduce bias.Reviewer 2 noted the open-label design is stated without justification; the limitation is partially acknowledged but the lack of rationale should be disclosed.
- 6.MEDIUMcopyeditClarify in the Table 3 footnote (or safety text) that one randomized docetaxel patient did not receive treatment, explaining the safety analysis n=41 versus the randomized n=42.The copyedit pass flagged this n discrepancy as potentially confusing to readers even though it is explained in the text.
- 7.MEDIUMcopyeditAlign the abstract safety percentages (42% and 49%) with the body values (42.2% and 48.8%), or state explicitly that they are rounded.The copyedit pass flagged a consistency ambiguity between the abstract and the body safety results.
- 8.LOWcopyeditAdd a clarifying note in the Discussion that the ipilimumab 3 mg/kg dose is a different agent/indication, to avoid confusion with the 3 mg/kg gotistobart arm that was terminated by the DMC.The copyedit pass flagged that the ipilimumab comparison could be misread as referring to the terminated gotistobart 3 mg/kg arm.
- 9.LOWreportingAdd a brief note in Table 1 or the Methods that body weight was not reported, or state that weight is not a required baseline variable for this clinical-trial context.Reviewer 2 noted weight is not reported; an explicit statement avoids the appearance of an omission.
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.