Alpha¹ by KaimenConfidential

Kaimen Rigor review

Elraglusib and chemotherapy in metastatic pancreatic ductal adenocarcinoma: a randomized controlled phase 2 trial.

Mahalingam D, Shroff RT, Carneiro BA, Ji Y, Coveler AL, Cervantes A, Sahai V, Ploquin A, Hiret S, LoConte NK, Percent IJ, Lopez CD, Pernot S, Kavan P, Mulcahy M, Carr R, Giles FJ, Seifarth C, Ugolkov A, Weiskittel T, Fine G, Jaros M, Mazar AP, Bekaii-Saab TS

DOI: 10.1038/s41591-026-04327-4
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/31174cab-7811-4612-896d-c45efaa80aca is authoritative.

Kaimen Rigor review

Kaimen Rigor reviewed

3 / 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★
StatisticsStatistic did not reproduce−0.5★

Model-reported confidence: highReviewer agreement: highEngine: 7.39.0Source: full text

Biggest issues

01Significance claim does not survive recomputation
highOverreachrecomputedFull detail
Recomputed HR 0.32 (95% CI 0.12–0.87), reported p=0.089
“HR 0.32; 95% CI 0.12 to 0.87; P = 0.089”
Find in source

Findings by category

101ArithmeticNumerical inconsistencies 202OverreachOverstated conclusions 103AuthenticityData authenticity concerns 04DisclosureNone found 05AvailabilityNone found 06Copyedit6 minor suggestions

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 transparently reported phase 2 randomized trial with strong scientific premise, detailed methods, and comprehensive reporting of demographics, ethics, and data availability. The main weakness is a statistical inconsistency detected in one recomputed test (HR 0.32 with reported p=0.089 vs recomputed p=0.024), which warrants caution in interpreting that specific result, and the primary endpoint revision without sample size re-estimation.

Both reviewers agreed on study type (interventional) and on all dimensions except statistical analysis, where the deterministic statistics check revealed an inconsistency not flagged by either reviewer. The statistics check covered only 7 tests; the rest are unverified. The paper is a human clinical trial, so many checklist items (e.g., cell line authentication, housing conditions) are not applicable.

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 6 tests: 5 consistent, 1 inconsistent (1 change significance at p<.05); 5 recomputed directly from the reported test statistics, 1 via agent-written checks.

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

CONSISTENTreported p = .010 · recomputed p = .002
Recomputed hazard ratio 0.62 (95% CI 0.46–0.84), reported p=0.01
“hazard ratio 0.62; 95% confidence interval 0.46 to 0.84; P = 0.01”
Find in source
Taken as given: 0.46–0.84 is a two-sided 95% confidence interval for the hazard ratio of 0.62, not a range, an IQR, or a different interval level; the hazard ratio is a RATIO measure, so the interval is symmetric on the log scale; p=0.01 is the p for THIS estimate, not for another comparison in the same sentence
Method: back the two-tailed p out of the log-scale CI width and compare it against the printed p
How we recomputed it: pCI(0.62, 0.46, 0.84, 1)
CONSISTENTreported p = .010 · recomputed p = .002
Recomputed HR 0.62 (95% CI 0.46–0.84), reported p=0.01
“HR 0.62; 95% CI 0.46 to 0.84; P = 0.01”
Find in source
Taken as given: 0.46–0.84 is a two-sided 95% confidence interval for the HR of 0.62, not a range, an IQR, or a different interval level; the HR is a RATIO measure, so the interval is symmetric on the log scale; p=0.01 is the p for THIS estimate, not for another comparison in the same sentence
Method: back the two-tailed p out of the log-scale CI width and compare it against the printed p
How we recomputed it: pCI(0.62, 0.46, 0.84, 1)
UNCOMPUTABLEreported p = .022 · recomputed p = .007
Recomputed HR 0.68 (95% CI 0.51–0.89), reported p=0.022
“HR 0.68; 95% CI 0.51 to 0.89; P = 0.022”
Find in source
Taken as given: 0.51–0.89 is a two-sided 95% confidence interval for the HR of 0.68, not a range, an IQR, or a different interval level; the HR is a RATIO measure, so the interval is symmetric on the log scale; p=0.022 is the p for THIS estimate, not for another comparison in the same sentence
Method: back the two-tailed p out of the log-scale CI width and compare it against the printed p
How we recomputed it: pCI(0.68, 0.51, 0.89, 1)
CONSISTENTreported p = .008 · recomputed p = .006
Recomputed HR 0.62 (95% CI 0.44–0.87), reported p=0.008
“HR 0.62; 95% CI 0.44 to 0.87; P = 0.008”
Find in source
Taken as given: 0.44–0.87 is a two-sided 95% confidence interval for the HR of 0.62, not a range, an IQR, or a different interval level; the HR is a RATIO measure, so the interval is symmetric on the log scale; p=0.008 is the p for THIS estimate, not for another comparison in the same sentence
Method: back the two-tailed p out of the log-scale CI width and compare it against the printed p
How we recomputed it: pCI(0.62, 0.44, 0.87, 1)
CONSISTENTreported p = .624 · recomputed p = .467
Recomputed HR 0.90 (95% CI 0.68–1.20), reported p=0.624
“HR 0.90; 95% CI 0.68 to 1.20; P = 0.624”
Find in source
Taken as given: 0.68–1.20 is a two-sided 95% confidence interval for the HR of 0.90, not a range, an IQR, or a different interval level; the HR is a RATIO measure, so the interval is symmetric on the log scale; p=0.624 is the p for THIS estimate, not for another comparison in the same sentence
Method: back the two-tailed p out of the log-scale CI width and compare it against the printed p
How we recomputed it: pCI(0.9, 0.68, 1.2, 1)
CONSISTENTreported p = .624 · recomputed p = .467Reviewers 1, 2
Check the PFS hazard ratio p-value from the reported HR and 95% CI.
“HR 0.90 (0.68 to 1.20) | 0.624”
Find in source
Taken as given: The HR is 0.90 with 95% CI 0.68 to 1.20.; The CI is a 95% confidence interval for the hazard ratio.; The p-value is two-sided.
Method: Recomputed the two-sided p-value from the hazard ratio and its 95% confidence interval using the normal approximation for the log hazard ratio.
How we recomputed it: pCI(0.90, 0.68, 1.20, 1)

lowinternal contradiction
The abstract reports 1-year survival rates of 44.1% vs 22.3%, while the Discussion states 44.1% vs 22.3% but also mentions a 2.9-month improvement; these are consistent, but the HR discrepancy is noted.
Abstract: 'The 1-year survival rates were 44.1% versus 22.3%, respectively.' Discussion: 'The 1-year landmark survival rate was 44.1% with elraglusib/GnP versus 22.3% with GnP'
Discussionreviewer’s wording
lowinternal contradiction
The abstract reports a hazard ratio of 0.62 for OS, while the Discussion states HR 0.63 for the same comparison.
Abstract: 'hazard ratio 0.62; 95% confidence interval 0.46 to 0.84; P = 0.01' vs Discussion: 'HR 0.63; P = 0.01'
Discussionreviewer’s wording

Overstated conclusions

2 findings · worst high

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.

Significance claim flips when recomputedRecomputed
Conclusions only partially backed by the presented evidenceAssessed

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

MORE SIGNIFICANT ON RECHECKINCONSISTENTreported p = .089 · recomputed p = .024
Recomputed HR 0.32 (95% CI 0.12–0.87), reported p=0.089
Recomputing from the paper’s own numbers lands below p = 0.05 — more significant than the printed value. Usually benign (the reported figure is conservative), but the two don’t match.
“HR 0.32; 95% CI 0.12 to 0.87; P = 0.089”
Find in source
Taken as given: 0.12–0.87 is a two-sided 95% confidence interval for the HR of 0.32, not a range, an IQR, or a different interval level; the HR is a RATIO measure, so the interval is symmetric on the log scale; p=0.089 is the p for THIS estimate, not for another comparison in the same sentence
Method: back the two-tailed p out of the log-scale CI width and compare it against the printed p
How we recomputed it: pCI(0.32, 0.12, 0.87, 1)

8 major claims checked against the paper's own evidence: 2 only partially supported (evidence backs part of the claim; gaps or caveats remain); the rest adequately supported.

partialReviewers 1, 2
Baseline circulating immune-related factors (CXCL2 and TRAIL ligands) were associated with improved survival in the elraglusib/GnP arm.
The claim is partially supported: CXCL2 showed significant correlation, but TRAIL showed only numerical benefit without statistical significance.
Evidence: Correlative analysis section states high CXCL2 significantly correlated, TRAIL displayed numerical benefit without statistical significance.
“Explorative correlative analyses demonstrated that baseline circulating immune-related factors (that is, CXCL2 and TRAIL ligands) were associated with improved survival in the elraglusib/GnP arm.”
AbstractFind in source
partialReviewer 1
The OS benefit is a true treatment effect and not due to differential treatment duration or confounding.
The claim is supported by several exploratory analyses (landmark, exposure-stratified), but these are post-hoc and not definitive.
Evidence: Discussion describes landmark analysis and exposure-stratified analyses showing preserved benefit.
“Taken together, these exploratory analyses consistently demonstrate that the OS benefit observed with elraglusib/GnP cannot be attributed to differential treatment duration, cumulative dosing or survivor bias”
DiscussionFind in source
supportedReviewers 1, 2
Elraglusib/GnP improved median OS by 2.9 months and decreased the risk of death by 38% versus GnP in the mITT population.
The claim is directly supported by the reported median OS (10.1 vs 7.2 months) and HR 0.62 (95% CI 0.46-0.84, P=0.01).
Evidence: Table 2 and Results section report median OS 10.1 vs 7.2 months, HR 0.62, P=0.01.
“elraglusib/GnP improved median OS by 2.9 months and decreased the risk of death by 38% versus GnP (median OS 10.1 months versus 7.2 months, respectively (hazard ratio 0.62; 95% confidence interval 0.46 to 0.84; P = 0.01)).”
AbstractFind in source
supportedReviewer 1
The 1-year survival rate was higher with elraglusib/GnP (44.1%) than with GnP (22.3%).
The claim is supported by the reported landmark survival rates in the Results and Table 2.
Evidence: Results report 1-year landmark survival rates of 44.1% and 22.3%.
“The 1-year survival rates were 44.1% versus 22.3%, respectively.”
AbstractFind in source
supportedReviewers 1, 2
The safety profile of elraglusib/GnP was manageable.
The claim is supported by the reported safety data, which show expected toxicities (neutropenia, fatigue) but no unexpected safety signals.
Evidence: Safety section reports TEAEs and grade 3+ events; no new safety concerns beyond known effects.
“The safety profile of elraglusib/GnP was manageable.”
AbstractFind in source
supportedReviewers 1, 2
Treatment was accompanied by increases in intratumoral cytotoxic immune cell populations.
The claim is supported by the immunophenotyping data from paired tumor samples, though the sample size is small (n=6).
Evidence: Biomarker analysis reports increased CD8+, granzyme B+, and CD56+ cells in elraglusib/GnP-treated tumors.
“Treatment was accompanied by increases in intratumoral cytotoxic immune cell populations.”
AbstractFind in source
supportedReviewer 2
The 1-year survival rates were 44.1% versus 22.3%.
These rates are reported in Table 2 and are consistent with the survival analysis.
Evidence: Table 2: 12-month survival rate 44.1% vs 22.3%.
“The 1-year survival rates were 44.1% versus 22.3%, respectively.”
AbstractFind in source
supportedReviewer 2
These findings support the clinical activity of elraglusib/GnP as first-line treatment in mPDAC.
The OS benefit and manageable safety support clinical activity, though the trial is phase 2 and not definitive.
Evidence: OS benefit (HR 0.62) and safety profile.
“Together, these findings support the clinical activity of elraglusib/GnP as first-line treatment in mPDAC and provide a biological context for the observed survival benefit.”
AbstractFind in source

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

ADEQUATESurrogate endpoint
The primary endpoint is overall survival (OS), a hard clinical outcome. The paper reports a statistically significant improvement in median OS (10.1 vs 7.2 months, HR 0.62, P=0.01) and 1-year survival rate (44.1% vs 22.3%). No surrogate is used as the primary basis for the efficacy claim.
“elraglusib/GnP improved median OS by 2.9 months and decreased the risk of death by 38% versus GnP (median OS 10.1 months versus 7.2 months, respectively (hazard ratio 0.62; 95% confidence interval 0.46 to 0.84; P = 0.01)).”
Find in source
ADEQUATEEffect size
The effect size is a 2.9-month improvement in median OS (from 7.2 to 10.1 months) with a hazard ratio of 0.62, which is statistically significant (P=0.01). This is anchored to a hard clinical outcome (overall survival) and is clinically meaningful in the context of metastatic pancreatic cancer, where median OS improvements of this magnitude are considered significant. The 1-year survival rate also improved from 22.3% to 44.1%.
“elraglusib/GnP improved median OS by 2.9 months and decreased the risk of death by 38% versus GnP (median OS 10.1 months versus 7.2 months, respectively (hazard ratio 0.62; 95% confidence interval 0.46 to 0.84; P = 0.01)).”
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 primary endpoint was revised after study initiation without re-estimating sample size, which may affect the interpretation of the p-value.
“Following regulatory feedback after study initiation, the primary efficacy analysis was revised to a comparison of survival curves using a log-rank test; the sample size was not reestimated.”
Statistical analysisFind in source

Reporting gaps

None found

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.

Checked — nothing surfaced.

Eight Dimensions8 pass

The introduction cites extensive prior work on GSK-3β as a therapeutic target, including preclinical models and early-phase clinical trials of elraglusib, and acknowledges both strengths and limitations (e.g., toxicity of existing regimens, failure of immunotherapy). The rationale for the study follows logically from the cited evidence, and the paper addresses limitations of prior research by noting the need for new agents and the exploratory nature of some analyses.

Sub-criteria

prior work citedADEQUATE

premise rationaleADEQUATE

limitations addressedADEQUATE

Per reviewer (2) — the votes behind the verdict

PASSReviewer 1· 90% conf

The paper establishes a strong scientific premise by citing prior preclinical and clinical evidence for GSK-3β inhibition in pancreatic cancer and clearly links this to the study objectives.

Evidence

paraphrase[Introduction, paragraphs 2-4]

“Glycogen synthase kinase-3 beta (GSK-3β), a serine/threonine kinase, has been identified as a potential therapeutic target in human pancreatic cancer.”

“Because of the promising survival trends, this trial transitioned to a phase 2 randomized trial that also explored weekly dosing of elraglusib with GnP.”

Introduction ¶4Find in source

“Some concerns include toxicity with the FOLFIRINOX and NALIRIFOX regimens, limiting their use to patients with an Eastern Cooperative Oncology Group (ECOG) performance status (PS) of 0 and 1”

Introduction ¶1Find in source

PASSReviewer 2· 90% conf

The paper establishes a strong scientific premise by citing prior preclinical and clinical evidence for GSK-3β inhibition in pancreatic cancer and clearly links this to the study objectives.

Evidence

paraphrase[Introduction, paragraphs 3-4]

“In a single-arm, phase 2 trial in 42 patients with previously untreated mPDAC, elraglusib/GnP led to a median OS of 11.9 months (95% CI 7.8 to 16.5) and 1-year survival rate of 48% (95% CI 32.1 to 62.3) in the intention-to-treat (ITT) population.”

“Because of the promising survival trends, this trial transitioned to a phase 2 randomized trial that also explored weekly dosing of elraglusib with GnP.”

Introduction ¶4Find in source

paraphrase[Introduction, paragraph 4]

“Paraphrase: The paper states that the promising survival trends from the single-arm trial led to the design of a randomized trial, implying the need for a controlled comparison.”

Randomization was stratified by center with block size three, and the unit is the patient. The open-label design is explicitly acknowledged and justified in the discussion. A sample size calculation is reported with assumptions (1-year survival 55% vs 35%, 80% power, alpha 0.05). Inclusion/exclusion criteria are detailed. The mITT population is defined as the primary analysis set, addressing missing data. Blinding is not applicable due to open-label design, which is stated. Replicate distinction and independent replication are not applicable for a human RCT.

Sub-criteria

controlsN/A

power analysisADEQUATE

blinding levelsADEQUATE

outlier handlingADEQUATE

randomization unitADEQUATE

inclusion exclusionADEQUATE

randomization methodADEQUATE

replicate distinctionN/A

independent replicationN/A

Per reviewer (2) — the votes behind the verdict

PASSReviewer 1· 85% conf

The study design is rigorous for a phase 2 trial: randomization method and unit are described, the open-label design is stated with rationale, a power analysis is provided, and inclusion/exclusion criteria are pre-specified.

Evidence

“Randomization for all stages of the study was stratified according to study center with a block size of three.”

MethodsFind in source

“Initial limitations of this study were its open-label design and regional treatment access”

Discussion ¶1Find in source

“The calculation assumptions included achieving a 1-year survival rate of 55% with elraglusib/GnP and 35% with GnP (20% difference between the arms) with 80% power using a chi-squared test for equal proportions at a two-sided significance level of 0.05.”

Statistical analysisFind in source

PASSReviewer 2· 85% conf

The study design is rigorous for a phase 2 trial: randomization is described with stratification, the primary analysis population is defined, and a power calculation is provided, though the primary endpoint was revised after study initiation.

Evidence

“Randomization for all stages of the study was stratified according to study center with a block size of three.”

MethodsFind in source

Statistical analysisFind in source

“Following regulatory feedback after study initiation, the primary efficacy analysis was revised to a comparison of survival curves using a log-rank test; the sample size was not reestimated.”

Statistical analysisFind in source

The paper reports sex (male 52.4%), age (median 66 years), ethnicity (White 83.2%), and other demographics in Table 1. Since both sexes are enrolled, sex_justified is not applicable. Age, weight, and health status (ECOG PS) are reported. Species/strain and housing conditions are not applicable for a human trial.

Sub-criteria

demographicsADEQUATE

sex reportedADEQUATE

sex justifiedN/A

age weight healthADEQUATE

housing conditionsN/A

species strain sourceN/A

Per reviewer (2) — the votes behind the verdict

PASSReviewer 1· 90% conf

Demographic and baseline characteristics are comprehensively reported, including age, sex, ethnicity, and disease status.

Evidence

“The key characteristics of treated patients were a median age of 66 years (range 42–86 years), male sex in 52.4% of patients and White ethnicity in 83.2% of patients”

ResultsFind in source

paraphrase[Table 1]

“ECOG PS, n (%) | 0 | 64 (41.3) | 31 (39.7)”

PASSReviewer 2· 90% conf

Demographic and baseline characteristics are comprehensively reported for the human trial population, including age, sex, ethnicity, and disease status.

Evidence

“Female | 75 (48.4) | 35 (44.9)”

Table 1Find in source

“Median (range) | 65.0 (42.0 to 86.0) | 68.0 (42.0 to 85.0)”

Table 1Find in source

paraphrase[Table 1]

“ECOG PS, n (%) | 0 | 64 (41.3) | 31 (39.7)”

The paper states that local institutional review boards or independent ethics committees approved the protocol, and all patients signed informed consent. Regulatory compliance with Good Clinical Practice and the Declaration of Helsinki is stated. The approval statement names the type of body (IRB/ethics committee) but does not provide a protocol number; however, this is adequate for a clinical trial report.

Sub-criteria

iacuc statementN/A

informed consentADEQUATE

irb ethics statementADEQUATE

regulatory complianceADEQUATE

Per reviewer (2) — the votes behind the verdict

PASSReviewer 1· 85% conf

Ethics approval and informed consent are adequately reported, with a named ethics body and compliance statement.

Evidence

“Local institutional review boards or the independent ethics committees of the participating centers approved the study protocol.”

EthicsFind in source

“All enrolled patients signed the informed consent form.”

EthicsFind in source

“The study was performed in compliance with Good Clinical Practice and the Declaration of Helsinki (v2013).”

EthicsFind in source

PASSReviewer 2· 90% conf

Ethics approval and informed consent are adequately reported, with a named ethics body and compliance with the Declaration of Helsinki.

Evidence

“Local institutional review boards or the independent ethics committees of the participating centers approved the study protocol.”

EthicsFind in source

“All enrolled patients signed the informed consent form.”

EthicsFind in source

“The study was performed in compliance with Good Clinical Practice and the Declaration of Helsinki (v2013).”

EthicsFind in source

Elraglusib is identified as a GSK-3β inhibitor with dose (9.3 mg/kg) and schedule. Gemcitabine and nab-paclitaxel are identified with doses. The manufacturer/source of elraglusib is implied (Actuate Therapeutics) but not explicitly stated in the methods; however, the drug is well-defined. Statistical software (SAS v9.4) is identified. Antibodies, cell lines, and mycoplasma testing are not applicable as this is a clinical trial without wet-lab assays.

Sub-criteria

mycoplasma testingN/A

reagents identifiedADEQUATE

organisms identifiedN/A

antibodies identifiedN/A

cell line authenticationN/A

software tools identifiedADEQUATE

Per reviewer (2) — the votes behind the verdict

PASSReviewer 1· 80% conf

The investigational product (elraglusib) and comparator (gemcitabine/nab-paclitaxel) are identified with dosing details, and the statistical software is named.

Evidence

“During the study, intravenous elraglusib 9.3 mg kg −1 was administered on day 1 for the once-weekly schedule or days 1 and 4 for the twice-weekly schedule.”

MethodsFind in source

“All statistical analyses were performed using SAS v.9.4”

Statistical analysisFind in source

PASSReviewer 2· 85% conf

The investigational product (elraglusib) and comparator chemotherapy are identified with dosing details, and the statistical software is named.

Evidence

“During the study, intravenous elraglusib 9.3 mg kg −1 was administered on day 1 for the once-weekly schedule or days 1 and 4 for the twice-weekly schedule.”

MethodsFind in source

“GnP treatment consisted of intravenous nab-paclitaxel 125 mg m − 2 followed by intravenous gemcitabine 1,000 mg m 2 on days 1, 8 and 15 in a 28-day cycle.”

MethodsFind in source

“All statistical analyses were performed using SAS v.9.4”

Statistical analysisFind in source

Tests are named (log-rank, Cox proportional hazards, Cochran–Mantel–Haenszel). Exact p-values are reported (e.g., P = 0.01). Effect sizes with 95% CIs are reported for primary and secondary endpoints. Statistical software is identified. Data presentation includes Kaplan-Meier curves and forest plots, appropriate for a trial. Assumptions are handled by design (Cox model, log-rank). Mathematical plausibility is not applicable for large-N continuous outcomes.

Sub-criteria

tests namedADEQUATE

exact p valuesADEQUATE

effect sizes ciADEQUATE

data presentationADEQUATE

software identifiedADEQUATE

assumptions verifiedADEQUATE

mathematical plausibilityN/A

Per reviewer (2) — the votes behind the verdict

PASSReviewer 1· 80% conf

Statistical methods are well described, with named tests, exact p-values, effect sizes with CIs, and software identified. The analysis is appropriate for a clinical trial.

Evidence

“Statistical tests: log-rank tests for time-to-event variables; Cox proportional hazards model for HRs; and Cochran–Mantel–Haenszel general association tests for responder variables.”

Table 2Find in source

“HR 0.62; 95% CI 0.46 to 0.84; P = 0.01”

ResultsFind in source

“median OS of 10.1 months (95% CI 7.7 to 12.5) with elraglusib/GnP versus 7.2 months (95% CI 5.7 to 9.0) with GnP (HR 0.62; 95% CI 0.46 to 0.84; P = 0.01”

ResultsFind in source

PASSReviewer 2· 80% conf

Statistical methods are appropriate and well-described, with exact p-values and effect sizes with confidence intervals reported; however, the primary endpoint revision and some subgroup analyses warrant caution.

Evidence

“Statistical tests: log-rank tests for time-to-event variables; Cox proportional hazards model for HRs; and Cochran–Mantel–Haenszel general association tests for responder variables.”

Table 2Find in source

“HR 0.62; 95% CI 0.46 to 0.84; P = 0.01”

ResultsFind in source

“HR 0.62 (0.46 to 0.84)”

Table 2Find in source

The data availability statement describes a managed-access process through ClinicalStudyDataRequest and a contact email, with a response timeframe. This is reported_and_adequate for identifiable patient data. Repository deposit and accession numbers are not applicable for patient-level data. Code sharing is not applicable as no bespoke code is mentioned.

Sub-criteria

code sharingN/A

accession numbersN/A

repository depositN/A

data availability statementADEQUATE

Per reviewer (2) — the votes behind the verdict

PASSReviewer 1· 85% conf

A data availability statement is present with a concrete access mechanism via ClinicalStudyDataRequest, which is adequate for patient-level data.

Evidence

“Trial data availability is according to the criteria and process described at www.clinicalstudydatarequest.com”

Data availabilityFind in source

PASSReviewer 2· 90% conf

A data availability statement is provided with a concrete access mechanism via a data-sharing platform, which is adequate for patient-level clinical trial data.

Evidence

paraphrase[Data availability]

“Trial data availability is according to the criteria and process described at www.clinicalstudydatarequest.com . The data described in this publication will be made available for any qualified request.”

The trial is registered (NCT03678883). Methods are detailed enough for replication. Limitations are explicitly discussed (open-label, early mortality, exploratory analyses). Conclusions are proportional, acknowledging the phase 2 nature and exploratory findings. Funding and competing interests are disclosed. A reporting guideline is not explicitly mentioned, but the paper follows CONSORT-like structure with a flow diagram.

Sub-criteria

funding coiADEQUATE

trial registrationADEQUATE

reporting guidelineINADEQUATE

methods completenessADEQUATE

all outcomes reportedADEQUATE

limitations discussedADEQUATE

conclusions proportionalADEQUATE

Per reviewer (2) — the votes behind the verdict

PASSReviewer 1· 85% conf

The manuscript is transparent: trial registration is provided, methods are detailed, limitations are discussed, and conclusions are proportional.

Evidence

“ClinicalTrials.gov registration: NCT03678883”

AbstractFind in source

“Initial limitations of this study were its open-label design and regional treatment access”

Discussion ¶1Find in source

“This work was supported by Actuate Therapeutics.”

Competing interestsFind in source

PASSReviewer 2· 85% conf

The manuscript is transparent: methods are detailed, the trial is registered, limitations are discussed, and funding/COI are disclosed.

Evidence

“ClinicalTrials.gov registration: NCT03678883”

AbstractFind in source

“Initial limitations of this study were its open-label design and regional treatment access as several patients withdrew from the study once randomized to the control arm to pursue alternative treatment options closer to home.”

Discussion ¶10Find in source

“This work was supported by Actuate Therapeutics.”

Competing interestsFind in source

Registered (3 IDs: ClinicalTrials.gov). No reporting guideline cited.

ClinicalTrials.gov

NCT03678883

ClinicalTrials.gov

NCT05077800

ClinicalTrials.gov

NCT06896188

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 54 references by DOI: 3 verified — 51 no DOI (shown, not verified).

NO DOI
Overview of pancreatic cancer epidemiology in Europe and recommendations for screening in high-risk populations
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Pancreatic Adenocarcinoma (Version 2.2025)
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Epidemiology of pancreatic cancer: global trends, etiology and risk factors
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Treatment landscape of metastatic pancreatic cancer
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Pancreatic cancer: ESMO Clinical Practice Guideline for diagnosis, treatment and follow-up
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
NALIRIFOX, FOLFIRINOX, and gemcitabine with nab-paclitaxel as first-line chemotherapy for metastatic pancreatic cancer: a systematic review and meta-analysis
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
FOLFIRINOX versus gemcitabine for metastatic pancreatic cancer
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
NALIRIFOX versus nab-paclitaxel and gemcitabine in treatment-naive patients with metastatic pancreatic ductal adenocarcinoma (NAPOLI 3): a randomised, open-label, phase 3 trial
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Real-world validation of the purity independent subtyping of tumors classifier for informing therapy selection in pancreatic ductal adenocarcinoma
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Real-world clinical outcomes and economic burden of metastatic pancreatic ductal adenocarcinoma: a systematic review
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Increased survival in pancreatic cancer with nab-paclitaxel plus gemcitabine
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Early results of the PASS-01 trial: pancreatic adenocarcinoma signature stratification for treatment-01
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Immunologic strategies in pancreatic cancer: making cold tumors hot
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Safety and activity of anti-PD-L1 antibody in patients with advanced cancer
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Phase 2 trial of single agent ipilimumab (anti-CTLA-4) for locally advanced or metastatic pancreatic adenocarcinoma
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Durvalumab with or without tremelimumab for patients with metastatic pancreatic ductal adenocarcinoma: a phase 2 randomized clinical trial
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
The CCTG PA.7 phase II trial of gemcitabine and nab-paclitaxel with or without durvalumab and tremelimumab as initial therapy in metastatic pancreatic ductal adenocarcinoma
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Glycogen synthase kinase-3β participates in nuclear factor κB-mediated gene transcription and cell survival in pancreatic cancer cells
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Glycogen synthase kinase-3 inhibition sensitizes pancreatic cancer cells to chemotherapy by abrogating the TopBP1/ATR-mediated DNA damage response
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Glycogen synthase kinase-3β inhibition with 9-ING-41 attenuates the progression of pulmonary fibrosis
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
GSK-3β in pancreatic cancer: spotlight on 9-ING-41, its therapeutic potential and immune modulatory properties
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Molecular pathways: revisiting glycogen synthase kinase-3β as a target for the treatment of cancer
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
From a natural product lead to the identification of potent and selective benzofuran-3-yl-(indol-3-yl)maleimides as glycogen synthase kinase 3beta inhibitors that suppress proliferation and survival of pancreatic cancer cells
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Microfluidic organoid cultures derived from pancreatic cancer biopsies for personalized testing of chemotherapy and immunotherapy
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Phase I study of elraglusib (9-ING-41), a glycogen synthase kinase-3β inhibitor, as monotherapy or combined with chemotherapy in patients with advanced malignancies
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Phase II study of elraglusib (9-ING-41), a GSK-3β inhibitor, in combination with gemcitabine plus nab-paclitaxel in previously untreated metastatic pancreatic cancer
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Machine learning and statistical prediction of overall survival (OS) from pre-dose plasma biomarkers in a randomized phase 2 trial (1801 Part 3B) of the GSK-3 inhibitor elraglusib in metastatic pancreatic ductal adenocarcinoma (mPDAC): Application toward patient enrichment
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Elraglusib (9-ING-41), a selective small-molecule inhibitor of glycogen synthase kinase-3 beta, reduces expression of immune checkpoint molecules PD-1, TIGIT and LAG-3 and enhances CD8 + T cell cytolytic killing of melanoma cells
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Glycogen synthase kinase-3 (GSK-3) regulation of inhibitory coreceptor expression in T-cell immunity
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
GSK3 inhibition drives maturation of NK cells and enhances their antitumor activity
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Exploration of modified progression-free survival as a novel surrogate endpoint for overall survival in immuno-oncology trials
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Mechanism-based treatment of cancer with immune checkpoint inhibitor therapies
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Evaluation of overall response rate and progression-free survival as potential surrogate endpoints for overall survival in immunotherapy trials
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Evaluation of objective response, disease control and progression-free survival as surrogate end-points for overall survival in anti-programmed death-1 and anti-programmed death ligand 1 trials
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Irreconcilable differences: the divorce between response rates, progression-free survival, and overall survival
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Glycogen synthase kinase-3 is an in vivo regulator of hematopoietic stem cell repopulation
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
How lithium treatment generates neutrophilia by enhancing phosphorylation of GSK-3, increasing HIF-1 levels and how this path is important during engraftment
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Current and future immunotherapeutic approaches in pancreatic cancer treatment
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Molecular pathways: targeting tumor-infiltrating myeloid-derived suppressor cells for cancer therapy
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Myeloid cell mediated immune suppression in pancreatic cancer
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Myeloid-derived suppressor cells in cancer: current knowledge and future perspectives
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Myeloid-derived suppressor cells in the era of increasing myeloid cell diversity
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Targeting myeloid-derived suppressor cells for cancer therapy
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
CXCL2: a key player in the tumor microenvironment and inflammatory diseases
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Barriers and opportunities in pancreatic cancer immunotherapy
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Updates in immunotherapy for pancreatic cancer
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Abstract C073: FOLFIRINOX with glycogen synthase kinase-3 beta (GSK-3β) inhibitor elraglusib and transforming growth factor-β (TGFβ) inhibitor losartan in untreated metastatic pancreatic ductal adenocarcinoma (PDAC): interim analysis of safety cohort
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
A Bayesian pick-the-winner design in a randomized phase II clinical trial
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
New response evaluation criteria in solid tumours: revised RECIST guideline (version 1.1)
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Using modified intention-to-treat as a principal stratum estimator for failure to initiate treatment
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
The add-arm design for unimodal response curve with unknown mode
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).

4 of 5 data/code links checked; 4 live; 1 not probed.

datahttps://clinicaltrials.gov/study/NCT03678883LIVEHTTP 200
Resolves, but the content could not be matched to the paper.
datahttps://clinicaltrials.gov/ct2/show/NCT03678883LIVEHTTP 200
Resolves, but the content could not be matched to the paper.
datahttps://clinicaltrials.gov/ct2/show/NCT05077800LIVEHTTP 200
Resolves, but the content could not be matched to the paper.
datahttps://clinicaltrials.gov/ct2/show/NCT06896188LIVEHTTP 200
Resolves, but the content could not be matched to the paper.
datahttp://www.clinicalstudydatarequest.comUNVERIFIED
Liveness indeterminate — content not checked.

Copyediting

6 minor

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

No major wording or formatting errors. 6 minor suggestions below.

6 copyedit issues flagged: mostly consistency, clarity, other.

MINORconsistencyAbstract
“hazard ratio 0.62; 95% confidence interval 0.46 to 0.84; P = 0.01”
→ Ensure consistent use of 'P' vs 'p' throughout the manuscript.
The abstract uses uppercase 'P' while some tables use lowercase 'p'.
MINORclarityResults, Patient disposition
“The presented results for elraglusib/GnP focus on the once-weekly dosing of elraglusib; patients from the selection and confirmation stages are included.”
→ Clarify whether the twice-weekly arm is excluded from all efficacy analyses or only from the primary analysis.
The sentence is slightly ambiguous about the inclusion of the twice-weekly arm.
MINORotherMethods, Statistical analysis
“The log-rank test was selected to appropriately account for censored survival times occurring before 1 year.”
→ Consider adding a reference for this statement.
The rationale for log-rank test is given but no citation is provided.
MINORtypoAbstract
“elraglusib/GnP improved median OS by 2.9 months and decreased the risk of death by 38% versus GnP (median OS 10.1 months versus 7.2 months, respectively (hazard ratio 0.62; 95% confidence interval 0.46 to 0.84; P = 0.01)).”
→ Consider rephrasing to avoid double parentheses.
Minor punctuation issue.
MINORconsistencyDiscussion
“HR 0.63; P = 0.01”
→ Ensure HR is consistent with the Results (HR 0.62).
HR reported as 0.63 in Discussion vs 0.62 in Results; likely rounding but should be consistent.
MINORclarityMethods, Statistical analysis
“The log-rank test was selected to appropriately account for censored survival times occurring before 1 year.”
→ Clarify that the log-rank test compares survival curves, not just 1-year rates.
Slightly ambiguous wording.

Kaimen Rigor Action ItemsFinal pass before journal submission

The published work is largely robust, but the statistical inconsistency in one reported p-value (HR 0.32, reported p=0.089 vs recomputed p=0.024) is a validity concern that warrants an erratum or independent re-analysis. An informed reader should weigh this alongside the primary endpoint revision without sample size re-estimation when interpreting the results.

1.
HIGHstatistics
Investigate and correct the inconsistent p-value for the HR 0.32 (95% CI 0.12–0.87) result: the reported p=0.089 does not match the recomputed p=0.024; verify the source data and issue a correction if needed.
A demonstrable inconsistency between the reported p-value and the recomputed value is a validity threat that could mislead readers about the significance of that finding.
2.
HIGHstatistics
In the Methods (Statistical analysis), explicitly discuss the impact of revising the primary endpoint from 1-year survival to log-rank test without re-estimating sample size, and state the potential reduction in power.
The primary endpoint revision without sample size re-estimation is a major methodological concern that affects the interpretation of the primary p-value.
3.
HIGHreporting
Add an explicit reference to the CONSORT checklist in the Methods or Reporting Summary, and include a completed CONSORT flow diagram in the main text.
Explicit adherence to a reporting guideline strengthens transparency and is expected for randomized trials.
4.
HIGHreporting
Reconcile the HR discrepancy between the Results (HR 0.62) and the Discussion (HR 0.63) for the same OS comparison, and ensure consistent reporting throughout.
Internal inconsistency in a key effect size undermines reader confidence and may indicate a rounding or transcription error.
5.
MEDIUMethics
In the Methods (Ethics approval and consent), provide the specific IRB/ethics committee names and protocol approval numbers for each participating center.
Full auditability of ethics approvals is expected for clinical trials and facilitates verification.
6.
MEDIUMreporting
In the Methods (Statistical analysis), describe how the proportional hazards assumption was checked for the Cox model, or state that it was assumed.
Verifying model assumptions is part of sound statistical reporting and helps readers assess the validity of the HR estimates.
7.
MEDIUMreporting
In the Results (Patient disposition), clarify whether the twice-weekly elraglusib arm is excluded from all efficacy analyses or only from the primary analysis.
The current wording is ambiguous about the inclusion of the twice-weekly arm, which could confuse readers about the analysis population.
8.
MEDIUMreporting
In the Discussion, explicitly state that the exploratory biomarker analyses are hypothesis-generating and were not adjusted for multiple comparisons.
Clarifying the exploratory nature of these analyses prevents overinterpretation of potentially spurious findings.
9.
MEDIUMdata code
In the Data Availability section, provide a direct hyperlink to the ClinicalStudyDataRequest portal and specify the expected timeline for data availability.
A direct link and timeline improve the usability of the data access statement for qualified researchers.
10.
MEDIUMreporting
In the Results, report the number of patients who withdrew after randomization and were lost to follow-up, to clarify the mITT vs ITT difference.
Detailed disposition data enhance transparency and allow readers to assess potential attrition bias.
11.
LOWcopyedit
Standardize the use of 'P' vs 'p' for p-values throughout the manuscript (Abstract uses uppercase, tables use lowercase).
Consistent notation is a basic copyedit requirement that improves professionalism.
12.
LOWcopyedit
Rephrase the Abstract sentence with double parentheses to avoid punctuation clutter.
Minor punctuation issues detract from readability.
13.
LOWreporting
In the Methods (Statistical analysis), add a citation for the rationale for using the log-rank test to account for censored survival times.
Citing a standard reference for the log-rank test strengthens the methodological justification.
14.
LOWreporting
In the Methods (Procedures), specify the manufacturer/source of elraglusib and the chemotherapy agents.
Identifying the source of investigational products is part of complete resource reporting.
15.
LOWreporting
In the Abstract, add confidence intervals for the 1-year survival rates to be consistent with the main text.
Consistency between abstract and main text improves clarity and completeness.

Full text (74,187 chars)Analyzed: Aug 15, 2026

RigorReviewer 1

RigorReviewer 2

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.

Elraglusib and chemotherapy in metastatic pancreatic ductal adenocarcinoma: a randomized controlled phase 2 trial.

DOI

10.1038/s41591-026-04327-4

Record issued

2026-08-15

Engine

7.39.0

Exported

2026-09-22