Alpha¹ by KaimenConfidential

Kaimen Rigor review

A generative AI-discovered TNIK inhibitor for idiopathic pulmonary fibrosis: a randomized phase 2a trial

Xu Z, Ren F, Wang P, Cao J, Tan C, Ma D, Zhao L, Dai J, Ding Y, Fang H, Li H, Liu H, Luo F, Meng Y, Pan P, Xiang P, Xiao Z, Rao S, Satler C, Liu S, Lv Y, Zhao H, Chen S, Cui H, Korzinkin M, Gennert D, Zhavoronkov A.

DOI: 10.1038/s41591-025-03743-2
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/dc17aebb-3ae8-4e41-8ad2-ec092c41534f is authoritative.

Kaimen Rigor review

Kaimen Rigor reviewed

1 / 5

Critical — serious problems across many checks

Interventional Study

Critical

High-priority

How this rating was calculated

Started at 5★ — deductions

IntegrityIntegrity concern ×6−3★
ClaimsOverstated claim ×3−1.5★
ReportingStatistical analysis not met−0.5★
ReportingKey resources partially met−0.25★

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

Not verified in this report

References were not verified against Crossref/OpenAlex.
No reported statistical tests were found to recompute.

Biggest issues

01Statistical reporting inadequate
highDisclosurereviewer-assessedFull detail
Statistical tests are named and effect estimates are reported with 95% CIs, but the paper contains demonstrable arithmetic/percentage errors in the TEAE reporting, and several p-values are reported only as thresholds.
“3 in 60 mg QD (20.4%)”
ResultsFind in source
02Internal contradictions in the reported numbers
mediumArithmeticreviewer-assessedFull detail
ALT-increase counts in the AE narrative contradict Table 2: the text lists ALT increase as '6 in 60 mg QD (33.3%) and 3 in 60 mg QD (16.7%)' (and lists 60 mg QD twice), while Table 2 shows ALT increase of 0, 0, 1, 0 across the four arms.
“alanine aminotransferase (ALT) increase (1 in placebo (5.9%), 1 in 30 mg QD (5.6%), 1 in 30 mg BID (5.6%), 6 in 60 mg QD (33.3%) and 3 in 60 mg QD (16.7%))”
Table 2Find in source
03Internal contradictions in the reported numbers
mediumArithmeticreviewer-assessedFull detail
The text reports two different numbers for ALT increase in the 60 mg QD group (6 and 3), while Table 2 shows only 1 case. This is a clear internal inconsistency that needs correction.
“alanine aminotransferase (ALT) increase (1 in placebo (5.9%), 1 in 30 mg QD (5.6%), 1 in 30 mg BID (5.6%), 6 in 60 mg QD (33.3%) and 3 in 60 mg QD (16.7%)).”
ResultsFind in source
04Internal contradictions in the reported numbers
mediumArithmeticreviewer-assessedFull detail
The ALT increase sentence lists two separate counts for the same 60 mg QD arm (6 events and 3 events), which is internally inconsistent as written.
ALT increase (1 in placebo (5.9%), 1 in 30 mg QD (5.6%), 1 in 30 mg BID (5.6%), 6 in 60 mg QD (33.3%) and 3 in 60 mg QD (16.7%))
Resultsreviewer’s wording
05Mathematically impossible statistic
mediumArithmeticreviewer-assessedFull detail
The reported percentage 20.4% for 3 of 18 patients in the 60 mg QD group is arithmetically impossible (3/18 = 16.7%).
“3 in 60 mg QD (20.4%)”
ResultsFind in source
06Conclusion reaches beyond the evidence
mediumOverreachreviewer-assessedFull detail
The AI-discovered agent is 'safe and effective' for IPF.
“Preliminary results from a phase 2a trial involving 71 patients suggest that a new agent, discovered and designed with artificial intelligence assistance, is safe and effective for the treatment of idiopathic pulmonary fibrosis.”
AbstractFind in source

3 further findings of this severity or below — every one is in the sections below, filed under its error type.

Findings by category

201ArithmeticNumerical inconsistencies 202OverreachOverstated conclusions 103AuthenticityData authenticity concerns 204DisclosureReporting gaps 05AvailabilityNone found 106CopyeditCopyediting

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

Automated narrative synthesis was unavailable for this run; this verdict is the majority vote of the independent reviewers, with their own suggestions listed as action items.

Fallback synthesis: the narrative synthesizer did not return a usable result, so the report is assembled directly from the reviewers.

Numerical inconsistencies

2 findings · worst medium

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
Mathematically impossible statisticAssessed

mediuminternal contradiction
ALT-increase counts in the AE narrative contradict Table 2: the text lists ALT increase as '6 in 60 mg QD (33.3%) and 3 in 60 mg QD (16.7%)' (and lists 60 mg QD twice), while Table 2 shows ALT increase of 0, 0, 1, 0 across the four arms.
“alanine aminotransferase (ALT) increase (1 in placebo (5.9%), 1 in 30 mg QD (5.6%), 1 in 30 mg BID (5.6%), 6 in 60 mg QD (33.3%) and 3 in 60 mg QD (16.7%))”
Table 2Find in source
mediuminternal contradiction
The text reports two different numbers for ALT increase in the 60 mg QD group (6 and 3), while Table 2 shows only 1 case. This is a clear internal inconsistency that needs correction.
“alanine aminotransferase (ALT) increase (1 in placebo (5.9%), 1 in 30 mg QD (5.6%), 1 in 30 mg BID (5.6%), 6 in 60 mg QD (33.3%) and 3 in 60 mg QD (16.7%)).”
ResultsFind in source
mediuminternal contradiction
The ALT increase sentence lists two separate counts for the same 60 mg QD arm (6 events and 3 events), which is internally inconsistent as written.
ALT increase (1 in placebo (5.9%), 1 in 30 mg QD (5.6%), 1 in 30 mg BID (5.6%), 6 in 60 mg QD (33.3%) and 3 in 60 mg QD (16.7%))
Resultsreviewer’s wording
mediumimpossible statistic
The reported percentage 20.4% for 3 of 18 patients in the 60 mg QD group is arithmetically impossible (3/18 = 16.7%).
“3 in 60 mg QD (20.4%)”
ResultsFind in source
lowimpossible statistic
The hypokalemia count/percentage pair '3 in 60 mg QD (20.4%)' is arithmetically inconsistent: 3/18 = 16.7%, so 20.4% cannot be derived from the stated count and group size.
“hypokalemia (2 in placebo (11.8%), 3 in 30 mg QD (16.7%), 5 in 30 mg BID (27.8%) and 3 in 60 mg QD (20.4%))”
ResultsFind in source

Overstated conclusions

2 findings · worst medium

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

Conclusions overstated beyond the evidenceAssessed
Conclusions only partially backed by the presented evidenceAssessed

12 major claims checked against the paper's own evidence: 3 not fully backed by the presented evidence (unsupported or overstated).

overstatedReviewer 1
The AI-discovered agent is 'safe and effective' for IPF.
Safety is supported, but 'effective' overreaches: the primary endpoint was safety, and the efficacy signal was exploratory/secondary with no confirmatory between-group p-value.
Evidence: No primary efficacy endpoint tested with a p-value; FVC improvement reported only as a per-group estimate with overlapping CIs.
“Preliminary results from a phase 2a trial involving 71 patients suggest that a new agent, discovered and designed with artificial intelligence assistance, is safe and effective for the treatment of idiopathic pulmonary fibrosis.”
AbstractFind in source
overstatedReviewer 2
Rentosertib is 'safe and effective for the treatment of idiopathic pulmonary fibrosis'.
Safety is supported, but 'effective' overreaches: the FVC improvement at 60 mg QD is a secondary endpoint showing a trend (+98.4 ml, CI 10.9-185.9 vs placebo -20.3 ml) with no reported between-group significance, in a small phase 2a cohort.
Evidence: Safety data (TEAE rates similar across arms) plus a secondary FVC trend at the highest dose; no primary-efficacy endpoint or significance test for FVC is reported.
“is safe and effective for the treatment of idiopathic pulmonary fibrosis”
AbstractFind in source
overstatedReviewer 3
The study represents a revolutionary shift in the streamlining of drug discovery.
The paper presents a single example of AI-discovered drug reaching phase 2a, which is promising but not yet a 'revolutionary shift' as it has not completed phase 3 or been approved.
Evidence: The claim is made in the introduction but not supported by the phase 2a results alone.
Our generative AI-powered approach streamlined preclinical candidate nomination to a mere 18 months and completion of phase 0/1 clinical testing to under 30 months from the initiation of target discovery, representing a revolutionary shift in the streamlining of drug discovery.
Introductionreviewer’s wording
partialReviewer 1
60 mg rentosertib QD increased FVC compared with placebo.
The 60 mg arm showed a within-group mean FVC increase of +98.4 ml with 95% CI excluding zero, but the between-group difference vs placebo was not statistically tested and the CIs overlap.
Evidence: Mean change +98.4 ml (95% CI 10.9 to 185.9) for 60 mg QD vs −20.3 ml (95% CI −116.1 to 75.6) for placebo; no p-value for the comparison.
“We observed increased forced vital capacity at the highest dosage with a mean change of +98.4 ml (95% confidence interval 10.9 to 185.9) for patients in the 60 mg rentosertib QD group, compared with −20.3 ml (95% confidence interval −116.1 to 75.6) for the placebo group.”
AbstractFind in source
partialReviewer 1
60 mg QD without SOC antifibrotic therapy significantly improved FVC.
The subgroup mean change has a CI excluding zero, but the subgroup analysis is post-hoc, small, and lacks multiplicity control, so 'significant' is not fully established.
Evidence: Subgroup result: +187.8 ml, 95% CI 68.6 to 306.9 ml.
“Patients receiving 60 mg rentosertib QD not concurrently taking SOC antifibrotic therapy exhibited significant improvement in FVC (+187.8 ml, 95% CI 68.6 to 306.9 ml)”
ResultsFind in source
partialReviewer 2
Downregulation of the IPF-associated serum protein profile supports the hypothesis that inhibition of TNIK modulates IPF pathophysiological pathways.
The proteomic data show dose- and time-dependent downregulation of fibrosis-associated proteins, but the inference that this is specifically due to TNIK inhibition is indirect (no direct TNIK-modulation readout in patients).
Evidence: Differential protein abundance analyses, ECM-organization pathway enrichment, and inverse correlations with FVC change (Fig. 4, Extended Data Figs. 6-8).
“Downregulation of this IPF-associated protein profile supports the hypothesis that inhibition of TNIK modulates IPF pathophysiological pathways”
Discussion ¶5Find in source
partialReviewer 3
Downregulation of fibrosis-associated proteins supports that TNIK inhibition modulates IPF pathways.
The proteomic data show correlation but not causation; the mechanistic link is inferred from pathway enrichment and prior knowledge.
Evidence: Results: 'Pathway enrichment analysis identified extracellular matrix organization as the most downregulated Reactome pathway gene set' and correlation with FVC change.
“Downregulated proteins associated with 30 mg BID and 60 mg QD treatment include known fibrosis-associated proteins such as MMP10, PTPRZ1, COL1A1, FAP, FN1, ROBO2, ASPN and LTBP2”
ResultsFind in source
supportedReviewer 1
Rentosertib is safe and well tolerated in patients with IPF.
The primary safety endpoint (TEAE rates) was similar across arms, and treatment-related SAEs were low, directly supporting the claim.
Evidence: Primary endpoint TEAE rates 72.2%, 83.3%, 83.3%, and 70.6% across arms; low treatment-related SAE rates.
“These results suggest that targeting TNIK with rentosertib is safe and well tolerated and warrants further investigation in larger-scale clinical trials of longer duration.”
AbstractFind in source
supportedReviewer 1
Downregulation of IPF-associated proteins supports TNIK pathway modulation.
The proteomic data directly show dose- and time-dependent downregulation of fibrosis-associated proteins and ECM pathway enrichment, supporting the mechanistic claim.
Evidence: Olink proteomics: 22 high-confidence differentially abundant proteins at 60 mg QD; COL1A1, FAP, FN1, MMP10 decrease with dose/time; Reactome ECM organization enrichment.
“Downregulation of this IPF-associated protein profile supports the hypothesis that inhibition of TNIK modulates IPF pathophysiological pathways and points to a potential serum protein signature as a biomarker for response to rentosertib treatment.”
DiscussionFind in source
supportedReviewer 2
Targeting TNIK with rentosertib is safe and well tolerated and warrants further investigation.
The safety/tolerability data (TEAE rates and SAE rates comparable across arms, detailed AE tables) adequately back this claim, which is also appropriately hedged.
Evidence: Primary safety endpoint (TEAE percentages across arms) and treatment-related SAE rates reported in Results and Table 2.
“These results suggest that targeting TNIK with rentosertib is safe and well tolerated and warrants further investigation in larger-scale clinical trials of longer duration.”
AbstractFind in source
supportedReviewer 2
Treatment with 60 mg rentosertib QD over 12 weeks was associated with a trend toward an increase in FVC.
The reported +98.4 ml change (95% CI 10.9 to 185.9) at 60 mg QD versus -20.3 ml in placebo supports a trend claim, and the paper labels it a trend rather than proof of efficacy.
Evidence: FVC change with 95% CI for the 60 mg QD and placebo groups in Results and Fig. 2.
“Treatment with 60 mg rentosertib QD over 12 weeks was associated with a trend toward an increase in FVC in patients with IPF.”
Discussion ¶2Find in source
supportedReviewer 2
This is the first phase 2a trial of an AI-discovered/AI-designed drug, and the first report of AI-enabled discovery of both a target and a compound.
The paper consistently frames rentosertib as first-in-class AI-generated and cites its own prior work; the novelty claim is presented as the study's premise and is internally consistent (though not independently verifiable here).
Evidence: Background statements and citation of prior phase 0/1 results (NCT05154240, CTR20221542).
“This was also, importantly, the first reported instance of AI platform-enabled discovery of both a disease-associated target and a compound for that target.”
Main, paragraph 2Find in source

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

ADEQUATESurrogate endpoint
Primary efficacy endpoint is FVC, a validated clinical outcome in IPF, not a surrogate biomarker.
“FVC is the gold-standard metric for assessing the lung function of patients with IPF and response to therapeutic intervention”
Find in source
ADEQUATEEffect size
Reported mean FVC increase of +98.4 ml (95% CI 10.9 to 185.9) in 60 mg QD group, percentage change of 2.82% meeting MCID of 2–6%.
“Patients receiving 60 mg rentosertib QD experienced a mean increase in FVC percentage change of 2.82% by our modeling, meeting the minimal clinically important difference (MCID) reported for FVC in IPF of 2–6%”
Find in source

Data authenticity concerns

1 finding · worst medium

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

6 integrity concerns flagged (0 high).

mediumother
Trial NCT05938920 was first submitted to ClinicalTrials.gov on 2023-06-28, after the registered study start date of 2023-06-19. Retrospective registration means the protocol and outcomes were not on the public record before the study ran, which is what prospective registration exists to establish.
NCT05938920
reviewer’s wording

Reporting gaps

2 findings · worst high

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.

Statistical reporting inadequateAssessed
Key resources under-identified (antibodies, cell lines, RRIDs)Assessed

Eight Dimensions6 pass · 1 warn · 1 fail

The introduction cites prior studies on TNIK as a target for IPF, AI-driven drug discovery, and limitations of current therapies. It acknowledges that no AI-discovered drug has progressed through phase 3, providing a clear rationale for the trial. The hypothesis that TNIK inhibition may treat IPF follows from the cited evidence.

Sub-criteria

prior work citedADEQUATE

premise rationaleADEQUATE

limitations addressedADEQUATE

Per reviewer (3) — the votes behind the verdict

PASSReviewer 1· 85% conf

The paper clearly establishes the scientific premise: AI-driven drug discovery has few clinical successes, IPF lacks disease-modifying therapies, and the authors previously identified TNIK via AI and advanced rentosertib through phase 0/1 trials.

Evidence

“Despite these advancements, few AI-discovered or AI-designed drugs have reached clinical trials.”

Main, paragraph 2Find in source

“Clinical trials investigating these two drugs in patients with IPF have demonstrated only a slowing of disease progression”

Main, paragraph 4Find in source

paraphrase[Main, paragraph 2]

“The question of whether AI can impart meaningful, sustained disruption to drug development has remained unanswered.”

PASSReviewer 2· 85% conf

The paper establishes a strong scientific foundation, citing the slow/costly drug-development context, prior AI-drug failures, IPF disease burden, and prior phase 0/1 results of rentosertib.

Evidence

“Clinical trials investigating these two drugs in patients with IPF have demonstrated only a slowing of disease progression”

Main, paragraph 1-3Find in source

“AI-discovered drugs have experienced similar levels of phase 2 trial failure as non-AI-discovered drugs”

Main, paragraph 1Find in source

“We reported the development and positive phase 0 and 1 trial results of rentosertib”

Main, paragraph 2Find in source

PASSReviewer 3· 95% conf

The paper clearly establishes its scientific foundation by citing prior work on AI drug discovery, TNIK biology, and IPF, and provides a logical rationale for the study.

Evidence

“Despite these advancements, few AI-discovered or AI-designed drugs have reached clinical trials. AI-discovered drugs have experienced similar levels of phase 2 trial failure as non-AI-discovered drugs”

Introduction ¶1Find in source

paraphrase[Introduction, paragraph 4]

“Our group used generative AI-driven discovery tools to identify Traf2- and Nck-interacting kinase (TNIK) de novo as a critical regulator of idiopathic pulmonary fibrosis (IPF) pathology”

paraphrase[Introduction, paragraph 3]

“AI-discovered drugs have experienced similar levels of phase 2 trial failure as non-AI-discovered drugs, and none has so far progressed through phase 3 trials”

Randomization was performed via interactive response technology in a 1:1:1:1 ratio. The study is double-blind, with blinding of subjects, investigators, and staff. Inclusion and exclusion criteria are detailed. Missing data were handled by multiple imputation. A sample size calculation was provided for safety (90% probability to detect an AE with 15% rate) but not for efficacy endpoints.

Sub-criteria

controlsN/A

power analysisINADEQUATE

blinding levelsADEQUATE

outlier handlingADEQUATE

randomization unitADEQUATE

inclusion exclusionADEQUATE

randomization methodADEQUATE

replicate distinctionN/A

independent replicationN/A

Per reviewer (3) — the votes behind the verdict

PASSReviewer 1· 80% conf

Core trial design elements are present: randomized 1:1:1:1 allocation via interactive response technology, double-blinding, detailed inclusion/exclusion criteria, and missing-data handling. Gaps include lack of randomization generation details and absence of a power-based sample size calculation.

Evidence

“Adults with IPF were randomly assigned in a 1:1:1:1 ratio via interactive response technology to receive oral rentosertib”

MethodsFind in source

“Subjects, investigators, site study staff, reviewers and everyone involved in study conduct or analysis were blinded with regard to the randomized treatment assignments until after data freeze.”

MethodsFind in source

“although a sample size calculation based on statistical power considerations was not performed”

Statistical analysisFind in source

PASSReviewer 2· 80% conf

A well-described randomized, double-blind, placebo-controlled design with detailed inclusion/exclusion criteria, blinding, ITT/missing-data handling, and a stated (feasibility-based) sample-size rationale; no formal power calculation was performed.

Evidence

“Adults with IPF were randomly assigned in a 1:1:1:1 ratio via interactive response technology”

MethodsFind in source

“Subjects, investigators, site study staff, reviewers and everyone involved in study conduct or analysis were blinded with regard to the randomized treatment assignments until after data freeze.”

MethodsFind in source

“although a sample size calculation based on statistical power considerations was not performed”

Statistical analysisFind in source

PASSReviewer 3· 90% conf

The study design is rigorous for a phase 2a trial, with clear randomization, blinding, and pre-specified criteria, though the sample size calculation was based on safety rather than efficacy.

Evidence

“Adults with IPF were randomly assigned in a 1:1:1:1 ratio via interactive response technology”

MethodsFind in source

“Subjects, investigators, site study staff, reviewers and everyone involved in study conduct or analysis were blinded with regard to the randomized treatment assignments until after data freeze”

MethodsFind in source

“Given an approximate sample size of 15 subjects per treatment arm, there exists a 90% probability of observing at least one AE if the true population rate is approximately 15%, which was sufficient to assess the feasibility of safety parameters, although a sample size calculation based on statistical power considerations was not performed.”

Statistical analysisFind in source

Table 1 reports sex, age, weight, BMI, smoking history, and baseline lung function (FVC, DLCO, LCQ) for all treatment groups. All participants are male (90.1%) and Asian, but this is a single-country study and does not require justification for a single-sex study as it is not by design.

Sub-criteria

demographicsADEQUATE

sex reportedADEQUATE

sex justifiedN/A

age weight healthADEQUATE

housing conditionsN/A

species strain sourceN/A

Per reviewer (3) — the votes behind the verdict

PASSReviewer 1· 90% conf

Baseline characteristics are well reported: sex, age, weight, BMI, race, smoking history, and antifibrotic use. Both sexes were enrolled, so a single-sex justification is not required.

Evidence

“Male sex – no. (%) | 18 (100.0) | 15 (83.3) | 16 (88.9) | 15 (88.2) | 64 (90.1)”

Table 1Find in source

“Age – years, mean ± s.d. | 65.8 ± 7.0 | 67.2 ± 7.8 | 65.7 ± 6.8 | 68.3 ± 5.3 | 66.7 ± 6.7”

Table 1Find in source

paraphrase[Table 1]

“Race – no. (%) | Asian | 18 (100.0) | 18 (100.0) | 18 (100.0) | 17 (100.0) | 71 (100.0)”

PASSReviewer 2· 85% conf

Baseline characteristics including sex, age, weight, BMI, race, smoking history and comorbid antifibrotic use are reported in a baseline table for a mixed-sex (predominantly male) cohort.

Evidence

“Male sex – no. (%) | 18 (100.0) | 15 (83.3) | 16 (88.9) | 15 (88.2) | 64 (90.1)”

Table 1Find in source

paraphrase[Table 1]

“Current antifibrotic treatment – no. (%) | Nintedanib | 6 (33.3) | 4 (22.2) | 9 (50.0) | 3 (17.6) | 22 (31.0)”

PASSReviewer 3· 95% conf

Biological variables including sex, age, weight, and baseline lung function are well-reported in Table 1.

Evidence

“Male sex – no. (%) | 18 (100.0) | 15 (83.3) | 16 (88.9) | 15 (88.2) | 64 (90.1)”

Table 1Find in source

“Age – years, mean ± s.d. | 65.8 ± 7.0 | 67.2 ± 7.8 | 65.7 ± 6.8 | 68.3 ± 5.3 | 66.7 ± 6.7”

Table 1Find in source

paraphrase[Table 1]

“FVC | Mean – ml, mean ± s.d. | 2,696 ± 444.2 | 2,617 ± 667.4 | 2,762 ± 684.2 | 2,246 ± 473.6 | 2,585 ± 600.4”

The paper states that the institutional review board or ethics committee at participating centers approved the protocols, and all patients provided written informed consent. The trial was conducted in accordance with the Declaration of Helsinki and ICH-GCP guidelines.

Sub-criteria

iacuc statementN/A

informed consentADEQUATE

irb ethics statementADEQUATE

regulatory complianceADEQUATE

Per reviewer (3) — the votes behind the verdict

WARNReviewer 1· 85% conf

Informed consent and regulatory compliance are adequately described, but the ethics approval statement does not name a specific IRB/ethics committee or provide protocol approval numbers, making it a generic rather than fully auditable statement.

Evidence

“The institutional review board or ethics committee at participating centers approved protocols and adhered to local laws before initiation of the clinical trial.”

MethodsFind in source

“All patients in this study provided written informed consent.”

MethodsFind in source

“The trial was conducted following the principles outlined in the Declaration of Helsinki and the International Council for Harmonization guidelines for Good Clinical Practice.”

MethodsFind in source

PASSReviewer 2· 80% conf

The trial reports IRB/ethics committee approval, written informed consent from all patients, and compliance with the Declaration of Helsinki and ICH-GCP.

Evidence

“The institutional review board or ethics committee at participating centers approved protocols and adhered to local laws before initiation of the clinical trial.”

MethodsFind in source

“All patients in this study provided written informed consent.”

MethodsFind in source

“The trial was conducted following the principles outlined in the Declaration of Helsinki and the International Council for Harmonization guidelines for Good Clinical Practice.”

MethodsFind in source

PASSReviewer 3· 95% conf

Ethical approvals are adequately reported, including IRB approval, written informed consent, and compliance with Declaration of Helsinki and ICH-GCP.

Evidence

“The institutional review board or ethics committee at participating centers approved protocols and adhered to local laws before initiation of the clinical trial.”

MethodsFind in source

“All patients in this study provided written informed consent.”

MethodsFind in source

“The trial was conducted following the principles outlined in the Declaration of Helsinki and the International Council for Harmonization guidelines for Good Clinical Practice.”

MethodsFind in source

Rentosertib is named with dose and regimen, and provided by Insilico Medicine. However, statistical software such as lme4 and clusterProfiler are mentioned without version numbers. SpiroSphere devices are identified but no version. Other resources like Olink panel are identified but not with catalog numbers. The drug is adequately identified, but software identification is incomplete.

Sub-criteria

mycoplasma testingN/A

reagents identifiedADEQUATE

organisms identifiedN/A

antibodies identifiedN/A

cell line authenticationN/A

software tools identifiedINADEQUATE

Per reviewer (3) — the votes behind the verdict

WARNReviewer 1· 70% conf

The investigational product is adequately identified, but statistical software/tools are named without versions, and no formulation/lot details for the drug are given.

Evidence

“The study treatment rentosertib was provided by Insilico Medicine or a designated contract research organization, packaged and labeled in accordance with the principles of Good Manufacturing Practice.”

MethodsFind in source

“The lme4 package was used for the generalized linear mixed model analysis.”

Statistical analysisFind in source

“using the Olink Explore 3072 panel”

ResultsFind in source

PASSReviewer 2· 70% conf

The investigational product (rentosertib) is named with manufacturer and dose regimen, and analysis software/custom code are identified; bench-science criteria are not applicable to this clinical trial.

Evidence

“The study treatment rentosertib was provided by Insilico Medicine or a designated contract research organization”

MethodsFind in source

“Custom code used to analyze Olink proteomics data is available via GitHub at https://github.com/HUICUI1992/Code-for-NM”

Data availabilityFind in source

WARNReviewer 3· 85% conf

The investigational product (rentosertib) is adequately identified with dose and regimen, but software tools are named without version numbers, which is a reporting gap.

Evidence

“oral rentosertib (at a dose of 30 mg (QD), 30 mg (BID) or 60 mg (QD))”

MethodsFind in source

“The lme4 package was used for the generalized linear mixed model analysis.”

Statistical analysisFind in source

ANCOVA, MMRM, Spearman, Pearson, paired t-test, and generalized linear mixed models are named; effect sizes are reported with CIs. However, two TEAE percentage errors are present: 3/18 is reported as 20.4% (actual 16.7%), and the ALT increase listing includes two separate entries for the same 60 mg QD arm (6 events at 33.3% and 3 events at 16.7%). These are demonstrable numeric errors, warranting a fail despite otherwise adequate reporting. Software versions are absent, and many p-values are given as thresholds ('P < 0.05').

Sub-criteria

tests namedADEQUATE

exact p valuesINADEQUATE

effect sizes ciADEQUATE

data presentationADEQUATE

software identifiedINADEQUATE

assumptions verifiedADEQUATE

mathematical plausibilityFAIL

Per reviewer (3) — the votes behind the verdict

FAILReviewer 1· 90% conf

Statistical tests are named and effect estimates are reported with 95% CIs, but the paper contains demonstrable arithmetic/percentage errors in the TEAE reporting, and several p-values are reported only as thresholds.

Evidence

“3 in 60 mg QD (20.4%)”

ResultsFind in source

paraphrase[Results, Primary safety endpoint]

“ALT increase (1 in placebo (5.9%), 1 in 30 mg QD (5.6%), 1 in 30 mg BID (5.6%), 6 in 60 mg QD (33.3%) and 3 in 60 mg QD (16.7%))”

“two-sided P = 0.0495”

ResultsFind in source

WARNReviewer 2· 75% conf

Statistical tests are well named and effect sizes are reported with 95% CIs, but several p-values are printed only as thresholds, software versions are omitted, and minor arithmetic inconsistencies appear in the AE text.

Evidence

“patients receiving 60 mg rentosertib QD showed improved FVC, with and +98.4 ml (95% CI 10.9 to 185.9)”

ResultsFind in source

“1, 8 and 22 high-confidence ( P adj < 0.05) differentially abundant proteins”

ResultsFind in source

“hypokalemia (2 in placebo (11.8%), 3 in 30 mg QD (16.7%), 5 in 30 mg BID (27.8%) and 3 in 60 mg QD (20.4%))”

ResultsFind in source

PASSReviewer 3· 90% conf

Statistical tests are named, assumptions are addressed by design, effect sizes with CIs are reported, and data presentation is appropriate for a clinical trial. Software versions are not reported.

Evidence

paraphrase[Methods, Statistical analysis]

“An analysis of covariance (ANCOVA) model was performed for endpoints including FVC. ... LCQ scores were analyzed using a mixed model for repeated measures.”

“our modeling indicated a significant increase of least-squares mean change of the LCQ scores of patients receiving 60 mg rentosertib compared with those receiving placebo (two-sided P = 0.0495).”

ResultsFind in source

“patients receiving 60 mg rentosertib QD showed improved FVC, with and +98.4 ml (95% CI 10.9 to 185.9)”

ResultsFind in source

The data availability statement names a repository and accession, describes a managed-access route for the Olink data, and states that the protocol and SAP are available on request. The custom code used for Olink proteomics analysis is available via a public GitHub repository. All applicable sub-criteria are adequate.

Sub-criteria

code sharingADEQUATE

accession numbersADEQUATE

repository depositADEQUATE

data availability statementADEQUATE

Per reviewer (3) — the votes behind the verdict

PASSReviewer 1· 95% conf

Data and code availability are fully and concretely described: deidentified Olink proteomics and FVC data are deposited in OMIX with accession OMIX008341, and custom analysis code is on GitHub.

Evidence

“Complete deidentified Olink proteomics data, and FVC data, have been deposited at the OMIX database under accession codes at OMIX008341 ( https://ngdc.cncb.ac.cn/omix/release/OMIX008341 ).”

Data availabilityFind in source

“Custom code used to analyze Olink proteomics data is available via GitHub at https://github.com/HUICUI1992/Code-for-NM .”

Data availabilityFind in source

“Study protocol and statistical analysis plan will be provided in a secure data sharing environment upon academic or research request.”

Data availabilityFind in source

PASSReviewer 2· 90% conf

Deidentified proteomics and FVC data are deposited in the OMIX database with an accession, and custom analysis code is shared on GitHub; the protocol/SAP are available on request.

Evidence

“Complete deidentified Olink proteomics data, and FVC data, have been deposited at the OMIX database under accession codes at OMIX008341 ( https://ngdc.cncb.ac.cn/omix/release/OMIX008341 )”

Data availabilityFind in source

“Custom code used to analyze Olink proteomics data is available via GitHub at https://github.com/HUICUI1992/Code-for-NM”

Data availabilityFind in source

PASSReviewer 3· 95% conf

Data and code availability are well-documented: proteomics and FVC data deposited at OMIX with accession, and custom code on GitHub.

Evidence

paraphrase[Data availability]

“Complete deidentified Olink proteomics data, and FVC data, have been deposited at the OMIX database under accession codes at OMIX008341 (https://ngdc.cncb.ac.cn/omix/release/OMIX008341).”

“Custom code used to analyze Olink proteomics data is available via GitHub at https://github.com/HUICUI1992/Code-for-NM”

Data availabilityFind in source

Methods are detailed enough for replication. The trial is registered at ClinicalTrials.gov (NCT05938920). All primary and secondary outcomes are reported. Limitations are discussed (small cohort, homogeneity, short follow-up). Conclusions are proportional. Funding and competing interests are disclosed. However, no explicit statement of adherence to CONSORT or other reporting guidelines is provided, only a vague reference to a 'Reporting summary'.

Sub-criteria

funding coiADEQUATE

trial registrationADEQUATE

reporting guidelineINADEQUATE

methods completenessADEQUATE

all outcomes reportedADEQUATE

limitations discussedADEQUATE

conclusions proportionalADEQUATE

Per reviewer (3) — the votes behind the verdict

WARNReviewer 1· 80% conf

Methods are detailed, the trial is registered, most outcomes are reported, limitations are discussed, and COI are declared. However, the reporting guideline reference is generic, funding/grant numbers are absent, and the abstract's 'safe and effective' wording overstates the exploratory efficacy signal.

Evidence

“ClinicalTrials.gov registration number: NCT05938920”

AbstractFind in source

“Further information on research design is available in the linked to this article.”

AbstractFind in source

“Preliminary results from a phase 2a trial involving 71 patients suggest that a new agent, discovered and designed with artificial intelligence assistance, is safe and effective for the treatment of idiopathic pulmonary fibrosis.”

AbstractFind in source

PASSReviewer 2· 80% conf

Methods are detailed, the trial is registered (NCT05938920), a reporting summary is referenced, negative results are reported, limitations are discussed, and funding/COI are declared; the abstract's 'safe and effective' wording slightly overreaches the evidence.

Evidence

“ClinicalTrials.gov registration number: NCT05938920”

AbstractFind in source

“The limitations of this study include the small cohort size of each arm, the geographical and demographic homogeneity of the participants (all were residents of China of similar race) and a short period of follow-up”

DiscussionFind in source

“is safe and effective for the treatment of idiopathic pulmonary fibrosis”

AbstractFind in source

PASSReviewer 3· 85% conf

The manuscript is well-structured with comprehensive methods, trial registration, and limitations discussion, but does not explicitly reference a reporting guideline checklist.

Evidence

“ClinicalTrials.gov registration number: NCT05938920 (https://clinicaltrials.gov/study/NCT05938920)”

AbstractFind in source

DiscussionFind in source

“Reporting summary Further information on research design is available in the linked to this article.”

AbstractFind in source

Registered (2 IDs: ClinicalTrials.gov). No reporting guideline cited.

ClinicalTrials.gov

NCT05938920

ClinicalTrials.gov

NCT05154240

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.

6 data/code links checked; 6 live.

datahttps://clinicaltrials.gov/study/NCT05938920LIVEHTTP 200
Resolves, but the content could not be matched to the paper.
datahttps://ngdc.cncb.ac.cn/omix/release/OMIX008341LIVEHTTP 200
Resolves, but the content could not be matched to the paper.
datahttps://ngdc.cncb.ac.cn/omix/LIVEHTTP 200
Resolves, but the content could not be matched to the paper.
datahttps://clinicaltrials.gov/ct2/show/NCT05154240LIVEHTTP 200
Resolves, but the content could not be matched to the paper.
datahttp://www.chinadrugtrials.org.cn/LIVEHTTP 202
Resolves, but the content could not be matched to the paper.
codeGitHubLIVEHTTP 200
https://github.com/HUICUI1992/Code-for-NM
Resolves to GitHub (code repository).

Copyediting

1 finding · worst low

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

Wording or formatting errors that need correctingAssessed

13 copyedit issues flagged (3 major): mostly consistency, grammar, clarity.

MAJORconsistencyResults, Primary safety endpoint
“3 in 60 mg QD (20.4%)”
→ Change to '3 in 60 mg QD (16.7%)'
3/18 = 16.7%, not 20.4%.
MAJORconsistencyResults, Primary safety endpoint
“ALT increase (1 in placebo (5.9%), 1 in 30 mg QD (5.6%), 1 in 30 mg BID (5.6%), 6 in 60 mg QD (33.3%) and 3 in 60 mg QD (16.7%))”
→ Reconcile the duplicate '60 mg QD' entries; one count likely belongs to a different group or subcategory.
The same arm appears twice with two different event counts.
MAJORconsistencyResults, Primary safety endpoint vs Table 2
“alanine aminotransferase (ALT) increase (1 in placebo (5.9%), 1 in 30 mg QD (5.6%), 1 in 30 mg BID (5.6%), 6 in 60 mg QD (33.3%) and 3 in 60 mg QD (16.7%))”
→ Reconcile narrative counts with Table 2 ALT-increase row (0, 0, 1, 0) and remove the duplicated 60 mg QD entry
Text and Table 2 disagree; '60 mg QD' appears twice for ALT increase.
MINORgrammarResults, Secondary lung function endpoints
“with and +98.4 ml”
→ Delete the stray 'and'.
Typographical error.
MINORgrammarResults, Pharmacokinetics
“indicates an increased exposure to rentosertib with increases time on treatment”
→ Change to 'with increased time on treatment'.
Grammar error.
MINORclarityResults, Secondary lung function endpoints
“with and +98.4 ml (95% CI 10.9 to 185.9)”
→ Remove 'and' -> 'with +98.4 ml (95% CI 10.9 to 185.9)'
Typo: 'with and'.
MINORgrammarResults, Pharmacokinetics
“indicates an increased exposure to rentosertib with increases time on treatment”
→ Change to 'with increased time on treatment'
Ungrammatical phrase.
MINORconsistencyResults, Primary safety endpoint
“3 in 60 mg QD (20.4%)”
→ Change percentage to 16.7% (3/18)
3/18 = 16.7%, not 20.4%.
MINORconsistencyExtended Data Fig. 5
“n = 10 at week 10 in 60 QD”
→ Change to 'n = 10 at week 12 in 60 QD'
Timepoint inconsistency appearing in both panels a and b.
MINORconsistencyResults, secondary lung function endpoints
“forced expiry in 1 s”
→ Change to 'forced expiratory volume in 1 second' (FEV1) for consistency with standard terminology.
The term 'forced expiry' is used once; elsewhere 'FEV1' is used correctly.
MINORclarityResults, secondary lung function endpoints
“patients receiving 60 mg rentosertib QD not concurrently taking SOC antifibrotic therapy exhibited significant improvement”
→ Add 'who were' before 'not concurrently' for clarity: 'patients receiving 60 mg rentosertib QD who were not concurrently taking SOC antifibrotic therapy'.
Minor grammatical improvement.
MINORotherTable 2 and Results text
“ALT increase: 6 in 60 mg QD (33.3%) and 3 in 60 mg QD (16.7%)”
→ Verify and correct the numbers for ALT increase in the 60 mg QD group; the text gives two different counts (6 and 3) for the same group, while the table shows 1. This is a data inconsistency that needs resolution.
See integrity_concerns for more details.
MINORpunctuationResults, secondary lung function endpoints
“with and +98.4 ml”
→ Remove 'and': 'with +98.4 ml' or 'with an increase of +98.4 ml'.
Typo: 'with and' should be 'with'.

Kaimen Rigor Action ItemsFinal pass before journal submission

1.
MEDIUMrigor
Correct the TEAE percentage error in Results: change '3 in 60 mg QD (20.4%)' to '3 in 60 mg QD (16.7%)' (3/18 = 16.7%).
2.
MEDIUMrigor
Reconcile the ALT increase listing in Results: '6 in 60 mg QD (33.3%) and 3 in 60 mg QD (16.7%)' duplicates the same arm; determine which count belongs to another group or clarify the subcategories.
3.
MEDIUMrigor
Name the central/lead IRB and each participating center's ethics committee, and include protocol approval numbers, in Methods so the ethics approval is auditable.
4.
MEDIUMrigor
State the randomization sequence generation method (e.g., computer-generated random numbers, block/stratification) in Methods, in addition to 'interactive response technology'.
5.
MEDIUMrigor
Replace the safety-feasibility sample size note with a formal power/sample-size statement for the primary safety endpoint, including alpha and power, or clearly label it as a feasibility calculation in the protocol.
6.
MEDIUMrigor
Report exact p-values or effect estimates with CIs for all inferential tests, including the Spearman/Pearson correlations in figure legends, instead of 'P < 0.05' thresholds.
7.
MEDIUMrigor
Add versions for R, lme4, clusterProfiler, and the Olink panel in Methods, and provide a software environment file (renv/conda) in the GitHub repository.
8.
MEDIUMrigor
Explicitly cite the CONSORT extension for pilot/phase 2 trials in the reporting summary, rather than a generic 'Reporting Summary' link.
9.
MEDIUMrigor
Add a Funding statement with grant numbers and sponsor support in the Acknowledgements/Funding section; competing interests are already declared.
10.
MEDIUMrigor
Tone down the abstract's 'safe and effective' to 'safe and well tolerated, with a preliminary efficacy signal', since the primary endpoint was safety and the FVC result was exploratory.
11.
MEDIUMrigor
Add a formal a priori power/sample-size calculation (with effect size, alpha, power) for the primary safety endpoint, or state explicitly that the trial is feasibility-only; edit the Methods 'Statistical analysis' and 'Study design' sections.
12.
MEDIUMrigor
Report exact p-values (2-3 significant figures) instead of thresholds such as 'P < 0.05' and 'P adj < 0.05' throughout the Results (proteomics and correlation analyses).
13.
MEDIUMrigor
Resolve the contradiction in ALT-increase counts between the Results narrative ('6 in 60 mg QD (33.3%) and 3 in 60 mg QD (16.7%)') and Table 2 (ALT increase row: 0, 0, 1, 0); correct the duplicated 60 mg QD entry.
14.
MEDIUMrigor
Correct the hypokalemia percentage in the Results AE narrative: '3 in 60 mg QD (20.4%)' should be 16.7% (3/18).
15.
MEDIUMrigor
Specify versions of the statistical software/packages used (R, lme4, clusterProfiler, lm) in the Methods 'Statistical analysis' section.
16.
MEDIUMrigor
Softening the abstract's 'safe and effective' to 'safe and potentially effective' to match the phase 2a secondary-endpoint trend evidence.
17.
MEDIUMrigor
Fix the typo 'with and +98.4 ml' in the Results 'Secondary lung function endpoints' section.
18.
MEDIUMrigor
Correct 'n = 10 at week 10 in 60 QD' to 'week 12' in Extended Data Fig. 5 (appears twice).
19.
MEDIUMrigor
Add version numbers for all software (R, lme4, clusterProfiler, etc.) in the Methods section to improve key_resources reporting.
20.
MEDIUMrigor
Explicitly state adherence to the CONSORT 2010 statement for randomized trials and provide a completed checklist as supplementary material to enhance reporting_transparency.
21.
MEDIUMrigor
Provide a formal sample size calculation for efficacy endpoints (e.g., FVC) based on expected effect size and power, not just for safety, to strengthen study_design.
22.
MEDIUMrigor
Clarify the potential impact of partial unblinding by bioanalytics staff who could identify placebo samples; discuss how this was managed to avoid bias.
23.
MEDIUMrigor
Deposit the study protocol and statistical analysis plan in a public repository (e.g., ClinicalTrials.gov or a data repository) rather than only on request, to improve data availability.
24.
MEDIUMrigor
In the limitations, explicitly discuss the generalizability of findings given the single-country, all-Asian, predominantly male cohort, and the potential impact on external validity.
25.
MEDIUMrigor
For the claim of 'revolutionary shift' in drug discovery, consider toning down to avoid overstatement, as a single successful phase 2a trial does not yet constitute a paradigm shift.

Full text (54,469 chars)Analyzed: Aug 10, 2026

RigorReviewer 1

RigorReviewer 2

RigorReviewer 3

RigorPre-Submission Reviewer

CopyeditAgent Scribe

CitationsCitation Integrity

StatisticsStatistics Consistency

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

PreregistrationDeterministic check· no model (rule-based)

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

How stars are deducted

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

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

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

The eight reporting dimensions

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

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

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

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

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

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

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

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

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

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

A generative AI-discovered TNIK inhibitor for idiopathic pulmonary fibrosis: a randomized phase 2a trial

DOI

10.1038/s41591-025-03743-2

Record issued

2026-08-10

Engine

7.29.0

Exported

2026-09-21