Verified Rigor RecordRigor Reviewed

Rigor-Reviewed Manuscript

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

Name: Verified Rigor Record — A generative AI-discovered TNIK inhibitor for idiopathic pulmonary fibrosis: a randomized phase 2a trial
Item: A generative AI-discovered TNIK inhibitor for idiopathic pulmonary fibrosis: a randomized phase 2a trial
Rating: 88
Author: Αlpha¹

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.

Paper source

Open DOI

Αlpha¹ Editorial

This rigor review was examined and confirmed by Αlpha¹ Editorial · July 4, 2026

Αlpha¹ Rigor Review

Rigor Reviewed

WARN

Warnings to address · Interventional Study

Eight dimensions — click a segment for details

Critical

High-priority

Model-reported confidence: moderateReviewer agreement: lowEngine: 7.4.0Source: full text

This Αlpha¹ Rigor Review uses AI 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 phase 2a trial is well-designed with clear reporting of methods, ethics, and data availability. However, it contains a critical internal inconsistency in the safety data (ALT increase counts) and inadequate specification of key resources (drug formulation, software versions). The abstract overstates efficacy for a small safety study.

Independent rigor evaluations from three reviewers were synthesized; the copyedit and integrity checks confirmed the critical inconsistency. Statistics recomputation covered 5 tests, all consistent. Two references were not found in the registry (potential fabrication signals). The paper is published, so this is a post-publication audit.

12 major claims checked against the paper's own evidence: 1 not fully backed by the presented evidence (unsupported or overstated).

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

4 integrity concerns flagged (0 high).

Rigor

Eight Dimensions6 pass · 1 warn · 1 fail

12 major claims1 overstated4 partial7 supported

Prior work on AI-driven drug discovery and TNIK is extensively cited (refs 1–21). The rationale linking TNIK to IPF and the need for a phase 2a trial is logically presented. Limitations of prior therapies (only slowing progression) and of the current study (small sample, short duration) are discussed. All three sub-criteria are adequately addressed.

Sub-criteria

prior work citedADEQUATE

premise rationaleADEQUATE

limitations addressedADEQUATE

Per reviewer (3) — the votes behind the verdict

PASSReviewer 1· 95% conf

The paper clearly establishes its scientific foundation, citing prior work on AI drug discovery, TNIK target identification, and IPF pathophysiology, including strengths and limitations of prior research.

Evidence

direct quote[Introduction, paragraphs 1-3]

“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 as it orchestrates multiple profibrotic and proinflammatory cellular programs.”

direct quote[Introduction, paragraph 4]

“IPF is an age-related, progressive lung disease... The median survival time after diagnosis is 2–4 years despite widespread availability of targeted antifibrotic therapies , , highlighting the ongoing need for novel, effective therapies to treat IPF.”

direct quote[Discussion, paragraph 7]

“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, which limit the assessment of long-term safety and efficacy.”

PASSReviewer 2· 95% conf

The paper thoroughly establishes the scientific premise, citing prior work on AI drug discovery, the TNIK target, the IPF disease landscape, and early-phase trial results, with a clear rationale and acknowledged limitations.

Evidence

direct quote[Introduction, paragraphs 1-3]

“Advances in artificial intelligence (AI) have demonstrated proof-of-concept innovation and potential to accelerate and decrease costs across all stages of pharmaceutical research and development”

direct quote[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”

direct quote[Discussion, paragraph 7]

“The limitations of this study include the small cohort size of each arm, the geographical and demographic homogeneity of the participants… and a short period of follow-up”

PASSReviewer 3· 95% conf

The paper establishes a clear scientific foundation, citing prior work on AI drug discovery, TNIK biology, and IPF pathophysiology, and logically links the premise to the study objectives.

Evidence

direct quote[Introduction, paragraphs 1-3]

“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...”

direct quote[Introduction, paragraph 2]

“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.”

paraphrase[Introduction, paragraph 3]

“The authors state that TNIK orchestrates profibrotic and proinflammatory programs, and that rentosertib is the first antifibrotic small-molecule TNIK inhibitor, filling a gap in IPF treatment.”

Randomization method (interactive response technology, 1:1:1:1 ratio), blinding (double-blind, investigators and subjects blinded), and inclusion/exclusion criteria are well described. The sample size justification (90% probability of detecting an AE at 15% rate) is provided, though a formal power calculation for efficacy was not performed. Missing data are handled via multiple imputation (MAR assumption). The ITT analysis population is defined.

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· 90% conf

The study design is robust: multicenter, double-blind, randomized with interactive response technology, clear inclusion/exclusion criteria, and appropriate handling of missing data. Sample size justification prioritizes safety feasibility over efficacy power, which is acceptable for a phase 2a trial.

Evidence

direct quote[Methods, Study design]

“This phase 2a, multicenter, double-blind, placebo-controlled, randomized, multidose trial of rentosertib in adults with IPF... Adults with IPF were randomly assigned in a 1:1:1:1 ratio via interactive response technology to receive oral rentosertib... or placebo... 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.”

direct quote[Methods, Participants]

“Eligible patients were 40 years of age or older with an IPF diagnosis... Eligible patients had to meet all three of the following criteria during the screening visit: (1) FVC >40% predicted of normal, (2) DLCO corrected for hemoglobin ≥25% and <80% predicted of normal, and (3) FEV1/FVC ratio >0.7...”

direct quote[Methods, Statistical analysis]

“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.”

PASSReviewer 2· 90% conf

The study design is generally rigorous with randomization, blinding, and pre-specified inclusion/exclusion criteria; however, the sample size justification is based on safety feasibility rather than a formal power calculation for efficacy, which is a minor gap.

Evidence

direct quote[Methods, Study design]

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

direct quote[Methods, Study design]

“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”

direct quote[Results, Statistical analysis]

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

WARNReviewer 3· 90% conf

The design is a standard phase 2a RCT with adequate randomization, blinding, and inclusion/exclusion criteria, but is underpowered for efficacy endpoints, lacks an a priori sample size calculation for superiority, and does not specify a pre-planned analysis for multiplicity or primary vs. secondary hypothesis testing.

Evidence

direct quote[Methods, Study design]

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

direct quote[Methods, Study design]

“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.”

direct quote[Methods, Statistical analysis]

Table 1 provides baseline characteristics including sex (male), age (mean ± s.d.), race (Asian), weight, BMI, smoking history, current antifibrotic treatment, FVC, DLCO, and LCQ scores. Age, weight, and health status (lung function criteria) are reported. Sex is reported, and since the trial enrolled both sexes (though predominantly male), a single-sex justification is not required.

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· 95% conf

Demographic and clinical characteristics are well reported, including sex, age, weight, BMI, smoking history, and baseline lung function. All applicable sub-criteria are adequately addressed.

Evidence

paraphrase[Table 1]

“Table 1 reports male sex for each group: 100% in 30 mg QD, 83.3% in 30 mg BID, 88.9% in 60 mg QD, 88.2% in placebo.”

direct quote[Table 1]

“Age – years, mean ± s.d. 65.8 ± 7.0 ... Weight – kg, mean ± s.d. 72.27 ± 7.40 ... FVC mean – ml, mean ± s.d. 2,696 ± 444.2”

direct quote[Table 1]

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

PASSReviewer 2· 95% conf

Subject characteristics including sex, age, weight, and detailed demographics are well reported in Table 1, meeting the criteria for a human clinical trial.

Evidence

direct quote[Table 1]

“Table 1 Characteristics of the patients at baseline … Male sex – no. (%) 18 (100.0) … Age – years, mean ± s.d. 65.8 ± 7.0 … Weight – kg, mean ± s.d. 72.27 ± 7.40”

direct quote[Table 1]

“Race – no. (%) Asian 71 (100.0)”

PASSReviewer 3· 95% conf

Patient demographics (age, sex, race, weight, BMI, smoking history, lung function, concurrent medications) are comprehensively reported in Table 1, but sex is predominantly male (90%) with no justification for this imbalance (though it may reflect the epidemiology of IPF). Age, weight, and baseline lung function are well documented.

Evidence

direct quote[Table 1]

“Male sex – no. (%) 18 (100.0), 15 (83.3), 16 (88.9), 15 (88.2), 64 (90.1) ... Age – years, mean ± s.d. 65.8±7.0 ... Race – no. (%) Asian 18 (100.0) ... 71 (100.0)”

absence[Table 1 and Methods]

No scientific justification for the male-predominant sample is given

The methods state that the institutional review board or ethics committee at participating centers approved the protocols, all patients provided written informed consent, and the trial followed the Declaration of Helsinki and ICH Good Clinical Practice guidelines. These statements satisfy all applicable sub-criteria.

Sub-criteria

iacuc statementN/A

informed consentADEQUATE

irb ethics statementADEQUATE

regulatory complianceADEQUATE

Per reviewer (3) — the votes behind the verdict

PASSReviewer 1· 95% conf

Ethical compliance is clearly stated: ethics committee approval, written informed consent, and adherence to Declaration of Helsinki and ICH-GCP guidelines are reported.

Evidence

direct quote[Methods, Study design]

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

direct quote[Methods, Study design]

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

direct quote[Methods, Study design]

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

WARNReviewer 2· 90% conf

The ethics statement does not name the specific institutional review board or provide a protocol number, which is a reporting gap; informed consent and regulatory compliance are adequately stated.

Evidence

direct quote[Methods, Study design]

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

direct quote[Methods, Study design]

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

direct quote[Methods, Study design]

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

PASSReviewer 3· 95% conf

Ethics approval is explicitly stated with reference to the Declaration of Helsinki and ICH-GCP, and informed consent was obtained. A named ethics committee is referenced (institutional review board/ethics committee at participating centers), and trial registration is provided.

Evidence

direct quote[Methods, Study design]

“The trial was conducted following the principles outlined in the Declaration of Helsinki and the International Council for Harmonization guidelines for Good Clinical Practice. The institutional review board or ethics committee at participating centers approved protocols and adhered to local laws before initiation of the clinical trial. All patients in this study provided written informed consent.”

Rentosertib is described as provided by Insilico Medicine; placebo is described but not identified with a manufacturing source. Nintedanib and pirfenidone are mentioned as concurrent medications but not otherwise identified. The Olink Explore 3072 panel is named, but no catalog numbers or validation references are given for the proteomic assay. Statistical software packages (lme4, clusterProfiler, R) are named but without version numbers. Power analysis software is not cited. This is adequate for a clinical trial context but could be more complete.

Sub-criteria

mycoplasma testingN/A

reagents identifiedINADEQUATE

organisms identifiedN/A

antibodies identifiedN/A

cell line authenticationN/A

software tools identifiedINADEQUATE

Per reviewer (3) — the votes behind the verdict

WARNReviewer 1· 85% conf

The investigational product (rentosertib) is adequately identified, but statistical software versions are not reported, and the GitHub repository for custom code lacks versioning details.

Evidence

direct quote[Methods, Study design]

“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.”

direct quote[Methods, Statistical analysis]

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

direct quote[Code availability]

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

FAILReviewer 2· 85% conf

The investigational drug (rentosertib) is not fully identified with manufacturer/source details, formulation, or lot numbers, and the statistical software is missing version numbers, leading to a fail on this dimension.

Evidence

direct quote[Methods, Study design]

direct quote[Methods, Statistical analysis]

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

WARNReviewer 3· 90% conf

The investigational product (rentosertib) and comparators (nintedanib, pirfenidone) are named and manufacturers are implied, but no catalog numbers, lot numbers, or full supplier details are provided for the drug or assays. Key software tools (e.g., lme4, clusterProfiler, Olink panel) are named but without version numbers.

Evidence

direct quote[Methods, Study design]

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

direct quote[Methods, Statistical analysis]

“The lme4 package was used for the generalized linear mixed model analysis. ... Pathway analyses were performed using the clusterProfiler R package”

Tests (ANCOVA, MMRM, Spearman, GLMM) are named and appropriate. Assumptions are implicitly handled by standard models (adequate). Effect sizes with 95% CIs are reported for key endpoints (adequate). Data presentation includes per-group n and error bars (adequate). However, exact p-values are not consistently reported (e.g., only one exact p of 0.0495 is given). Software versions are missing (inadequate). Most critically, the text reports 'ALT increase (… 6 in 60 mg QD (33.3%) and 3 in 60 mg QD (16.7%))' while Table 2 shows only 1 ALT increase in the 60 mg QD group, a clear internal contradiction. This mathematical_plausibility issue constitutes a demonstrable error, triggering an overall fail per the scoring rule.

Sub-criteria

tests namedADEQUATE

exact p valuesINADEQUATE

effect sizes ciADEQUATE

data presentationADEQUATE

software identifiedINADEQUATE

assumptions verifiedADEQUATE

mathematical plausibilityINADEQUATE

Per reviewer (3) — the votes behind the verdict

PASSReviewer 1· 80% conf

Statistical methods are described appropriately for a phase 2a trial, with named tests, effect sizes with confidence intervals, and appropriate handling of missing data. Exact p-values are not consistently reported, and software versions are omitted, but these are minor issues for a safety-focused early-phase study.

Evidence

direct quote[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. ... Spearman correlation analysis was used to assess the association between FVC change and AUC 0– t or C trough .”

direct quote[Results, Secondary lung function endpoints]

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

direct quote[Results, Secondary lung function endpoints]

“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)”

FAILReviewer 2· 90% conf

While most statistical methods are appropriately named and effect sizes with confidence intervals are reported, an internal inconsistency in the reported ALT increase counts between the text and Table 2 constitutes a demonstrable error, leading to an overall fail.

Evidence

direct quote[Results, primary safety endpoint paragraph vs Table 2]

“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%))”

direct quote[Table 2]

“ALT increase – no. (%) 0 (0) 0 (0) 1 (5.6) 0 (0)”

direct quote[Results, secondary lung function endpoints]

“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)”

WARNReviewer 3· 90% conf

Statistical tests are named, exact p-values are reported for some endpoints (e.g., LCQ p=0.0495), and effect sizes with 95% CIs are given for FVC change. However, test assumptions are not verified, no multiplicity correction is applied to secondary endpoints, and the primary safety endpoint lacks a formal hypothesis test.

Evidence

direct quote[Methods, Statistical analysis]

“An analysis of covariance (ANCOVA) model was performed for endpoints including FVC.”

direct quote[Results, Secondary lung function endpoints]

“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).”

direct quote[Methods, Statistical analysis]

“All statistical comparisons were made using two-sided tests.”

The data availability statement is concrete, naming the OMIX database (accession OMIX008341) for proteomics and FVC data. The study protocol is available in a secure data-sharing environment. Custom code for Olink analysis is shared via a public GitHub repository. These satisfy all four sub-criteria. Although patient-level data access is managed, this is appropriate for privacy reasons.

Sub-criteria

code sharingADEQUATE

accession numbersADEQUATE

repository depositADEQUATE

data availability statementADEQUATE

Per reviewer (3) — the votes behind the verdict

PASSReviewer 1· 95% conf

Data availability is well documented: de-identified Olink proteomics and FVC data are deposited in the OMIX database with an accession number, the study protocol is available on request, and custom code is shared on GitHub.

Evidence

direct quote[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 ).”

direct quote[Data availability]

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

direct quote[Code availability]

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

PASSReviewer 2· 95% conf

Proteomics data are deposited in a public repository (OMIX) with a stable accession, the study protocol is available upon managed access, and custom code is shared on GitHub.

Evidence

direct quote[Data availability]

“Complete deidentified Olink proteomics data, and FVC data, have been deposited at the OMIX database under accession codes at OMIX008341”

direct quote[Code availability]

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

direct quote[Data availability]

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

PASSReviewer 3· 95% conf

Proteomics and FVC data are deposited in the OMIX database with an accession code, and custom analysis code is shared on GitHub. Study protocol and statistical analysis plan are available on request, which is appropriate for a clinical trial with patient-level data.

Evidence

direct quote[Data availability section]

“Complete deidentified Olink proteomics data, and FVC data, have been deposited at the OMIX database under accession codes at OMIX008341”

direct quote[Code availability section]

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

direct quote[Data availability section]

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

The methods section includes full protocol details. Trial registration number is provided. A Nature Portfolio reporting summary is accessible. All outcomes (safety, FVC, DLCO, FEV1, LCQ, 6MWD, PK, proteomics) are reported, including null results. Limitations are explicitly acknowledged. The conclusion appropriately calls for larger longer trials. Competing interests are stated. No overclaiming detected.

Sub-criteria

funding coiADEQUATE

trial registrationADEQUATE

reporting guidelineADEQUATE

methods completenessADEQUATE

all outcomes reportedADEQUATE

limitations discussedADEQUATE

conclusions proportionalADEQUATE

Per reviewer (3) — the votes behind the verdict

PASSReviewer 1· 90% conf

Reporting is thorough: methods are detailed, the trial is registered (NCT05938920), a reporting summary is linked, all pre-specified outcomes are reported, limitations are discussed, conclusions are proportional, and funding/competing interests are disclosed.

Evidence

direct quote[Abstract]

“ClinicalTrials.gov registration number: NCT05938920”

direct quote[Methods, Reporting summary]

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

direct quote[Discussion, paragraph 7]

PASSReviewer 2· 90% conf

The trial is registered, methods are detailed, all outcomes are reported, limitations are discussed, conclusions are proportional, and funding/competing interests are stated.

Evidence

direct quote[Abstract and Methods]

“ClinicalTrials.gov registration number: NCT05938920”

direct quote[Results sections]

“Treatment-related serious adverse event rates were low and comparable across treatment groups”

direct quote[Discussion, paragraph 7]

“The limitations of this study include the small cohort size of each arm, the geographical and demographic homogeneity of the participants… and a short period of follow-up”

WARNReviewer 3· 90% conf

Methods are detailed and reproducible, the trial is registered, funding/conflict statements are present, and limitations are discussed. However, no reporting guideline checklist (CONSORT) is explicitly referenced, and the conclusions slightly overstate the evidence (suggesting efficacy from a small phase 2a safety study).

Evidence

direct quote[Abstract, Methods]

“ClinicalTrials.gov registration number: NCT05938920”

direct quote[Discussion, paragraph 5]

“The limitations of this study include the small cohort size of each arm, the geographical and demographic homogeneity of the participants... and a short period of follow-up, which limit the assessment of long-term safety and efficacy.”

direct quote[Abstract]

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

direct quote[Methods, Reporting summary section]

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

13 copyedit issues flagged (1 major): mostly consistency, clarity, typo.

Checked 67 references: 65 verified — 2 unresolved.

Reproducibility

3 data/code links checked; 3 live.

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

Αlpha¹ Rigor Action ItemsFinal pass before journal submission

For post-publication audit: the paper is robust in design and reporting but contains a significant numerical error and overstatement that warrant an erratum from the authors. An informed reader should weigh the ALT discrepancy and the limited efficacy evidence before citing the paper as evidence of efficacy.

1.
HIGHstatistics
Resolve the ALT increase discrepancy between the text (Results, primary safety endpoint) and Table 2. Correct the numbers and ensure consistency throughout the paper.
This is a demonstrable error that undermines data integrity and requires an erratum.
2.
HIGHother
For the two references not found in the registry ("Machine learning-aided generative molecular design" and "Chondroadherin binds to type II collagen"), verify the citations and replace or correct them if they are incorrect or fabricated.
Unresolved references are a potential fabrication signal and must be rectified.
3.
HIGHreporting
Add the specific IRB name(s) and protocol number(s) to the Methods section (Ethics statement).
Clinical trials should name the approving ethics committee for transparency; the current generic statement is inadequate.
4.
HIGHother
Add formulation details (e.g., tablet strength, lot number) for rentosertib in the Methods section.
Reproducibility requires full specification of the investigational product.
5.
HIGHother
Add version numbers for all software packages (R, lme4, clusterProfiler, etc.) in the Methods (Statistical analysis) section.
Software versions are essential for reproducibility of the analyses.
6.
HIGHreporting
Tone down the abstract's claim of efficacy: replace 'safe and effective' with 'safe and potentially effective' or 'warrants further investigation'.
The study was not powered for efficacy; the abstract overstates the evidence from a single dose signal in a small cohort.
7.
MEDIUMstatistics
Report exact p-values for all primary and key secondary endpoints, not just threshold-based statements.
Exact p-values allow readers to fully evaluate the evidence; currently only one exact p-value is reported.
8.
MEDIUMstatistics
Verify the hypokalemia percentage in the 60 mg QD group (20.4%): 3/18 = 16.7%, not 20.4%. Correct if necessary.
Numerical inconsistency may indicate a typo that should be corrected.
9.
MEDIUMreporting
Add an explicit reference to the CONSORT 2010 checklist in the Methods and provide the completed checklist as a supplementary file.
CONSORT is the standard reporting guideline for RCTs; its absence is a gap in transparency.
10.
MEDIUMstatistics
Add a statement verifying key assumptions for the ANCOVA and MMRM models (normality, homogeneity of variance) or justify robustness.
Assumption verification is expected for parametric models, especially with small sample sizes.
11.
LOWcopyedit
Fix grammar: 'increases time' -> 'increased time' in the Pharmacokinetic section.
Minor grammatical error that affects readability.
12.
LOWcopyedit
Add commas for clarity in long lists, e.g., in Table 2 footnote and discussion of TEAEs.
Minor punctuation improvements enhance clarity.

Full text (54,229 chars)Analyzed: Jul 4, 2026

Αlpha¹ assesses methodological rigor, not the importance of the findings. See how we calibrate →