AI Peer Review for Oncology Manuscripts
From PFS-versus-OS arguments to immortal time bias, oncology reviewers know exactly where to look. So does ManuscriptMind.
Oncology has its own catalogue of reviewer red flags. Progression-free survival presented as a stand-in for overall survival, biomarker cutoffs chosen to maximize a p-value, single-arm trials leaning on historical controls, and immortal time bias quietly inflating a survival curve.
ManuscriptMind evaluates your oncology manuscript against the standards reviewers actually apply: endpoint validity, correct survival methodology, honest handling of subgroups, and conclusions that match the strength of the design.
What reviewers flag in Oncology papers
Surrogate endpoints in place of overall survival
Reporting progression-free survival or objective response rate and implying a survival benefit. Reviewers ask whether the surrogate has been validated for this tumor type and setting.
Immortal time bias
Classifying patients by an event that can only occur if they survive long enough, which mechanically favors the treated group. A frequent and fatal flaw in observational oncology analyses.
Optimal-cutpoint biomarker dredging
Searching for the biomarker threshold that produces the most significant split, then reporting it without correction or validation. Reviewers recognize the inflated effect immediately.
Single-arm trials with historical controls
Comparing outcomes to historical or external cohorts without addressing differences in staging, supportive care, or era effects.
Underpowered subgroup claims
Highlighting a positive result in a small biomarker-defined subgroup that the study was never powered to detect.
Statistical pitfalls specific to Oncology
- Immortal time bias from time-dependent exposures treated as fixed at baseline
- Optimal cutpoint selection without multiplicity correction or external validation
- Progression-free survival interpreted as evidence of overall survival benefit
- Multiplicity across exploratory biomarker subgroups
Reporting guidelines we check against
What ManuscriptMind checks in your Oncology manuscript
- Whether your primary endpoint justifies the survival or benefit claim you make
- Time-to-event setup for immortal time bias and informative censoring
- How biomarker cutoffs were chosen and whether they were validated
- Comparability of any external or historical control group
- Whether subgroup conclusions exceed the study's power
Review your Oncology manuscript before you submit
Upload your paper and get structured, severity-classified feedback in minutes. Methodology, statistics, and literature issues flagged with specific fixes. Free during beta.
Frequently asked questions
Does it catch immortal time bias?
Yes. Immortal time bias is one of the most common and damaging errors in observational oncology research, and ManuscriptMind specifically looks for time-dependent exposures that have been treated as fixed at baseline.
Can it tell me if PFS is an acceptable endpoint for my study?
It flags when progression-free survival or response rate is being used to support claims that really require overall survival, and prompts you to justify the surrogate for your tumor type and setting.
Does it understand biomarker and cutpoint analyses?
Yes. It looks for optimal-cutpoint dredging, missing validation, and uncorrected multiplicity across biomarker-defined subgroups.
Will my unpublished oncology data stay private?
ManuscriptMind never trains on your manuscripts and deletes data on request. Confidentiality of unpublished clinical data is treated as essential.