Journal of Clinical Oncology Response to Reviewers: JCO Revision Guide

Make reviewer text and author response visually distinct. Bold the complete comment and keep the reply in ordinary text, unless the decision letter specifies a different format. Use labels in addition to color.

Dear Editor,

Thank you for the opportunity to revise our manuscript. We focused on the
issues emphasized in the decision letter: interpretation of the primary
endpoint, multiplicity in the subgroup analyses, and the balance between
clinical benefit and treatment-related toxicity.

Reviewer 1, Comment 1: [Paste the complete comment.]

Response: We agree that the original abstract emphasized relative effect
without absolute clinical context. We now report the absolute difference,
hazard ratio, and 95% confidence interval in the abstract and Results. The
changes appear on page 2, lines 34-41 and page 9, lines 244-261. The updated
estimate is shown in Figure 2 and Table 2.

If the response changes a result, quote the revised sentence so the reviewer does not have to infer the new interpretation.

Each response should show whether the revision changes statistical confidence, clinical importance, or both. Keep that distinction visible as you work through endpoints, subgroups, safety, and reporting.

Endpoints and estimands

State exactly what was measured, over what time horizon, and under what handling of competing events, crossover, censoring, or missing data. If the reviewer identifies a mismatch between protocol, registry, statistical plan, and manuscript, reconcile it openly.

When a sensitivity analysis changes the estimate, report the direction and magnitude. Avoid saying the result is "unchanged" when only the threshold status is unchanged.

Effect size and clinical meaning

JCO readers need more than a p-value. Report relative and absolute effects where the design permits, with confidence intervals and relevant baseline risk. Explain whether the observed difference is large enough to matter for patients, clinicians, or trial design.

Do not retrofit a minimum clinically important difference after seeing the data. If the threshold was not prespecified, describe the interpretation as contextual.

Subgroups and biomarkers

For a requested subgroup analysis, state whether it was prespecified, the sample size and event count, the interaction test, and how multiplicity was handled. A significant result in one subgroup and a non-significant result in another does not establish heterogeneity.

If the analysis is exploratory, label it in the response, figure caption, Results, and Discussion. Avoid using it to recommend treatment selection without validation.

Safety and tolerability

Respond with exposure-adjusted context when appropriate, denominators, severity grades, timing, dose changes, discontinuations, and serious events. Explain whether missing follow-up or differential exposure changes the comparison.

If a treatment improves an efficacy endpoint but increases clinically meaningful harm, revise the conclusion to show the tradeoff rather than presenting safety as a secondary footnote.

Reporting and reproducibility

Update the appropriate reporting checklist, flow diagram, protocol or analysis-plan links, trial registration, data-sharing statement, and supplementary methods. The response should identify where each item changed.

For retrospective and observational studies, make cohort construction and exclusions traceable. For trials, keep protocol-defined and manuscript-defined endpoints aligned.

Do not use politeness to obscure a changed result. A direct, bounded answer is easier for the editor to trust.

Some requests would create more noise than evidence: a severely underpowered subgroup, an outcome not reliably captured, an unplanned cut point chosen after inspecting the data, or a model that violates the study design.

Use this sequence:

1. State the clinical question behind the request.
2. Explain the design or statistical limitation.
3. Provide a valid descriptive or sensitivity analysis when useful.
4. Label it exploratory and avoid decision language.
5. Add the unresolved question to the limitations.

Do not decline with "outside scope" alone. Show that you considered the clinical question and chose the most valid answer available.

Reviewer 2, Comment 3: The treatment effect may be greater in patients with biomarker-positive disease. Please present separate efficacy analyses.

Response: We agree that biomarker status is clinically relevant. This subgroup was not prespecified, and the biomarker-positive group includes [number] patients and [number] events. We added an exploratory treatment-by-biomarker interaction model rather than comparing separate within-group p-values. The interaction estimate and confidence interval appear in Figure S3 and Table S5. Because the study is not powered to establish predictive utility, we removed patient-selection language from the abstract and describe the result as hypothesis-generating on page 14, lines 402-419. Methods are on page 22, lines 681-714.

The response answers the question without turning a small exploratory cell into a validated biomarker.

Reconcile the response against:

• protocol, registry, and statistical analysis plan
• participant flow and all analysis denominators
• endpoint definitions and time windows
• effect estimates, confidence intervals, and multiplicity statements
• subgroup labels, cut points, and interaction tests
• adverse-event denominators, grades, and exposure
• tables, figures, appendices, and reporting checklists
• data-sharing and code statements
• abstract and conclusion language

Generate page and line references from the final clean manuscript. Check that the marked copy does not contain unresolved track changes or author notes.

Across our JCO pre-submission and revision reviews, three patterns repeatedly weaken otherwise serious clinical responses. These are Manusights editorial observations, not official ASCO acceptance rules.

The JCO subgroup answer skips the interaction

Authors often respond to a reviewer by showing a small p-value in one group and a large p-value in another. That comparison does not test whether treatment effects differ. We audit the JCO forest plot, subgroup table, Methods, and response for the interaction estimate, prespecification status, event counts, and multiplicity. When those pieces are absent, the right revision is usually exploratory framing and removal of patient-selection language, not a stronger claim based on separate significance tests.

The JCO abstract preserves relative benefit without absolute context

A revised model may produce a persuasive hazard ratio while the absolute difference is small, uncertain, or dependent on follow-up. In our review work, we compare the abstract with the survival figure, numbers at risk, median follow-up, absolute event rates, and confidence intervals. The JCO response should tell the editor whether the clinical interpretation changes when absolute context is visible. If it does, the title and conclusion need the same calibration.

The JCO safety reply changes denominators

Safety tables may use treated patients, evaluable patients, cycles, or person-time, while efficacy uses randomized or eligible participants. A response can appear reassuring because the denominator changes silently. We trace every JCO safety percentage to its analysis population and exposure window, then compare discontinuation, dose modification, serious events, and follow-up. The response should explain denominator differences and avoid "manageable" or "acceptable" unless the actual event pattern and study design support that judgment.

We also compare the JCO response with the protocol and registry when a revised endpoint, cutoff, or analysis population appears. A post hoc definition can be clinically interesting, but it cannot be presented as though it governed the original design. The response should identify when the analysis was introduced, why it was added, and how it changes interpretation. That disclosure belongs in the Methods, figure caption, and limitations, not only in correspondence that readers will never see.

Finally, we test whether the JCO abstract gives readers the same benefit-harm picture as the response. If a revision adds a serious toxicity signal, short follow-up caveat, or uncertain subgroup interaction, the abstract cannot preserve an unconditional efficacy conclusion. The strongest revisions use one calibrated clinical sentence across the abstract, response, and Discussion.

These patterns create a gap between the careful qualifications in the response letter and the stronger clinical story that remains in the manuscript.

The response protects the headline at all costs

If corrected analysis weakens the estimate, revise the headline. Transparency is more persuasive than threshold-preserving language. State the new estimate and explain which clinical interpretation no longer follows from it.

Exploratory results become clinical recommendations

Label exploratory analyses consistently and keep treatment-selection claims out until validation exists. The same label should appear in the response, Results, figure caption, abstract, and Discussion.

Reporting fixes stay in the rebuttal

Put endpoint, cohort, and model clarification in the manuscript, checklist, and supplement too. Readers will not see the private correspondence, so the published record must contain every definition needed to interpret the result.

The reviewer cannot find the changed evidence

Use exact page, line, figure, and table references and verify them after final pagination. When a result changes, quote the key revised sentence so the reviewer can confirm the new interpretation immediately.

Most major revisions remain conditional, and rejection is still possible when the primary endpoint interpretation is unstable, subgroup claims exceed the interaction evidence, or safety and benefit are presented with incompatible denominators. A complete response file does not repair a clinically overclaimed conclusion.

Think twice before resubmitting if the paper's central treatment-selection claim depends on an unplanned, underpowered subgroup and the revision has no external validation. The claim needs to become exploratory or the evidence needs to improve.

• Every editor and reviewer comment is reproduced and answered.
• Each response states the action, result, interpretation, and exact location.
• Endpoint and analysis populations match protocol, registry, tables, and figures.
• Effect sizes include uncertainty and relevant absolute context.
• Subgroup claims match prespecification, interaction testing, and power.
• Safety denominators and exposure windows are explicit.
• Reviewer comments and author replies are visually distinct.
• The clean manuscript, marked copy, response, appendices, and reporting checklist agree.

Use the free revision readiness scan to test whether the revised clinical claim survives the editor's decision issues.

We reviewed current ASCO journal author and data-sharing materials, established rebuttal guidance, public JCO response examples, and recurring clinical-interpretation problems in Manusights revision work. Use this page when a JCO decision requires one consistent account of endpoints, effect size, subgroup evidence, safety, and clinical meaning.

Reviewed July 12, 2026.

• Journal of Clinical Oncology author center
• ASCO journals data-sharing policy
• PLOS Computational Biology: Ten Simple Rules for Writing a Response to Reviewers
• Nature Computational Science guidance on responding to peer review

Official ASCO sources support journal reporting and author requirements. The response matrices and failure patterns are Manusights editorial guidance and do not predict a JCO decision.