Stage-Dependent Role of Eicosanoids in Colorectal Cancer

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Article titled “Stage-Dependent Role of Eicosanoids in Colorectal Cancer“ reports on the study conducted to show the potential of serum eicosanoid level analysis as biomarkers for colorectal cancer staging, as well as tentative drug targets. The manuscript is well written, providing adequate literature background, with elaborate statistical analysis.

The main limitation is the absence of a control group, which would give the interpretation of the results more solid ground. However, this limitation is clearly stated in the text. Minor English editing required

I recommend the publication of the article

Author Response

We would like to thank the reviewer for such a positive reception of our article.

We have corrected the article for language mistakes.

Reviewer 2 Report

Comments and Suggestions for Authors

The manuscript “Stage-Dependent Role of Eicosanoids in Colorectal Cancer” examines how disease progression is associated with measurable changes in serum eicosanoid profiles. Among the analyzed compounds, TXB₂, PGD₂, and LTB₄ demonstrate potential as stage-related biomarkers, with TXB₂ showing the strongest association with advanced disease, while PGE₂ appears to have limited diagnostic value.

My suggestions are as follows:

1. Please explain the relevance of prostaglandins, considering that multiple prostaglandins are monitored (PGD₂, PGE₂, PGF₂α, etc.).

2. In Table 1, TNM and TNM < III data are presented as  (median [Q1–Q3]). I suggest presenting the results as mean ± SD instead.

General comment: References should be formatted using square brackets.

Author Response

We thank you for the reviewing of our article. Here we would like to address your comments.

1. Please explain the relevance of prostaglandins, considering that multiple prostaglandins are monitored (PGD₂, PGE₂, PGF₂α, etc.).

An additional paragraph has been added to the introduction section of the article providing more background for the prostaglandins.

2. In Table 1, TNM and TNM < III data are presented as  (median [Q1–Q3]). I suggest presenting the results as mean ± SD instead.

We have added means and SD to Table 1 as the Reviewer suggested. Along with medians and quartiles, this more fully reports the characteristics of the quantitative features.

General comment: References should be formatted using square brackets.

The reference numbers are now in square brackets.

Reviewer 3 Report

Comments and Suggestions for Authors

Dear Authors,

the manuscript entitled “Stage-Dependent Role of Eicosanoids in Colorectal Cancer” investigates the association between circulating eicosanoid profiles and colorectal cancer progression using a targeted metabolomics approach combined with regression-based modeling strategies. The topic is timely and relevant, and the use of two complementary feature-selection approaches adds robustness to the statistical inference.

The study is methodologically sound and addresses an important biological question. However, some aspects related to interpretation, clarity of presentation, and framing of the conclusions require revision to improve readability and to better align the claims with the presented data. In particular:

• The absence of a healthy control group limits the interpretation of eicosanoids as biomarkers. While the data convincingly show stage-dependent associations within CRC patients, claims regarding biomarker potential, especially in Abstract and Conclusion, should be toned down and more clearly framed as associative rather than diagnostic or prognostic.
• The frequent use of inverse odds ratios (1/OR) throughout the Results and Discussion may reduce clarity. Reporting odds ratios consistently in a single direction would improve interpretability for the reader.
• Lines 212-224: The inverse association between LTB₄ levels and local tumor advancement contrasts with several studies suggesting a tumor-promoting role of this mediator. This discrepancy deserves a more detailed discussion, considering possible explanations such as systemic versus local effects, immune modulation, or stage-specific mechanisms.
• Lines 245-254: Given the sample size and events-per-variable ratio, a brief discussion of feature-selection stability is recommended, with results framed more explicitly as hypothesis-generating.

Based on the comments above, I recommend major revision of the manuscript.

Author Response

We would like to thank the Reviewer for thorough review of our article. Below we address the comments.

• The absence of a healthy control group limits the interpretation of eicosanoids as biomarkers. While the data convincingly show stage-dependent associations within CRC patients, claims regarding biomarker potential, especially in Abstract and Conclusion, should be toned down and more clearly framed as associative rather than diagnostic or prognostic.
  • Thank you. We have revised the abstract and conclusions to moderate the strength of the claims and ensure they accurately reflect the presented data.
• The frequent use of inverse odds ratios (1/OR) throughout the Results and Discussion may reduce clarity. Reporting odds ratios consistently in a single direction would improve interpretability for the reader.
  • We agree. To improve readability, we revised the manuscript to report odds ratios in a single, consistent direction throughout. Specifically, all ORs are now oriented such that OR > 1 indicates higher odds of more advanced disease (higher ordinal category for TNM/T/N, or the event TNM ≥ III for binary models). When the native model parameterization implied the opposite direction, estimates and corresponding confidence intervals were inverted (1/OR) solely for interpretability. We removed inverse-OR notation from the Results and Discussion and clarified the reporting convention in the Methods and table footnotes.
• Lines 212-224: The inverse association between LTB₄ levels and local tumor advancement contrasts with several studies suggesting a tumor-promoting role of this mediator. This discrepancy deserves a more detailed discussion, considering possible explanations such as systemic versus local effects, immune modulation, or stage-specific mechanisms.
  • We have expanded the discussion to provide more explanation to this issue.
• Lines 245-254: Given the sample size and events-per-variable ratio, a brief discussion of feature-selection stability is recommended, with results framed more explicitly as hypothesis-generating.
  • We agree and revised limitations section by adding appropriate information. Given the limited sample size and the risk of instability in high-dimensional selection, we treated variable selection explicitly as sensitivity analysis rather than definitive discovery. We used two complementary selection paradigms—Elastic Net penalization and likelihood-based stepwise LRT—to evaluate whether selected metabolites are robust to the selection mechanism. We now state more clearly that the multivariable findings are hypothesis-generating, and we highlight signals that were consistent across model tiers (M0/M1) and outcomes as the most credible candidates for follow-up. We did not apply bootstrap resampling because, under our events-per-variable constraints, repeated refitting led to frequent convergence issues and highly variable selected sets; thus, bootstrap would be unlikely to yield stable, interpretable stability metrics in this setting. Instead, we provide transparent reporting of coefficient tables in the Supplementary Materials.

Round 2

Reviewer 3 Report

Comments and Suggestions for Authors

Dear Authors,
thank you for submitting the revised version of the manuscript. I have checked the changes and I confirm that my previous comments have been adequately addressed. The manuscript has improved in clarity and interpretation and is now suitable for publication.

Comments on the Quality of English Language