Serum Antioxidant Capacity Predicts Prognosis in Patients with Metastatic Colorectal Cancer: An Original Cohort Study

Round 1

Reviewer 1 Report

This study investigated the relationship between baseline serum antioxidant capacity (biological antioxidant potential, BAP) and chemotherapy response as well as prognosis in patients with metastatic colorectal cancer (mCRC). The research offers a unique perspective and sound clinical entry point, suggesting that systemic antioxidant buffering capacity may be one of the key determinants of chemotherapy sensitivity.

1.Measurements of BAP and d-ROMs in this study were performed prior to primary tumor resection, while systemic chemotherapy for all patients in the cohort was administered after surgery. Surgical stress itself, along with the removal of the primary tumor, can drastically alter the systemic redox state of patients, a point the authors themselves acknowledged in the Discussion section regarding potential resection-related changes. Using preoperative biomarkers to predict chemotherapy response over a long postoperative period introduces significant temporal confounding bias, and the authors need to fully explain the core reasons why blood samples were not collected immediately prior to the initiation of chemotherapy.

2.This study is a single-center, retrospective analysis with a small sample size, carrying the inherent selection bias and limited generalizability of this study design. The authors should conduct a more systematic and detailed discussion of the study’s limitations in the manuscript, rather than only addressing them briefly in passing.

3.On lines 265-266 of the main text, the authors made a clear error in the definition of the abbreviation DSS, writing it as "drug exposure duration satisfaction (DSS)", while in both the Abstract and Methods sections, DSS is explicitly defined as disease-specific survival. The authors should correct this typo immediately to ensure consistent definition of the abbreviation throughout the manuscript.

4.ROC analysis showed that the area under the curve (AUC) of BAP for predicting 3-year DSS was only 0.509, which is nearly equivalent to the 0.5 threshold of random discrimination. While this marker reached statistical significance as a continuous variable in Cox regression analysis, it has extremely poor clinical utility as a binary marker for risk stratification. The authors are advised to develop a combined predictive model integrating BAP, CEA, and core tumor molecular subtypes to improve the clinical application value of this biomarker.

5.There is significant heterogeneity in the chemotherapy regimens across the study cohort, including oral 5-FU monotherapy in 10 patients, doublet combination regimens in 74 patients, and concurrent molecular targeted agents in a subset of patients, such as the anti-angiogenic agent bevacizumab and the anti-EGFR agent panitumumab. Different targeted agents have fundamentally distinct regulatory effects on oxidative stress in the tumor microenvironment, yet the multivariate analysis in this study only adjusted for the number of chemotherapy regimens, without stratified analysis or adjustment for regimen type and targeted agent use. The authors should supplement these relevant analyses to rule out the interference of regimen heterogeneity on the study conclusions.

6.The Discussion section is currently presented as large blocks of continuous text with an unclear logical hierarchy, making it difficult for readers to quickly capture the core viewpoints. The authors are recommended to add subheadings to the Discussion section, for example, dividing it into subsections such as Mechanisms of Tumor Oxidative Stress and Chemotherapy Resistance, Clinical Significance of BAP as a Predictive Biomarker, and Study Limitations, which will greatly improve the logical structure and readability of the manuscript.

7.The formatting of some terms is inconsistent throughout the full text, most notably for derivatives of reactive oxygen metabolites, which are sometimes written as the plural "d-ROMs" and other times as the singular "d-ROM". The authors should read through the full text to unify the formatting and expression of terms, ensuring the rigor of the manuscript.

Author Response

EDITOR’S SPECIFIC COMMENTS: 

1. Measurements of BAP and d-ROMs in this study were performed prior to primary tumor resection, while systemic chemotherapy for all patients in the cohort was administered after surgery. Surgical stress itself, along with the removal of the primary tumor, can drastically alter the systemic redox state of patients, a point the authors themselves acknowledged in the Discussion section regarding potential resection-related    Using  preoperative  biomarkers  to  predict  chemotherapy response over a long postoperative period introduces significant temporal confounding bias, and the authors need to fully explain the core reasons why blood samples were not collected immediately prior to the initiation of chemotherapy.

 

[Response]

We appreciate the reviewer’s insightful comment and the opportunity to clarify this important point.

We agree that blood sampling immediately prior to the initiation of systemic chemotherapy would be methodologically preferable to minimize potential temporal confounding. However, in the present retrospective cohort, postoperative serum samples obtained immediately before chemotherapy were not systematically collected or preserved in most patients. Therefore, analyses based on such time points were not feasible without substantially reducing the sample size and introducing selection bias. For this reason, we used uniformly available preoperative samples as baseline measurements.

We also acknowledge that surgical stress and removal of the primary tumor may influence systemic redox status. As described in the Discussion section of the submitted manuscript (lines 515–519), we have previously reported that although d-ROM levels decrease after tumor resection in association with tumor burden, BAP values do not show significant changes at one month postoperatively (Reference 32). Based on these findings, we considered that preoperative BAP may reasonably reflect systemic antioxidant capacity after postoperative recovery, although we recognize that residual temporal confounding cannot be completely excluded.

We have revised the limitations section of the Discussion to address this point, as follows:
“Finally, the systemic redox state may change during surgery and chemotherapy. Prospective studies with longitudinal measurements of d-ROMs and BAP throughout the treatment period are needed to clarify whether changes in oxidative stress and antioxidant capacity are associated with treatment efficacy, acquired resistance, and survival outcomes.”

 

 

2. This study is a single-center, retrospective analysis with a small sample size, carrying the inherent selection bias and limited generalizability of this study design. The authors should conduct a more systematic and detailed discussion of the study’s limitations in the manuscript, rather than only addressing them briefly in passing.

 

[Response]

We appreciate these helpful suggestions.

We have revised the limitations section of the Discussion to address this point, as follows:
“Finally, the systemic redox state may change during surgery and chemotherapy. Prospective studies with longitudinal measurements of d-ROMs and BAP throughout the treatment period are needed to clarify whether changes in oxidative stress and antioxidant capacity are associated with treatment efficacy, acquired resistance, and survival outcomes.”

 

3.On lines 265-266 of the main text, the authors made a clear error in the definition of the abbreviation DSS, writing it as "drug exposure duration satisfaction (DSS)", while in both the Abstract and Methods sections, DSS is explicitly defined as disease-specific survival. The authors should correct this typo immediately to ensure consistent definition of the abbreviation throughout the manuscript.

 

[Response]

We thank the reviewer for the careful review.

We have corrected the definition of DSS in lines 265–266 to “disease-specific survival”

 

4.ROC analysis showed that the area under the curve (AUC) of BAP for predicting 3-year DSS was only 0.509, which is nearly equivalent to the 0.5 threshold of random discrimination. While this marker reached statistical significance as a continuous variable in Cox regression analysis, it has extremely poor clinical utility as a binary marker for risk stratification. The authors are advised to develop a combined predictive model integrating BAP, CEA, and core tumor molecular subtypes to improve the clinical application value of this biomarker.

 

[Response]

We thank the reviewer for the careful review of the manuscript.

We agree that the ROC analysis indicates very limited discriminative ability of BAP for predicting 3-year disease-specific survival (DSS), with an AUC of 0.509. Accordingly, we have explicitly stated in the Limitations that the ROC-derived cutoff for BAP may be unreliable and that BAP has extremely limited clinical utility when used as a standalone binary marker for risk stratification.

We also corrected an inaccurate description in the previous version of the manuscript. In the Discussion, we had stated that BAP was analyzed as a continuous variable in multivariable Cox regression; however, in the DSS analyses BAP was handled as a binary variable using a cutoff value. To avoid any misunderstanding, we informed the Assistant Editor of this error after submission, and the sentence has been deleted in the revised manuscript.

To provide clinically interpretable evidence beyond binary risk grouping, we added two new Results subsections: “DSS According to Therapeutic Response” (Section 3.5) and “Multivariable Analysis of Prognosis According to Therapeutic Response” (Section 3.6). These analyses show that therapeutic response at 4 months is strongly associated with 3-year DSS (69.0% for PR/SD vs. 13.6% for PD; log-rank P < 0.001) and remains an independent predictor in multivariable Cox analysis (HR 7.641; 95% CI 3.693–15.810; P < 0.001). Importantly, we confirmed that higher baseline BAP is independently associated with poorer therapeutic response in multivariable analysis when BAP is treated as a continuous variable. Taken together, these findings support the interpretation that BAP may influence prognosis indirectly through treatment resistance, while the ROC-derived dichotomization itself performs poorly; we added this point to the Limitations.

Regarding the suggestion to develop a combined predictive model integrating BAP, carcinoembryonic antigen (CEA), and core tumor molecular subtypes, we additionally evaluated CEA using the conventional cutoff of 5 ng/mL. In this stage IV-only cohort, DSS did not differ between the CEA < 5 and CEA ≥ 5 groups (51.7% vs. 47.7%; log-rank P = 0.550), suggesting limited usefulness of CEA for risk stratification in the present dataset. In our prior studies including a wider range of stages, we reported the prognostic utility of CEA and its combination with cancer stem cell markers in circulating tumor cells (Sawai et al., Cancers 2024; Sawai et al., Int. J. Mol. Sci. 2025). However, restricting the analysis to stage IV cases may attenuate the prognostic gradient of CEA. In addition, comprehensive molecular subtype data were not available for all patients, and the current sample size is limited for developing and validating a multivariable prediction model without substantial risk of overfitting. We therefore did not construct a combined model in this revision, but we agree that developing and validating an integrated model in a larger cohort with molecular profiling is an important future direction.

All additions and changes are highlighted in yellow in the revised manuscript.

 

5.There is significant heterogeneity in the chemotherapy regimens across the study cohort, including oral 5-FU monotherapy in 10 patients, doublet combination regimens in 74 patients, and concurrent molecular targeted agents in a subset of patients, such as the anti-angiogenic agent bevacizumab and the anti-EGFR agent panitumumab. Different targeted agents have fundamentally distinct regulatory effects on oxidative stress in the tumor microenvironment, yet the multivariate analysis in this study only adjusted for the number of chemotherapy regimens, without stratified analysis or adjustment for regimen type and targeted agent use. The authors should supplement these relevant analyses to rule out the interference of regimen heterogeneity on the study conclusions

 

[Response]

We appreciate the reviewer’s suggestions.

We have revised the classification of chemotherapy regimens into three categories: oral 5-FU agents, doublet regimens, and doublet regimens combined with molecular targeted therapy. Accordingly, the items in Table 5 have been updated. In addition, in Section 3.4 (Multivariate Analysis of Therapeutic Response), we reanalyzed the data using these three regimen categories, which yielded the following result: β = 0.282 (95% CI, 0.0001–0.0008; P = 0.009). All changes have been highlighted in yellow.

 

6.The Discussion section is currently presented as large blocks of continuous text with an unclear logical hierarchy, making it difficult for readers to quickly capture the core viewpoints. The authors are recommended to add subheadings to the Discussion section, for example, dividing it into subsections such as Mechanisms of Tumor Oxidative Stress and Chemotherapy Resistance, Clinical Significance of BAP as a Predictive Biomarker, and Study Limitations, which will greatly improve the logical structure and readability of the manuscript.

 

[Response]

We appreciate the reviewer’s suggestions.

I divided the Discussion section into subsections.

 

7.The formatting of some terms is inconsistent throughout the full text, most notably for derivatives of reactive oxygen metabolites, which are sometimes written as the plural "d-ROMs" and other times as the singular "d-ROM". The authors should read through the full text to unify the formatting and expression of terms, ensuring the rigor of the manuscript.

 

We thank the reviewer for the careful review.
We corrected the terminology throughout the manuscript for consistency.

Reviewer 2 Report

The article addresses a significant gap in mCRC management by proposing a simple, blood-based biomarker (BAP) to predict chemotherapy resistance and prognosis. The introduction effectively explains the dual role of ROS in cancer therapy and provides a strong biological basis for why high antioxidant capacity could be detrimental. The study demonstrates good internal consistency, with higher BAP levels associated with both poorer treatment response and worse disease-specific survival. The use of standardized, reproducible assays (d-ROMs and BAP) and validated response criteria (RECIST v1.1) strengthens the technical quality.

Thank you for this important and clinically relevant study. You have convincingly shown that higher baseline BAP (antioxidant capacity) is associated with poorer chemotherapy response and worse survival in stage IV CRC. This is a valuable negative prognostic biomarker that could be easily implemented.

 

Reviewer recommendations

1-The authors should clarify in the title or abstract that this is an original cohort study.

2-The introduction provides an excellent mechanistic overview. However, the discussion of prior clinical studies is brief. Add 2-3 sentences summarizing previous studies that measured antioxidant capacity (TEAC, FRAP, or BAP) in GI cancers, highlighting how the present findings concord or diverge.

3-The assay protocols (d-ROMs, BAP) are detailed. However, the specific chemotherapy regimens are not provided, only "doublet" (FOLFOX? FOLFIRI?) and "oral 5-FU agent". Add a supplementary table listing exact regimens (ex., mFOLFOX6, FOLFIRI) and targeted agents (bevacizumab, cetuximab) per patient.

4-Linear regression for ordinal response scores and Cox regression for survival are appropriate. The use of ROC to define cutoffs is common but problematic when AUC ~0.5. Consider using median BAP as the cutoff instead of ROC-derived cutoff when AUC < 0.6, and report both analyses in supplementary material.

5- Section 3.4 ("Multivariate Analysis...") is very short. Merge with 3.3 or expand to include the actual beta coefficient and CI as reported in the text (already present, but format as a table for clarity). A simple schematic figure showing "Low BAP → better response and survival" vs "High BAP → resistance and worse survival" would significantly enhance reader comprehension.

6-The language is clear, grammatically correct, and appropriately academic. Only the one translation error noted above ("drug exposure duration satisfaction") needs correction.

Specific comments and required corrections

References: Reference #33 (Therasse et al., 2000) for RECIST is outdated. Please cite the current RECIST v1.1 publication (Eisenhauer et al., Eur J Cancer 2009).

Reference #41 is incomplete ("Comparative Analysis of Serum (Anti)Oxidative Status Parameters in Healthy Persons"). Please provide the full citation or remove if unpublished.

 

1. Tables and Figures:

Table 1: The formatting is broken (column alignment is off, "Cases" column missing). The table should be re-formatted so that "n" appears under "Cases" for each variable.

Table 2: The p-value for BAP (0.003) is correctly reported, but please clarify in the legend whether this is from a Mann-Whitney U test (appropriate for non-parametric data).

Figure 2 and 3: The Y-axis is labeled "Cumulative survival" but should be "Cumulative Disease-Specific Survival" for precision. The figure legends are also incomplete (Figure 2 legend cuts off; Figure 3 legend is missing entirely from the provided PDF).

Recommendations for enhancement

The following suggestions are intended to further strengthen an already solid manuscript:

Address methodological heterogeneity: Patients received heterogeneous postoperative regimens (oral 5-FU alone, doublet, doublet + targeted therapy). Perform a sensitivity analysis restricted to the 74 patients who received doublet regimens (with/without targeted therapy) to confirm that the BAP effect is independent of regimen intensity. If sample size permits, this would considerably strengthen the conclusion.

Emphasize bioavailability and measurement challenges

The BAP assay measures ferric-reducing antioxidant capacity, which primarily reflects uric acid, albumin, and small molecules, not directly intracellular GSH or NRF2 activity. In the Discussion (limitations section), explicitly state that BAP is a circulating, non-specific antioxidant measure and may not fully represent tumor cell-intrinsic defenses (e.g., NRF2-driven GSH synthesis). Consider adding a sentence on whether correlating BAP with tumor tissue markers (e.g., NRF2, GCLC) would be valuable future work.

Harmonize structure of cancer-specific sections: The Results section jumps from multivariate analysis of response (3.4) directly to DSS with ROC-derived cutoffs (3.5), but the ROC cutoffs are not described in the Methods (they are only in Results).  Move the ROC analysis (AUCs, cutoffs, sensitivity/specificity) to the Methods section (2.4 or 2.5) as a prespecified analytical plan, rather than introducing it in Results.

Sharpen the "Future perspectives" Section: The Discussion mentions limitations (AUC near 0.5, lack of longitudinal sampling) but does not provide a clear "Future Directions" subsection. Add a short paragraph before the Conclusions: "Future prospective studies should incorporate serial BAP measurements during chemotherapy to determine whether changes in antioxidant capacity correlate with acquired resistance. Additionally, combining BAP with tumor NRF2 expression or circulating tumor DNA (ctDNA) could improve prognostic accuracy beyond BAP alone."

Please address the following mandatory corrections:

1. Correct the translation error on page 8, line 277: "drug exposure duration satisfaction" → "disease-specific survival (DSS)".
2. Re-format Table 1 so that the "n" (number of cases) appears in a dedicated column.
3. Provide the complete Figure 3 legend
4. Update reference #33 to the current RECIST v1.1 citation.

 

Optional but strongly encouraged revisions:

-Add a sensitivity analysis restricted to the 74 patients on doublet chemotherapy.

-Include a supplementary table of exact chemotherapy regimens.

-Expand the limitations section to note that BAP measures circulating, not intratumoral, antioxidant capacity.

This work is a solid contribution to the field of redox biology in oncology. With the minor revisions above, it will be suitable for publication.

The article addresses a significant gap in mCRC management by proposing a simple, blood-based biomarker (BAP) to predict chemotherapy resistance and prognosis. The introduction effectively explains the dual role of ROS in cancer therapy and provides a strong biological basis for why high antioxidant capacity could be detrimental. The study demonstrates good internal consistency, with higher BAP levels associated with both poorer treatment response and worse disease-specific survival. The use of standardized, reproducible assays (d-ROMs and BAP) and validated response criteria (RECIST v1.1) strengthens the technical quality.

Thank you for this important and clinically relevant study. You have convincingly shown that higher baseline BAP (antioxidant capacity) is associated with poorer chemotherapy response and worse survival in stage IV CRC. This is a valuable negative prognostic biomarker that could be easily implemented.

 

Reviewer recommendations

1-The authors should clarify in the title or abstract that this is an original cohort study.

2-The introduction provides an excellent mechanistic overview. However, the discussion of prior clinical studies is brief. Add 2-3 sentences summarizing previous studies that measured antioxidant capacity (TEAC, FRAP, or BAP) in GI cancers, highlighting how the present findings concord or diverge.

3-The assay protocols (d-ROMs, BAP) are detailed. However, the specific chemotherapy regimens are not provided, only "doublet" (FOLFOX? FOLFIRI?) and "oral 5-FU agent". Add a supplementary table listing exact regimens (ex., mFOLFOX6, FOLFIRI) and targeted agents (bevacizumab, cetuximab) per patient.

4-Linear regression for ordinal response scores and Cox regression for survival are appropriate. The use of ROC to define cutoffs is common but problematic when AUC ~0.5. Consider using median BAP as the cutoff instead of ROC-derived cutoff when AUC < 0.6, and report both analyses in supplementary material.

5- Section 3.4 ("Multivariate Analysis...") is very short. Merge with 3.3 or expand to include the actual beta coefficient and CI as reported in the text (already present, but format as a table for clarity). A simple schematic figure showing "Low BAP → better response and survival" vs "High BAP → resistance and worse survival" would significantly enhance reader comprehension.

6-The language is clear, grammatically correct, and appropriately academic. Only the one translation error noted above ("drug exposure duration satisfaction") needs correction.

Specific comments and required corrections

References: Reference #33 (Therasse et al., 2000) for RECIST is outdated. Please cite the current RECIST v1.1 publication (Eisenhauer et al., Eur J Cancer 2009).

Reference #41 is incomplete ("Comparative Analysis of Serum (Anti)Oxidative Status Parameters in Healthy Persons"). Please provide the full citation or remove if unpublished.

 

1. Tables and Figures:

Table 1: The formatting is broken (column alignment is off, "Cases" column missing). The table should be re-formatted so that "n" appears under "Cases" for each variable.

Table 2: The p-value for BAP (0.003) is correctly reported, but please clarify in the legend whether this is from a Mann-Whitney U test (appropriate for non-parametric data).

Figure 2 and 3: The Y-axis is labeled "Cumulative survival" but should be "Cumulative Disease-Specific Survival" for precision. The figure legends are also incomplete (Figure 2 legend cuts off; Figure 3 legend is missing entirely from the provided PDF).

Recommendations for enhancement

The following suggestions are intended to further strengthen an already solid manuscript:

Address methodological heterogeneity: Patients received heterogeneous postoperative regimens (oral 5-FU alone, doublet, doublet + targeted therapy). Perform a sensitivity analysis restricted to the 74 patients who received doublet regimens (with/without targeted therapy) to confirm that the BAP effect is independent of regimen intensity. If sample size permits, this would considerably strengthen the conclusion.

Emphasize bioavailability and measurement challenges

The BAP assay measures ferric-reducing antioxidant capacity, which primarily reflects uric acid, albumin, and small molecules, not directly intracellular GSH or NRF2 activity. In the Discussion (limitations section), explicitly state that BAP is a circulating, non-specific antioxidant measure and may not fully represent tumor cell-intrinsic defenses (e.g., NRF2-driven GSH synthesis). Consider adding a sentence on whether correlating BAP with tumor tissue markers (e.g., NRF2, GCLC) would be valuable future work.

Harmonize structure of cancer-specific sections: The Results section jumps from multivariate analysis of response (3.4) directly to DSS with ROC-derived cutoffs (3.5), but the ROC cutoffs are not described in the Methods (they are only in Results).  Move the ROC analysis (AUCs, cutoffs, sensitivity/specificity) to the Methods section (2.4 or 2.5) as a prespecified analytical plan, rather than introducing it in Results.

Sharpen the "Future perspectives" Section: The Discussion mentions limitations (AUC near 0.5, lack of longitudinal sampling) but does not provide a clear "Future Directions" subsection. Add a short paragraph before the Conclusions: "Future prospective studies should incorporate serial BAP measurements during chemotherapy to determine whether changes in antioxidant capacity correlate with acquired resistance. Additionally, combining BAP with tumor NRF2 expression or circulating tumor DNA (ctDNA) could improve prognostic accuracy beyond BAP alone."

Please address the following mandatory corrections:

1. Correct the translation error on page 8, line 277: "drug exposure duration satisfaction" → "disease-specific survival (DSS)".
2. Re-format Table 1 so that the "n" (number of cases) appears in a dedicated column.
3. Provide the complete Figure 3 legend
4. Update reference #33 to the current RECIST v1.1 citation.

 

Optional but strongly encouraged revisions:

-Add a sensitivity analysis restricted to the 74 patients on doublet chemotherapy.

-Include a supplementary table of exact chemotherapy regimens.

-Expand the limitations section to note that BAP measures circulating, not intratumoral, antioxidant capacity.

This work is a solid contribution to the field of redox biology in oncology. With the minor revisions above, it will be suitable for publication.

Author Response

 Reviewer recommendations

• The authors should clarify in the title or abstract that this is an original cohort study.

 

[Response]

Thank you for this helpful comment. We have revised the title to clarify that this manuscript reports an original cohort study. We also revised the Abstract to explicitly state that the study was an original single-center observational cohort study. These changes have been made in the revised manuscript.

 

• The introduction provides an excellent mechanistic overview. However, the discussion of prior clinical studies is brief. Add 2-3 sentences summarizing previous studies that measured antioxidant capacity (TEAC, FRAP, or BAP) in GI cancers, highlighting how the present findings concord or diverge.

 

[Response]

Thank you for this helpful suggestion. We have expanded Section 4.3, “Clinical Evidence Linking Antioxidant Capacity to Treatment Response,” in the Discussion by providing a more detailed description of previous findings on ABTS, FRAP, and TEAC, based on References 25 and 38. We have also added a description of gastric cancer studies that linked antioxidant-related indices with chemotherapy response or prognosis. These revisions are highlighted in yellow in the revised manuscript.

 

• The assay protocols (d-ROMs, BAP) are detailed. However, the specific chemotherapy regimens are not provided, only "doublet" (FOLFOX? FOLFIRI?) and "oral 5-FU agent". Add a supplementary table listing exact regimens (ex., mFOLFOX6, FOLFIRI) and targeted agents (bevacizumab, cetuximab) per patient.

 

[Response]

We thank the reviewer for this helpful comment. We have added a supplementary table listing the exact chemotherapy regimens and targeted agents used in the study cohort. The revised table specifies oral fluoropyrimidine regimens, cytotoxic doublet regimens, and doublet regimens combined with targeted agents, including bevacizumab, panitumumab, and ramucirumab.

 

• Linear regression for ordinal response scores and Cox regression for survival are appropriate. The use of ROC to define cutoffs is common but problematic when AUC ~0.5. Consider using median BAP as the cutoff instead of ROC-derived cutoff when AUC < 0.6, and report both analyses in supplementary material.

 

[Response]

We thank the reviewer for this important comment. Following the reviewer’s suggestion, we performed an additional survival analysis using the median BAP value as the cutoff, given the limited discriminatory ability of the ROC-derived cutoff. The median BAP value in this cohort was 2480 μmol/L, and patients were divided into high- and low-BAP groups according to this threshold.

We have revised Section 2.4, “Survival Outcome,” in the Materials and Methods to state that survival analysis was also performed using the median BAP value as an alternative cutoff. In this supplementary analysis, there was no significant difference in survival outcomes between the high- and low-BAP groups. We have added these results to Section 3.7, “DSS According to Redox Biomarker Levels,” and have included the corresponding survival curves in the Supplementary Material.

 

• Section 3.4 ("Multivariate Analysis...") is very short. Merge with 3.3 or expand to include the actual beta coefficient and CI as reported in the text (already present, but format as a table for clarity). A simple schematic figure showing "Low BAP → better response and survival" vs "High BAP → resistance and worse survival" would significantly enhance reader comprehension.

 

[Response]

Thank you for this helpful suggestion. We added Table 3 to present the multivariate linear regression results more clearly, including the beta coefficient, 95% confidence interval, and P-value. We also added a new schematic figure, now shown as Figure 2, to summarize the observed relationship between baseline BAPand chemotherapy response. The figure illustrates that low BAP was associated with better response, whereas high BAP was associated with treatment resistance.

 

• The language is clear, grammatically correct, and appropriately academic. Only the one translation error noted above ("drug exposure duration satisfaction") needs correction.

 

[Response]

We thank the reviewer for the careful review.

We have corrected the definition of DSS in lines 265–266 to “disease-specific survival”

 

 

Specific comments and required corrections

 

References: Reference #33 (Therasse et al., 2000) for RECIST is outdated. Please cite the current RECIST v1.1 publication (Eisenhauer et al., Eur J Cancer 2009).

 

[Response]

Thank you for your comment. We have updated Reference #33 by replacing the outdated RECIST citation by Therasse et al. (2000) with the current RECIST version 1.1 guideline by Eisenhauer et al. (2009). The manuscript and reference list have been revised accordingly.

 

Reference #41 is incomplete ("Comparative Analysis of Serum (Anti)Oxidative Status Parameters in Healthy Persons"). Please provide the full citation or remove if unpublished.

 

[Response]

Thank you for your comment. We apologize for the incomplete citation. We have corrected Reference #41 by adding the full bibliographic information in the revised reference list. We have also checked the corresponding in-text citation to ensure consistency with the corrected reference.

 

Tables and Figures:

Table 1: The formatting is broken (column alignment is off, "Cases" column missing). The table should be re-formatted so that "n" appears under "Cases" for each variable.

 

[Response]

Thank you for pointing this out. We have reformatted Table 1 and added the missing “Cases” column.

 

Table 2: The p-value for BAP (0.003) is correctly reported, but please clarify in the legend whether this is from a Mann-Whitney U test (appropriate for non-parametric data).

 

[Response]

Thank you for this helpful comment. We have clarified the Table 2 legend to state that continuous variables, including BAP, were compared using the Mann–Whitney U test. This clarification has been added to the Table 2 legend.

 

Figure 2 and 3: The Y-axis is labeled "Cumulative survival" but should be "Cumulative Disease-Specific Survival" for precision. The figure legends are also incomplete (Figure 2 legend cuts off; Figure 3 legend is missing entirely from the provided PDF).

 

[Response]
Thank you for this helpful comment. We have revised the Y-axis label in Figures 2 , 3and4 to “Cumulative Disease-Specific Survival” for precision. We have also completed the figure legends.

 

Recommendations for enhancement

 

The following suggestions are intended to further strengthen an already solid manuscript:

Address methodological heterogeneity: Patients received heterogeneous postoperative regimens (oral 5-FU alone, doublet, doublet + targeted therapy). Perform a sensitivity analysis restricted to the 74 patients who received doublet regimens (with/without targeted therapy) to confirm that the BAP effect is independent of regimen intensity. If sample size permits, this would considerably strengthen the conclusion.

 

[Response]
We thank the reviewer for this important suggestion. To address the potential influence of methodological heterogeneity in postoperative chemotherapy regimens, we performed an additional sensitivity analysis restricted to the 74 patients who received doublet regimens with or without targeted therapy.

In the multivariable linear regression analysis for therapeutic response, the model was adjusted for age, sex, number of metastatic organs, number of chemotherapy regimens, and BAP level. In this restricted cohort, higher BAP remained independently associated with a poorer therapeutic effect score (β = 0.393; 95% CI, 0.280–0.963; P < 0.001).

We also performed a survival analysis for disease-specific survival (DSS) in the same cohort. The optimal cutoff value of BAP for 3-year DSS was 2646, and the area under the receiver operating characteristic curve was 0.543. The 3-year DSS rate was 56.5% in patients with BAP ≤2646 and 35.7% in those with BAP >2646. The difference between the two groups was statistically significant by the log-rank test (P = 0.026).

Because only 32 disease-specific deaths occurred in this subgroup, we used a parsimonious multivariable Cox proportional hazards model adjusted for age, sex, number of chemotherapy regimens, and BAP level. High BAP remained an independent predictor of poorer DSS (HR, 2.223; 95% CI, 1.080–4.576; P = 0.030). These findings support the robustness of the association between high BAP and poor outcomes, even after restricting the analysis to patients treated with doublet regimens. However, because the subgroup size and number of events were limited, we have acknowledged that further validation in a larger cohort is warranted.

 

Emphasize bioavailability and measurement challenges

The BAP assay measures ferric-reducing antioxidant capacity, which primarily reflects uric acid, albumin, and small molecules, not directly intracellular GSH or NRF2 activity. In the Discussion (limitations section), explicitly state that BAP is a circulating, non-specific antioxidant measure and may not fully represent tumor cell-intrinsic defenses (e.g., NRF2-driven GSH synthesis). Consider adding a sentence on whether correlating BAP with tumor tissue markers (e.g., NRF2, GCLC) would be valuable future work.

 

[Response]

Thank you for this important comment. In the limitations section of the revised Discussion, we clarified that BAP is a circulating, non-specific ferric-reducing antioxidant capacity assay and is influenced by serum constituents such as uric acid, albumin, and other low-molecular-weight antioxidants. We also emphasized that serum BAP may not fully reflect tumor cell-intrinsic antioxidant defense mechanisms, including NRF2-driven glutathione synthesis. Finally, we added that future studies correlating BAP with tumor tissue markers, such as NRF2 and glutamate-cysteine ligase catalytic subunit, would be valuable to clarify whether circulating antioxidant capacity reflects tumor-intrinsic redox adaptation.

 

Harmonize structure of cancer-specific sections: The Results section jumps from multivariate analysis of response (3.4) directly to DSS with ROC-derived cutoffs (3.7), but the ROC cutoffs are not described in the Methods (they are only in Results).  Move the ROC analysis (AUCs, cutoffs, sensitivity/specificity) to the Methods section (2.4 or 2.5) as a prespecified analytical plan, rather than introducing it in Results.

 

[Response]

Thank you for this helpful suggestion. We agree that the ROC-derived cutoff values should be described in the Methods section as part of the prespecified analytical plan, rather than being introduced for the first time in the Results section. Accordingly, we moved the ROC analysis details, including the AUCs, cutoff values, sensitivity, and specificity for d-ROMs and BAP, from Section 3.7 to Section 2.4, “Survival Outcome.”

 

Sharpen the "Future perspectives" Section: The Discussion mentions limitations (AUC near 0.5, lack of longitudinal sampling) but does not provide a clear "Future Directions" subsection. Add a short paragraph before the Conclusions: "Future prospective studies should incorporate serial BAP measurements during chemotherapy to determine whether changes in antioxidant capacity correlate with acquired resistance. Additionally, combining BAP with tumor NRF2 expression or circulating tumor DNA (ctDNA) could improve prognostic accuracy beyond BAP alone."

 

[Response]

Thank you for this helpful suggestion. We have added a new subsection, Section 4.6. Future Perspectives, immediately before the Conclusions. In this section, we now emphasize the need for prospective studies incorporating serial BAP measurements during chemotherapy to determine whether changes in antioxidant capacity are associated with acquired resistance or treatment failure. We also added that combining BAP with tumor-based biomarkers, such as NRF2 expression, and circulating biomarkers, such as circulating tumor DNA (ctDNA), may improve prognostic accuracy beyond BAP alone.

 

 

 

 

Reviewer 3 Report

This very focussed paper describes an interesting and important observation that baseline antioxidant capacity is associated with a poorer response to chemotherapy in pateints with metastatic CRC. The paper is well written and contains no superfluous information. The potential impact might be enhanced by a more declamatory title reflecting the abstract conclusion. 

I have no additional comments.

Author Response

We thank the reviewer for the positive and constructive comments on our manuscript.