Prognostic Impact of Baseline Neutrophil-to-Lymphocyte Ratio and Its On-Treatment Change on Survival Outcomes in Advanced Small-Cell Lung Cancer: A Retrospective Analysis

Comments 1:

In this retrospective analysis, the prognostic role of NLR in advanced SCLC is evaluated. According to the authors' results, pretreatment high NLR was associated to time to treatment failure and overall survival. Additionally, an increase in NLR within the first 6 months after treatment was associated to OS. Additionally, combined assessment of pre-treatment NLR and NLR increase on treatment defined a low prognostic group of patients.
 Although NLR has been widely evaluated as a prognostic factor in solid tumors the inclusion of NLR fluctuation during treatment is of interest. The paper is overall well written and informative and the methodology is elegant. However, I have minor suggestion that may improve the overall quality of the paper. The introduction is a little bit misleading as it focuses, expecially in the first parts, on predictive factors of immuno checkpoint inhibitors response. However, the study is not powered to analyse predictive factors as the number of patients receiving ICI is limited.

Response 1:

Thank you for this insightful comment. We agree that the original Introduction may have overemphasized predictive biomarkers for immune checkpoint inhibitor (ICI) response, which could be misleading given the limited proportion of patients receiving ICIs in this cohort. Accordingly, we have revised the Introduction to reduce and generalize the discussion of PD-L1 and tumor mutational burden, emphasizing instead the lack of practical biomarkers in SCLC and the prognostic relevance of systemic inflammatory markers. In addition, we have explicitly acknowledged in the Limitations section that only a minority of patients received ICIs.

These revisions can be found in the Introduction (page 2, lines 57–60) and Discussion—Limitations (page 12, lines 311–314).

Comments 2:

In figure 2 p value is <0.01 while in the text is 0.001.

Response 2:

Thank you for pointing out this inconsistency. We have carefully checked the statistical outputs and standardized the reporting format across the manuscript. The P-value in the text has been revised to “P < 0.01” to match Figure 2.

This correction has been made in the Results section (page 5, lines 149).

Comments 3:

Overall in the kaplan meier plots time is grouped in decile. As most of the patients are censored within the first year maybe a 6 months groups with a 32 months end may be more informative.

Response 3:

 Thank you for this helpful suggestion. We agree that the original Kaplan–Meier plots using default decile time grouping reduced readability, particularly given that most censoring occurred within the first year. Accordingly, we have regenerated all Kaplan–Meier figures with the maximum time axis limited to 32 months to better reflect the follow-up distribution and improve interpretability. Due to graphical settings in the statistical software, time ticks could not be displayed at exact 6-month intervals; therefore, 5-month intervals were used instead. We hope the reviewer will agree that this modification improves clarity while maintaining accurate representation of the data.

The revised figures have been applied to Figures 2–4 and Supplementary Figures 1–3.

Comments 4:

 In the combined analysis group a, b and c may be called in a more informative way. Like NLR<5 and deltaNLR<0; NLR>5 or delta NLR>0, NLR>5 and deltaNLR>0. This can be easier to read.

Response 4:

 We agree with this helpful suggestion. The group labels have been revised throughout the manuscript and figures to explicitly describe the NLR/ΔNLR combinations instead of alphabetical labels.

These changes have been made in the Results section (page 8, lines 204–211; lines 220-221), Figure 4 legend, and Table 4.

Comments 5:

The association between continuous NLR and TTF and OS is not discussed. Especially based on the different treatment groups.

Response 5:

 Thank you for this important comment. We agree that the implications of continuous NLR analyses stratified by treatment modality required clearer discussion. We have expanded the Discussion to explicitly interpret the continuous NLR findings across chemoimmunotherapy and chemotherapy-alone groups and their biological implications.

 This addition appears in the Discussion (page 11, lines 297–301).

Comments 1 & 2:

1) The paper does not sufficiently recognize previous work done on the NLR and small cell lung cancer including both a large meta-analysis and a validation of several proposed scales.  In particular the two references should be discussed:

BMC Pulm Med. 2024 Dec 22;24:630.

Front Immunol. 2025 Nov 7:16:1681658.

The MDACC scoring methodology was validated on a larger cohort with SCLC already.

2) As noted above, the MDACC criteria were validated on a larger cohort.  Ishihara and colleagues should explain in detail how their cutoff criteria of a NLR value of 5 was derived and how this may be superior or more informative than the previously validated scales.

Response 1 & 2:

We really appreciate these important comments. We agree that the current consensus regarding the prognostic role of NLR in SCLC, as well as validated composite scoring systems such as the MDACC-NLR score, should be more clearly acknowledged and positioned relative to the present study.

Accordingly, we have revised the Introduction and Discussion to explicitly incorporate recent evidence, including the meta-analysis in BMC Pulmonary Medicine and the MDACC-NLR validation study in Frontiers in Immunology, which support NLR as an established inflammation-based prognostic marker and confirm the utility of composite NLR-based risk scores in larger SCLC cohorts.

We have also clarified the rationale for selecting an NLR cutoff of 5. In this study, the cutoff was predefined based on prior clinical literature and meta-analytic evidence rather than derived from cohort-specific ROC analysis, which is statistically suboptimal for censored survival endpoints. Importantly, we do not claim superiority over validated composite scales such as the MDACC score. Rather, the present study addresses a complementary objective by focusing on a single, routinely available biomarker and its early on-treatment dynamics.

To emphasize this distinction, we have expanded the Discussion to note that composite indices provide comprehensive baseline risk stratification, whereas our approach evaluates whether baseline NLR combined with its dynamic change can offer a simple and repeatedly assessable prognostic framework during treatment.

Comments 1:

The authors have chosen an arbitrary cutoff of 5 for the NLR. While they state that this is objective, there needs to be a better rationale for this choice. Would the differences in clinical outcomes change if the cutoff were different? This could be reported.

Response 1:

Thank you for this important comment. We agree that further clarification of the rationale for selecting an NLR cutoff of 5 and evaluation of alternative thresholds are warranted. In this study, the cutoff was predefined based on prior clinical literature and meta-analytic evidence rather than derived from cohort-specific ROC analysis.

To address the reviewer’s concern, we have now explicitly reported exploratory analyses using alternative cutoffs (3 and 4), which showed consistent associations with both TTF and OS. Among these thresholds, NLR ≥5 provided the greatest survival separation while maintaining a clinically meaningful subgroup size.

These clarifications have been added to the Results (page 5, lines156-160) and Discussion (page 11, lines272-287)

Comments 2:

 The authors group patients into 3 groups for analysis. One with NLR<5 and ΔNLR≤0, one with one of the factors positive, and the third group with both factors positive. It would be important to know which factor (NLR or the change in NLR) is the more important determinant of clinical outcome. Reanalysis of the data would answer this question.

Response 2:

 We appreciate this important suggestion. To clarify the relative prognostic contribution of baseline NLR and ΔNLR, we have emphasized the multivariate analyses showing that baseline NLR was consistently associated with both TTF and OS, whereas ΔNLR was independently associated with OS but not TTF. These findings indicate that baseline NLR represents the primary prognostic determinant, with on-treatment change providing complementary stratification.

 This clarification has been added to the Supplementary figure 3, Results (page 8, lines 200-202) and Discussion (page 11, 272-297).

Comments 3:

 The authors should speculate as to why there are differences in the correlations with chemoimmunotherapy, compared to just chemotherapy.

Response 3:

 Thank you for this careful comment. We agree that biological interpretation is warranted. We have expanded the Discussion to propose that systemic inflammatory status may more directly reflect host immune competence and tumor–immune interactions relevant to immune checkpoint inhibitor efficacy.

　We have added the corresponding explanation in the Discussion (page 11, lines 297-301).

Comments 4:

Finally, if possible is there any mutational data with these patients that would allow a correlation of the specific subclass of SCLC based on mutations, with NLR values?

Response 4:

Thank you for this thoughtful question. We agree that evaluation of molecular subclassification in relation to NLR represents an important next step. However, molecular subclassification data were not available in this retrospective cohort because tumor genomic or transcriptional profiling was not routinely performed during the study period. Such analyses would require a large number of cases and substantial additional complexity; therefore, this remains an important topic for future investigation. We have clarified this limitation in the manuscript.

This statement has been added to the Limitations (page 11-12, lines 307-311).