Comparative Prognostic Performance of CARWL and Naples Prognostic Score in Stage IIIC Non-Small Cell Lung Cancer Treated with Definitive Chemoradiotherapy

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This manuscript presents a retrospective comparison of the CARWL score and the Naples Prognostic Score (NPS) in 795 patients with AJCC 8th edition stage IIIC non-small cell lung cancer (NSCLC) treated with definitive concurrent chemoradiotherapy (CCRT). The authors demonstrate that both indices significantly stratify overall survival (OS), progression-free survival (PFS), and locoregional progression-free survival (LRPFS). The principal finding is that CARWL showed greater prognostic separation than NPS, with a larger OS difference between favorable and unfavorable groups (19.3 vs. 12.3 months) and a higher Harrell’s concordance index (0.672 vs. 0.603).

The study addresses an important clinical question. Despite similar TNM stage and treatment approaches, patients with unresectable stage III NSCLC frequently experience markedly different outcomes. Therefore, practical host-related prognostic markers may provide clinically meaningful information beyond conventional staging systems. The large cohort size and restriction to stage IIIC disease are strengths of the study. However, several issues regarding biological interpretation, clinical applicability, and statistical methodology require further clarification.

Major Comments

1. Distinguishing prognostic association from biological mechanism

The Discussion frequently interprets CARWL as a biologically grounded marker reflecting systemic inflammation, immune status, nutritional decline, and cancer cachexia. While this interpretation is plausible, the present study only demonstrates prognostic association and does not directly establish biological mechanisms.

In patients with stage III NSCLC, elevated CRP and reduced albumin may result from multiple processes including tumor-related inflammation, obstructive pneumonia, chronic pulmonary disease, occult infection, frailty, or other comorbidities. Likewise, pretreatment weight loss may reflect cachexia, but may also be influenced by dysphagia, chronic smoking-related disease, socioeconomic factors, or reduced oral intake due to respiratory symptoms.

I recommend that the authors more clearly distinguish prognostic association from mechanistic interpretation and avoid implying that CARWL specifically measures cachexia biology.

1. Is CARWL superior because of the score itself or because it incorporates weight loss?

The manuscript concludes that CARWL is superior to NPS. However, it remains unclear whether the observed advantage derives from the composite score itself or primarily from inclusion of pretreatment weight loss.

Weight loss is already a well-established adverse prognostic factor in locally advanced NSCLC and may independently capture functional decline, frailty, and cachexia-related vulnerability. Therefore, an important unanswered question is whether CARWL provides incremental prognostic information beyond its individual components.

If possible, the authors should perform analyses comparing:

• CAR alone
• Weight loss alone
• CARWL

Such analyses would help determine whether the composite score truly adds prognostic value beyond established predictors.

1. CARWL may reflect treatment tolerance as much as tumor biology

A key clinical issue that is not addressed is whether CARWL predicts outcome through aggressive tumor biology, impaired treatment delivery, or both.

Patients with significant weight loss and systemic inflammation may be more likely to experience:

• chemotherapy dose reductions,
• fewer delivered chemotherapy cycles,
• treatment interruptions,
• severe esophagitis,
• hospitalization,
• reduced treatment compliance.

The manuscript reports treatment-related toxicities but does not evaluate treatment completion according to CARWL category.

This issue is particularly important in a definitive CCRT population because treatment intensity itself strongly influences outcomes. If available, the authors should report treatment completion rates, delivered chemotherapy intensity, and treatment interruptions according to CARWL group.

1. Tumor burden is emphasized conceptually but not directly assessed

The Introduction appropriately highlights limitations of TNM staging and argues that anatomical staging does not fully reflect tumor burden. However, the analysis does not include any volumetric parameters such as:

• gross tumor volume (GTV),
• planning target volume (PTV),
• metabolic tumor volume,
• total lesion glycolysis.

This is particularly relevant in stage IIIC disease, where patients may have markedly different intrathoracic tumor volumes despite sharing the same stage classification.

Because tumor burden is central to the rationale of the study, incorporation of volumetric parameters would substantially strengthen the biological interpretation of the findings. If such data are unavailable, this limitation should be explicitly acknowledged.

1. Limited multivariable adjustment

The multivariable models include T stage, CARWL, and NPS but do not appear to include several clinically relevant variables such as:

• age,
• ECOG performance status,
• histology,
• smoking history,
• chemotherapy regimen,
• treatment completion,
• pulmonary function,
• comorbidity burden.

Although some variables may not have reached statistical significance in univariate analyses, their exclusion limits interpretation of independent prognostic effects.

The authors should clarify the variable-selection strategy and consider presenting a more comprehensive multivariable model.

1. Clinical utility remains insufficiently defined

An important question for readers is how CARWL would be used in clinical practice.

The manuscript demonstrates prognostic stratification but does not clearly define whether CARWL is intended as:

• a prognostic marker,
• a treatment-selection tool,
• a marker for intensified supportive care,
• a criterion for closer surveillance,
• a stratification factor in future clinical trials.

Clarification of the intended clinical application would substantially improve the translational relevance of the work.

1. Relevance in the PACIFIC era

The study population was treated between 2010 and 2020, largely preceding widespread adoption of consolidation durvalumab.

This issue extends beyond simple temporal generalizability. Inflammatory and nutritional biomarkers may influence not only prognosis after CCRT but also host immune competence and responsiveness to immunotherapy.

The authors should discuss whether CARWL might retain prognostic significance in patients receiving consolidation immunotherapy and whether future validation should focus on contemporary CCRT-plus-durvalumab cohorts.

1. The claim of superiority should be moderated

Although CARWL consistently outperformed NPS, the absolute improvement was modest.

For example:

• CARWL C-index: 0.672
• NPS C-index: 0.603

These values indicate moderate discrimination rather than highly accurate prognostic prediction.

Accordingly, statements describing CARWL as having “superior prognostic performance” should be tempered. The present data support modestly improved discrimination rather than definitive superiority.

1. External validation remains essential

Because CARWL was originally developed by the same research group, the present study should be regarded as additional institutional validation rather than independent external validation.

The findings are promising but require confirmation in multicenter cohorts, particularly those treated according to contemporary standards including consolidation immunotherapy.

Minor Comments

1. Please clarify how pretreatment weight loss data were collected and verified.
2. Please describe missing-data handling in greater detail.
3. Were patients with active obstructive pneumonia, recent antibiotic exposure, or inflammatory pulmonary conditions excluded? These factors may substantially influence CRP and albumin levels.
4. A subgroup analysis comparing T3N3 and T4N3 disease may provide additional insight regarding the robustness of CARWL across different disease burdens.
5. Reporting treatment completion, chemotherapy cycles delivered, and treatment interruptions by CARWL group would strengthen clinical interpretation.
6. Several sections of the Discussion repeat similar concepts regarding inflammation and cachexia. Condensation would improve readability.
7. Figure legends could be expanded to allow interpretation without referring back to the Methods section.

Overall Assessment

This is a clinically relevant study involving a large and relatively homogeneous cohort of stage IIIC NSCLC patients treated with definitive CCRT. The findings support the prognostic value of both CARWL and NPS and suggest that CARWL may provide modestly improved risk stratification. However, the manuscript would be substantially strengthened by deeper consideration of biological interpretation, treatment tolerance, volumetric tumor burden, and contemporary clinical applicability. I recommend Major Revision.

Comments on the Quality of English Language

The English could be improved to more clearly express the research.

Author Response

Response to Reviewer 1.

We sincerely thank Reviewer 1 for the thorough and constructive evaluation of our manuscript. The comments regarding biological interpretation, treatment delivery, multivariable adjustment, clinical utility, PACIFIC-era relevance, and external validation were highly valuable. We have revised the manuscript accordingly and believe these changes have substantially improved its methodological clarity, balance, and clinical relevance.

Major Comments

Comment 1. Distinguishing prognostic association from biological mechanism

The Discussion frequently interprets CARWL as a biologically grounded marker reflecting systemic inflammation, immune status, nutritional decline, and cancer cachexia. While this interpretation is plausible, the present study only demonstrates prognostic association and does not directly establish biological mechanisms.

In patients with stage III NSCLC, elevated CRP and reduced albumin may result from multiple processes, including tumor-related inflammation, obstructive pneumonia, chronic pulmonary disease, occult infection, frailty, or other comorbidities. Likewise, pretreatment weight loss may reflect cachexia, but may also be influenced by dysphagia, chronic smoking-related disease, socioeconomic factors, or reduced oral intake due to respiratory symptoms.

I recommend that the authors more clearly distinguish prognostic association from mechanistic interpretation and avoid implying that CARWL specifically measures cachexia biology.

Response 1. Thank you for this important comment. We agree that the present study demonstrates prognostic association rather than direct biological mechanism. We also acknowledge that elevated CRP, reduced albumin, and pretreatment weight loss may reflect multiple clinical processes, including tumor-related inflammation, occult infection, chronic pulmonary disease, frailty, impaired oral intake, and other comorbid conditions. Notably, the Discussion already emphasized that the proposed biological interpretation of CARWL should be regarded as plausible and hypothesis-generating rather than definitive. Nevertheless, to further improve clarity, we have revised and condensed the relevant Discussion section to more clearly distinguish prognostic association from mechanistic interpretation and to avoid implying that CARWL specifically measures cachexia biology. We now emphasize that CARWL should be interpreted primarily as a host-related prognostic index reflecting overlapping inflammatory, nutritional, and weight-loss domains, while future clinical and mechanistic studies are needed to clarify the biological pathways underlying its prognostic significance.

Comment 2. Is CARWL superior because of the score itself or because it incorporates weight loss?. The manuscript concludes that CARWL is superior to NPS. However, it remains unclear whether the observed advantage derives from the composite score itself or primarily from the inclusion of pretreatment weight loss. Weight loss is already a well-established adverse prognostic factor in locally advanced NSCLC and may independently capture functional decline, frailty, and cachexia-related vulnerability. Therefore, an important unanswered question is whether CARWL provides incremental prognostic information beyond its individual components. If possible, the authors should perform analyses comparing:

• CAR alone
• Weight loss alone
• CARWL

Such analyses would help determine whether the composite score truly adds prognostic value beyond established predictors.

Response 2. Thank you for this thoughtful comment. We agree that determining the relative contribution of CAR, pretreatment WL, and their composite CARWL score is important for understanding the score’s prognostic structure. However, CARWL is an established prognostic scoring system with predefined components and categories, and the primary aim of the present study was to compare the prognostic performance of the established CARWL score with that of the established NPS rather than to redevelop, recalibrate, or decompose CARWL into its individual components. The individual prognostic relevance of CAR and WL in LA-NSCLC has already been demonstrated in previous studies, including the original CARWL report. Therefore, we retained the original CARWL structure to preserve methodological consistency and allow direct comparison with NPS. We have clarified this point in the Discussion and acknowledge that future studies may further evaluate the incremental contribution of individual CARWL components.

Comment 3. CARWL may reflect treatment tolerance as much as tumor biology

A key clinical issue that is not addressed is whether CARWL predicts outcome through aggressive tumor biology, impaired treatment delivery, or both.

Patients with significant weight loss and systemic inflammation may be more likely to experience:

• chemotherapy dose reductions,
• fewer delivered chemotherapy cycles,
• treatment interruptions,
• severe esophagitis,
• hospitalization,
• reduced treatment compliance.

The manuscript reports treatment-related toxicities but does not evaluate treatment completion according to the CARWL category.

This issue is particularly important in a definitive CCRT population because treatment intensity itself strongly influences outcomes. If available, the authors should report treatment completion rates, delivered chemotherapy intensity, and treatment interruptions according to CARWL group.

Response 3. Thank you for this important comment. We agree that CARWL may reflect not only host- and tumor-related biological vulnerability but also aspects of treatment tolerance and treatment delivery during definitive CCRT. To address this concern, we performed an additional analysis of treatment delivery according to CARWL group, including chemotherapy cycles delivered, radiotherapy interruption rates and duration, hospitalization during CCRT, and acute grade 3–4 toxicities. These results are now presented in Supplementary Table S1. Notably, no significant differences were observed across CARWL groups with respect to chemotherapy cycles delivered (P = 0.92), radiotherapy interruption rates (P = 0.59), interruption duration (P = 0.82), hospitalization during CCRT (P = 0.46), or acute grade 3–4 toxicities (P = 0.71). These findings suggest that the prognostic value of CARWL cannot be readily explained by measurable differences in treatment delivery or acute treatment tolerance alone. We have added this clarification to the revised Discussion.

Comment 4 Tumor burden is emphasized conceptually but not directly assessed

The Introduction appropriately highlights limitations of TNM staging and argues that anatomical staging does not fully reflect tumor burden. However, the analysis does not include any volumetric parameters such as:

• gross tumor volume (GTV),
• planning target volume (PTV),
• metabolic tumor volume,
• total lesion glycolysis.

This is particularly relevant in stage IIIC disease, where patients may have markedly different intrathoracic tumor volumes despite sharing the same stage classification.

Because tumor burden is central to the rationale of the study, the incorporation of volumetric parameters would substantially strengthen the biological interpretation of the findings. If such data are unavailable, this limitation should be explicitly acknowledged.

Response 4. Thank you for this important observation. We agree that volumetric measures such as gross tumor volume (GTV), planning target volume (PTV), metabolic tumor volume, and total lesion glycolysis may provide a more comprehensive assessment of tumor burden than anatomical staging alone. Unfortunately, these parameters were not consistently available across the study period and therefore could not be incorporated into the present analysis. We acknowledge that patients with the same stage IIIC classification may exhibit substantial heterogeneity in tumor volume despite sharing similar TNM characteristics. Accordingly, we have added this issue to the Limitations section and agree that future studies integrating volumetric and metabolic tumor-burden parameters should evaluate whether the prognostic value of CARWL remains independent of these measures.

Comment 5. Limited multivariable adjustment

The multivariable models include T stage, CARWL, and NPS but do not appear to include several clinically relevant variables such as:

• age,
• ECOG performance status,
• histology,
• smoking history,
• chemotherapy regimen,
• treatment completion,
• pulmonary function,
• comorbidity burden.

Although some variables may not have reached statistical significance in univariate analyses, their exclusion limits interpretation of independent prognostic effects.

The authors should clarify the variable-selection strategy and consider presenting a more comprehensive multivariable model.

Response 5. Thank you for this important methodological comment. We agree that clarification of the variable-selection strategy and adjustment for clinically relevant covariates strengthen the interpretation of the independent prognostic effects. In the primary analyses, variables significantly associated with outcomes in univariate testing were incorporated into the multivariable models to preserve model parsimony and avoid unnecessary overfitting. In response to the reviewer’s suggestion, we additionally performed a sensitivity multivariable Cox regression analysis including age, sex, ECOG performance status, smoking history, histology, T stage, chemotherapy cycles, radiotherapy interruption, CARWL score, and NPS. These results are now presented in Supplementary Table S2. Importantly, both CARWL and NPS remained independently associated with overall survival after adjustment for these clinically relevant variables, supporting the robustness of the observed prognostic associations. Detailed pulmonary-function parameters and comorbidity burden were not consistently available for inclusion in the model and have been acknowledged as limitations in the revised Discussion. The Statistical Analysis section has also been revised to clarify the variable-selection strategy.

Comment 6. Clinical utility remains insufficiently defined

An important question for readers is how CARWL would be used in clinical practice.

The manuscript demonstrates prognostic stratification but does not clearly define whether CARWL is intended as:

• a prognostic marker,
• a treatment-selection tool,
• a marker for intensified supportive care,
• a criterion for closer surveillance,
• a stratification factor in future clinical trials.

Clarification of the intended clinical application would substantially improve the translational relevance of the work.

Response 6. Thank you for this important comment. We agree that the potential clinical application of CARWL should be more clearly defined. The present findings support the use of CARWL primarily as a prognostic stratification tool rather than as a treatment-selection instrument. Because CARWL is based on simple and routinely available clinical and laboratory parameters, it may help identify patients at higher risk of adverse outcomes despite sharing the same anatomical stage. Accordingly, patients with unfavorable CARWL scores may warrant closer surveillance, earlier supportive-care interventions, and consideration for enrollment in clinical trials. In addition, CARWL may serve as a useful stratification factor in future clinical studies evaluating personalized treatment approaches in LA-NSCLC. To clarify these potential applications, we have revised the relevant Discussion section.

Comment 7. Relevance in the PACIFIC era

The study population was treated between 2010 and 2020, largely preceding widespread adoption of consolidation durvalumab.

This issue extends beyond simple temporal generalizability. Inflammatory and nutritional biomarkers may influence not only prognosis after CCRT but also host immune competence and responsiveness to immunotherapy.

The authors should discuss whether CARWL might retain prognostic significance in patients receiving consolidation immunotherapy and whether future validation should focus on contemporary CCRT-plus-durvalumab cohorts.

Response 7. Thank you for this important comment. We agree that the relevance of CARWL in the contemporary PACIFIC era extends beyond simple temporal generalizability. Inflammatory, nutritional, and weight-loss-related biomarkers may influence not only prognosis after CCRT but also host immune competence and responsiveness to immunotherapy. Because the present cohort was treated largely before the routine adoption of consolidation durvalumab, we were unable to directly evaluate the prognostic performance of CARWL in patients receiving immunotherapy. We have therefore expanded the Discussion to emphasize that future validation studies should focus on contemporary CCRT-plus-durvalumab cohorts and should investigate whether CARWL retains prognostic significance in the setting of consolidation immunotherapy.

Comment 8. The claim of superiority should be moderated

Although CARWL consistently outperformed NPS, the absolute improvement was modest.

For example:

• CARWL C-index: 0.672
• NPS C-index: 0.603

These values indicate moderate discrimination rather than highly accurate prognostic prediction.

Accordingly, statements describing CARWL as having “superior prognostic performance” should be tempered. The present data support modestly improved discrimination rather than definitive superiority.

Response 8. Thank you for this important comment. We agree that the observed improvement in discrimination, although consistent across analyses, should be interpreted as modest rather than definitive evidence of superior predictive performance. Accordingly, we have tempered the relevant statements throughout the manuscript to avoid overstatement. The revised text now emphasizes that CARWL demonstrated a modest but consistent improvement in prognostic discrimination compared with NPS, rather than definitive superiority. We also clarified that the observed C-index values indicate moderate discriminatory performance and that external validation is required before CARWL can be considered a clinically definitive prognostic tool.

Comment 9. External validation remains essential

Because CARWL was originally developed by the same research group, the present study should be regarded as additional institutional validation rather than independent external validation.

The findings are promising but require confirmation in multicenter cohorts, particularly those treated according to contemporary standards including consolidation immunotherapy.

Response 9. Thank you for this important comment. We agree that the present study should be considered an additional institutional validation rather than an independent external validation of the CARWL score. Although the large cohort size, homogeneous stage IIIC population, and consistent treatment approach strengthen the findings, confirmation in independent multicenter cohorts remains essential. We have revised the Discussion to more explicitly acknowledge this limitation and to emphasize the need for external validation, particularly in contemporary cohorts treated with consolidation durvalumab following definitive CCRT.

Minor Comments

Comment 1. Please clarify how pretreatment weight loss data were collected and verified.

Response 1. Thank you for this comment. The definition and ascertainment of pretreatment WL followed the methodology used in the original CARWL study, as the present study aimed to compare the established CARWL and NPS scoring systems rather than redefine their components. Briefly, pretreatment WL was determined from institutional clinical records by comparing body weight at CCRT initiation with the documented body weight during the preceding six months. This clarification has been added to the Methods section.

Comment 2. Please describe missing-data handling in greater detail.

Response 2. Thank you for this important methodological comment. The study was conducted using a complete-case approach. Patients lacking any information required for eligibility assessment, calculation of CARWL or NPS, treatment characterization, or survival analyses were excluded during cohort assembly. Because both CARWL and NPS require complete laboratory and clinical data for score assignment, imputation methods were not used. To clarify this issue, a statement describing the complete-case approach has been added to the Statistical Analysis section.

Comment 3. Were patients with active obstructive pneumonia, recent antibiotic exposure, or inflammatory pulmonary conditions excluded? These factors may substantially influence CRP and albumin levels.

Response 3. Thank you for this important comment. Patients with active infectious diseases or those receiving immunosuppressive therapy within 30 days before CCRT initiation were excluded because such conditions may substantially affect immune-inflammatory biomarkers, including CRP and albumin levels. We agree that obstructive pneumonia, recent antibiotic exposure, and other inflammatory pulmonary conditions may also influence these parameters. However, because of the retrospective nature of the study, detailed and standardized information regarding recent antibiotic use and all inflammatory pulmonary conditions was not consistently available for analysis. To address this concern, we have clarified the exclusion criteria in the Methods section and acknowledged the potential influence of unmeasured inflammatory conditions as a limitation of the study.

Comment 4. A subgroup analysis comparing T3N3 and T4N3 disease may provide additional insight regarding the robustness of CARWL across different disease burdens.

Response 4. Thank you for this thoughtful suggestion. We agree that evaluating the performance of CARWL according to T3N3 and T4N3 disease may provide additional insight regarding its robustness across different levels of intrathoracic tumor burden. However, the primary objective of the present study was to compare the prognostic performance of CARWL and NPS in a homogeneous stage IIIC NSCLC cohort rather than to conduct multiple subgroup analyses. Because the study was not specifically designed or powered for detailed subgroup comparisons, such analyses may yield unstable estimates and increase the risk of spurious findings. Nevertheless, T stage was incorporated into the multivariable analyses, where it remained independently associated with outcomes. We have acknowledged this issue in the Discussion and agree that future studies incorporating more detailed measures of tumor burden should evaluate the consistency of CARWL across T-stage subgroups.

Comment 5. Reporting treatment completion, chemotherapy cycles delivered, and treatment interruptions by CARWL group would strengthen clinical interpretation.

Response 5. Thank you for this valuable suggestion. We agree that treatment delivery and treatment tolerance are clinically relevant considerations when interpreting the prognostic significance of CARWL. In response to the reviewer’s comment, we performed an additional analysis according to CARWL group, including chemotherapy cycles delivered, radiotherapy interruption rates and duration, hospitalization during CCRT, and acute grade 3–4 toxicities. These results are now presented in Supplementary Table S1. Notably, no significant differences were observed among the CARWL groups with respect to chemotherapy cycles delivered (P = 0.92), radiotherapy interruption rates (P = 0.59), interruption duration (P = 0.82), hospitalization during CCRT (P = 0.46), or acute grade 3–4 toxicities (P = 0.71). We have also incorporated these findings into the Discussion to facilitate clinical interpretation of the observed prognostic associations.

Comment 6. Several sections of the Discussion repeat similar concepts regarding inflammation and cachexia. Condensation would improve readability.

Response 6. Thank you for this helpful suggestion. We agree that some parts of the Discussion repeated similar concepts regarding systemic inflammation, nutritional decline, and cancer cachexia. Accordingly, the Discussion has been revised and condensed to improve readability, reduce conceptual redundancy, and more clearly distinguish prognostic association from mechanistic interpretation.

Comment 7. Figure legends could be expanded to allow interpretation without referring back to the Methods section.

Response 7. Thank you for this helpful suggestion. We agree that the figure legends should allow readers to interpret the figures independently. Accordingly, the legends have been expanded to define the plotted survival endpoints, score groups, color coding, and C-index interpretation without requiring readers to refer back to the Methods section.

 

 

 

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

This article is interesting, but there are some points to revise before publishing.

1. Why are there restrictions on eligibility criteria based on age and BMI, even though this is a retrospective study? Is this related to the statement that the cases were collected prospectively partway through the study? Please explain in detail, including the statement that the study was prospective.
2. Please describe in detail how you differentiated between chemotherapy regimens. In a prospective registration, there should be clear criteria for making those decisions.
3. Could you show the detail of chemotherapy regimen in table1?
4. Please explain in detail why you divided the data at age 70 instead of the median.
5. Both scores appear to have a difference between 0 and 1/2, but there doesn't seem to be much difference between 1 and 2. Please examine whether there is a difference between the 1 and 2 groups for each score. If not, please discuss whether it is appropriate to divide into three groups in the first place.

Author Response

Response to Reviewer 2

We sincerely thank Reviewer 2 for the careful assessment of our manuscript and for the practical suggestions regarding study design clarification, chemotherapy-regimen reporting, age categorization, and interpretation of score groups. We have addressed each comment in detail and revised the manuscript to improve transparency, readability, and interpretability.

Comment 1.  Why are there restrictions on eligibility criteria based on age and BMI, even though this is a retrospective study? Is this related to the statement that the cases were collected prospectively partway through the study? Please explain in detail, including the statement that the study was prospective.

Response 1. Thank you for this important comment. We agree that the original wording may have caused confusion regarding the study design. Although the present study was retrospective in design, it was conducted using a prospectively maintained institutional database in which demographic, clinical, laboratory, treatment, and follow-up data were routinely recorded according to predefined institutional standards. Therefore, the age (18–80 years) and BMI (>18.5 kg/m²) criteria were not retrospectively imposed for the purposes of this study but reflected the institutional eligibility criteria routinely applied for curative-intent definitive concurrent chemoradiotherapy during the study period. These criteria were used to maintain a clinically homogeneous cohort and to minimize potential confounding related to frailty, severe undernutrition, and treatment tolerance. To avoid ambiguity, we have revised the Methods section to clarify that this was a retrospective analysis of a prospectively maintained institutional database and to explain the rationale for the age and BMI eligibility criteria.

Comment 2. Please describe in detail how you differentiated between chemotherapy regimens. In a prospective registration, there should be clear criteria for making those decisions.

Response 2.  Thank you for this valuable comment. Chemotherapy selection was protocol-driven and followed the institutional multidisciplinary treatment guidelines in effect during the study period. Cisplatin-based regimens were preferred whenever clinically appropriate, whereas carboplatin-containing regimens were reserved for patients in whom cisplatin was considered unsuitable because of contraindications or safety concerns. Regimen allocation was based on predefined clinical considerations, including performance status, comorbidities, organ-function parameters, and anticipated treatment tolerance. The choice of accompanying agent (docetaxel, paclitaxel, or vinorelbine) similarly followed institutional treatment protocols and contemporary standards of care. To clarify this issue, additional details regarding chemotherapy administration and regimen selection have been added to the Methods section.

Comment 3. Could you show the detail of chemotherapy regimen in table1?

Response 3. Thank you for this valuable suggestion. We agree that detailed reporting of chemotherapy administration improves the transparency and interpretability of the study findings. Accordingly, Table 1 has been revised to include the distribution of concurrent chemotherapy regimens, including cisplatin- and carboplatin-based combinations with docetaxel, paclitaxel, or vinorelbine, as well as the number of concurrent chemotherapy cycles received by the patients. In addition, a brief description of chemotherapy administration has been incorporated into the Results section. As shown in the revised Table 1, the distributions of chemotherapy regimens and chemotherapy cycles were comparable across both CARWL and NPS groups, indicating that treatment characteristics were generally well balanced among the prognostic subgroups.

Comment 4.  Please explain in detail why you divided the data at age 70 instead of the median.

Response 4. Thank you for this thoughtful comment. Age was categorized using a 70-year threshold because this cutoff is widely used in thoracic and geriatric oncology research to distinguish older patients who may differ in treatment tolerance, comorbidity burden, toxicity profiles, and clinical outcomes. Our intention was to preserve clinical interpretability and facilitate comparison with previous studies rather than to apply a cohort-specific statistical cutoff. We did not use the median age because median-based dichotomization is population-dependent and may reduce comparability across studies. We agree that modeling age as a continuous variable may be preferable in prognostic analyses; however, in the present study age was included only as a descriptive baseline characteristic rather than as a primary prognostic variable. To clarify this rationale, we have added an explanatory note to Table 1.

Comment 5.  Both scores appear to have a difference between 0 and 1/2, but there doesn't seem to be much difference between 1 and 2. Please examine whether there is a difference between the 1 and 2 groups for each score. If not, please discuss whether it is appropriate to divide into three groups in the first place.

Response 5. Thank you for this important comment. We agree that examining the separation between adjacent score categories is relevant for interpreting the prognostic performance of both indices. However, both CARWL and NPS are established scoring systems with predefined three-tier classifications, and the primary aim of the present study was to compare their relative prognostic performance rather than to redefine or question their original scoring structures. Therefore, we retained the established three-group classifications to preserve methodological consistency with the original studies and to enable direct comparison between the two indices.

Following the reviewer’s suggestion, we additionally examined the survival differences between adjacent groups. Although the magnitude of separation was more pronounced between the favorable group and the remaining groups, stepwise numerical deterioration was observed from group 0 to group 1 and from group 1 to group 2 for both scoring systems across survival endpoints. These findings support the use of the original three-tier classifications, while also indicating that the strongest clinical separation occurs between the most favorable and less favorable prognostic categories. This point has been clarified in the revised Discussion.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

I thank the authors for their thorough and thoughtful revision of the manuscript. The responses have adequately addressed my major concerns, and the manuscript has been substantially strengthened.

In particular, the additional analyses regarding treatment delivery and treatment tolerance, the expanded multivariable sensitivity analyses, the clearer discussion of the biological interpretation of CARWL, and the more balanced discussion of its clinical applicability and relevance in the contemporary immunotherapy era have significantly improved the scientific rigor and clinical relevance of the study.

I also appreciate the authors’ efforts to moderate statements regarding the superiority of CARWL and to more clearly acknowledge the need for independent external validation.

Although some limitations remain inherent to the retrospective single-institution design, particularly regarding the absence of volumetric tumor burden measures and external validation, I believe these limitations are appropriately acknowledged and do not detract from the overall value of the study.

The manuscript now provides a balanced and clinically meaningful comparison between CARWL and NPS in a large cohort of stage IIIC NSCLC patients treated with definitive chemoradiotherapy.

I have no further major concerns and support acceptance of the manuscript in its current form.

Reviewer 2 Report

Comments and Suggestions for Authors

The authors have answered the questions I posed, so I will leave the final decision on whether or not to accept it to the editor.