Impact of Pre-Treatment NLR and Other Hematologic Biomarkers on the Outcomes of Early-Stage Non-Small-Cell Lung Cancer Treated with Stereotactic Body Radiation Therapy
Round 1
Reviewer 1 Report
I have read with great interest this manuscript concerning the potential impact of immunological biomarkers for lung SBRT. The paper is well written, data are properly collected and presented. I recommend it for publication in the current version. I only suggest to include among the references the following article in support of the efficacy of lung SBRT
-Radiol Med. 2018 Jun;123(6):406-414. doi: 10.1007/s11547-018-0858-7
Author Response
Reply to Reviewer 1
Dear reviewer,
We sincerely thank reviewer 1 for the review and welcome the comments.
As recommended, we have referenced the suggested article - Radiol Med. 2018 Jun;123(6):406-414. doi: 10.1007/s11547-018-0858-7.
The article can be found in the bibliography section as reference no 14.
Reviewer 2 Report
This study explores the association of NLR, MLR and PLR with RFS and OS in early stage NSCLC treated with SABR. NLR was associated with RFS and OS.
Comments:
The study is generally well written. However, the sample size is somewhat small to answer this clinical question. Additionally, the variable time frame of labs and treatment start is a significant limitation.
Did patients have PET staging, and was SUV considered in the MVA?
Charleson Comorbidity Index (or other similar comorbidity indices) should have also been considered in MVA as potentially the NLR is a surrogate for other medical comorbidities.
Was local control evaluated as well? Was this associated with NLR?
Was tumour size considered?
Additional look into more risk factors that could be infleuncing OS and RFS is warranted.
Author Response
Reviewer 2 - Comments and Suggestions for Authors
This study explores the association of NLR, MLR and PLR with RFS and OS in early stage NSCLC treated with SABR. NLR was associated with RFS and OS.
Comments:
The study is generally well written. However, the sample size is somewhat small to answer this clinical question. Additionally, the variable time frame of labs and treatment start is a significant limitation.
Did patients have PET staging, and was SUV considered in the MVA?
Charleson Comorbidity Index (or other similar comorbidity indices) should have also been considered in MVA as potentially the NLR is a surrogate for other medical comorbidities.
Was local control evaluated as well? Was this associated with NLR?
Was tumour size considered?
Additional look into more risk factors that could be infleuncing OS and RFS is warranted.
Reply to Reviewer 2
We sincerely thank reviewer 2 for the review and welcome the comments. We are pleased to share we have performed additional analysis and revisions as per your comments.
Please see summary of revisions below.
Comment 1: The study is generally well written. However, the sample size is somewhat small to answer this clinical question. Additionally, the variable time frame of labs and treatment start is a significant limitation.
Response to comment 1: We appreciate the comment and agree with the feedback.
We have elaborated these limitations of current study of small sample size in limitations section. We hope our study and previously reported studies (reference 23,24,25) with comparable small number of patients will add to growing evidence of this topic.
We selected the 3 months’ time frame from CBC to SBRT was chosen to be consistent with previous literature and enhance comparability of present study with previously reported literature. We completely agree with the comment it could be a limitation and have elaborated this aspect in limitations section. The relevant limitations section reads as “In addition, CBC draws from patients in this study were up to 3 months prior to initiation of SBRT, unlike other studies in surgical and chemotherapy series where CBC was done within one week to 1 month before treatment. It is unclear if the wider lead-time of CBC testing in our study had any material impact on the study results.”
Comment 2: Did patients have PET staging, and was SUV considered in the MVA?
Response to comment 2: We confirm the patient had Pre-treatment staging assessments included Positron emission tomography (PET) and/or computed tomography (CT), chest – abdomen and cranial imaging (CT or MRI). Pre-treatment PET SUV max values were available in 59 of 61 cases. The mean SUV value was 10.4 with standard deviation on 6.4.
We planned to assess PET SUV for univariate and multivariate analysis. In current analysis, the PET SUV was assessed in univariate analysis and did not have significant effect on overall survival (OS), recurrence free survival (RFS) or local recurrence free survival (LRFS). It was not included in multivariate model for current analysis. Please note we have amended the manuscript and table 1 to reflects these changes.
Comment 3: Charleson Comorbidity Index (or other similar comorbidity indices) should have also been considered in MVA as potentially the NLR is a surrogate for other medical comorbidities.
Response to comment 3: We agree and welcome the comment. Given the retrospective nature of the study, these CCI details could not be retrieved for entire population. This remains a limitation of the current study and confounding could not be excluded. We evaluated ECOG status in the current study and was included in the multivariate model. We have revised the limitation section to reflect confounding could not be excluded.
Revised section reads as “Furthermore, the impact of baseline comorbidity (Charleson Comorbidity Index or other similar comorbidity indices) and immunomodulatory or anti-inflammatory medications could not be evaluated, and confounding could not be excluded.”
We intend to assess and evaluate the impact of baseline comorbidity and immunomodulatory or anti-inflammatory medications in future projects that we plan to conduct using additional details / databases.
Comment 4: Was local control evaluated as well? Was this associated with NLR?
Response to comment 4: We welcome the suggestion from reviewer 2. As requested, we have conducted additional analysis to evaluate local recurrence free survival. The results and discussion section are amended to reflect these changes.
New addition section 3.4 reads as “On univariable analysis higher NLR was associated with poor RFS (P=0.01; NLR^1 HR-0.73; 95% CI 0.44-1.21; NLR^2 HR-1.05; 95% CI 1-1.1). The relationship between NLR and LRFS was nonlinear, and a polynomial function was used (Figure 6). The best LRFS was associated with NLR between 2.0-4.0 and the worst was above values 6. There was no statistically significant association between both MLR and PLR and LRFS in univariate analysis (P=0.18; HR-2.71; 95% CI 0.63-11.69 and P=0.958; HR-1.01; 95% CI 0.68-1.50 respectively). On multivariable analysis, higher NLR was also associated with poor LRFS (P=0.021; NLR^1 HR-0.74; 95% CI 0.44-1.23; NLR^2 HR-1.05; 95% CI 1-1.1), and the delta AIC between two models was 4.09, implying a moderate impact of NLR on RFS. The 1-index of adequacy was 0.61 and scaled integrated brier of 0.14. In our cohort MLR and PLR were not associated with LRFS in multivariable models (P=0.212; HR-2.72; 95% CI 0.56-13.14 and P=0.935; HR-0.98; 95% CI 0.66-1.47). For MLR and PLR, the delta AIC was less than 2, suggesting weak or no impact. LRFS survival curves were calculated based on the 10th, 50th and 90th percentile of NLR, MLR and PLR are shown in figure (Figure 7), which demonstrates the larger impact of NLR on LRFS than MLR and PLR through larger differences in LRFS estimates. The larger impact of NLR relative to MLR and PLR is also demonstrated with higher 1 minus index of adequacy values and scaled Brier increases (Table 4).”
Additional major changes are made to relevant discussion section and reads as “In contrast to previous reports, our data suggest a NLR correlated with RFS and LRFS.23,24 The previous studies reported cutoff values, and this may have decreased the sensitivity to detect relationship.
We did not use cutoff values for NLR, and our data demonstrated that, unlike for OS, the relationship of NLR with RFS and LRFS was nonlinear, with the risk of relapse increased when NLR values fell outside the optimal range. In our group of patients with early-stage NSCLC treated with SBRT had a lower RFS with NLR values between 3.5 and 4.5. Similarly, we found lower LRFS with NLR values between 2 and 4. To the best of our knowledge, this is the first study reporting the nonlinear relationship between NLR-RFS. And NLR-LRFS. The nonlinear relationship may indicate an optimal immunological environment is needed for optimal SBRT action and tumoricidal mechanisms.”
Comment 5: Was tumour size considered?
Response to comment 5: Thank you for the comment. We confirm the tumour size was consider in this study. The median tumour size was 2.1 cm with range of 1.1 to 5.2 cm. These changes were added to table 1. Additionally, ITV as a volumetric measure of tumour was included included in univariate and multivariate analysis.
Concluding repones to reviewer 2:
We thank reviewer 2 for the review and comments. As elaborated above, we have performed additional analysis and major revisions to improve the content and relevant topics in this article. We hope these changes will benefit the article and offer greater clarity to the readership.
Author Response File: 
Author Response.docx
