Longitudinal Assessment of FT3 to FT4 Conversion Ratio in Predicting the Efficacy of First-Line Pembrolizumab-Based Therapy in Advanced Non-Small Cell Lung Cancer: A Propensity-Score Matching Analysis of Data from the National Drug Monitoring Agency

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Nice work, congrats!

Abstract: Ok, but I would prefer it shorter

Introduction: Ok

Materials: Ok

Results: Ok, but the difference (0,18-0,22) is not quite enough to establish a straight forward idea. Few patients more or less could change the %

Discussion: accepting all the limitations you said I add one, have you checked the probability of autoimmune diseases, including thyroiditis? Please consider this option to put it in the research.

Conclusion: Ok

 

Author Response

Response to Reviewer 1 Comments

Nice work, congrats!

Thank you for these insightful comments.

Abstract: Ok, but I would prefer it shorter

We do agree that the abstract was too descriptive and redundant. Following the Reviewer's suggestion, we substantially shortened its length (from 541 to 420 words).

Introduction: Ok

Materials: Ok

Results: Ok, but the difference (0,18-0,22) is not quite enough to establish a straight forward idea. Few patients more or less could change the %

The difference in the FT3/FT4 ratio between time points 1 and 2 does not actually appear very pronounced. In this regard, the Reviewer's observation is also supported by the magnitude of IQRs that partially overlap. However, this difference was statistically significant (p<0.001) at univariate analysis across all comparisons. Moreover, the changes in FT3 and FT4 values were much more prominent than their ratio. This would suggest that it is the statistical index of the “ratio” between two continuous variables that attenuates its magnitude and thus the variability. Given the small tendency for variation in a ratio measure, we could argue that statistically significant change could be considered viable even if based on the deviation of only a few patients.

Discussion: accepting all the limitations you said I add one, have you checked the probability of autoimmune diseases, including thyroiditis? Please consider this option to put it in the research.

Autoimmune thyroid diseases are likely to have had a marginal impact, as patients with abnormal TSH at baseline or those taking levothyroxine supplementation were not considered eligible for the current research. Moreover, as much as we tried to limit the effects of immune-related adverse events on the change in fT3/fT4 ratio by excluding from survival analysis patients developing clinically relevant hypothyroidism, we cannot rule out the impact of subclinical toxicities. According to the Reviewer's suggestion, we have added this last consideration as an additional limitation of the study.

Conclusion: Ok

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

This study explores the relationship between thyroid function, particularly the fT3/fT4 conversion ratio, and survival outcomes in patients with advanced NSCLC undergoing pembrolizumab-based therapy. I have a few suggestions:

1. The authors calculate the fT3/fT4 ratio for each patient at baseline (Time Point 1) and after 12 weeks (Time Point 2). How is the change in the fT3/fT4 ratio associated with short-term efficacy, such as the ORR?

2. Time Point 2 is set at 12 weeks post-treatment. I am curious about the typical interval for efficacy evaluations. How many patients had already experienced disease progression before Time Point 2? Why was 12 weeks specifically chosen for Time Point 2? Would it be more appropriate to use the first efficacy evaluation time point for this analysis?

3. I am also interested in the authors' perspective on the insufficient correlation between TSH levels and prognosis.

Author Response

Response to Reviewer 2 Comments

This study explores the relationship between thyroid function, particularly the fT3/fT4 conversion ratio, and survival outcomes in patients with advanced NSCLC undergoing pembrolizumab-based therapy. I have a few suggestions:

We greatly appreciate your review of our manuscript and the insightful suggestions.

The authors calculate the fT3/fT4 ratio for each patient at baseline (Time Point 1) and after 12 weeks (Time Point 2). How is the change in the fT3/fT4 ratio associated with short-term efficacy, such as the ORR?

We did not consider the overall response rate (ORR) as a reliable measure of efficacy because of the unavailability of an independent review of treatment response. Following the best recommendations for real-world studies, our efficacy analysis relied on time-dependent measures such as PFS and OS, which depend on events (disease progression and mortality) that each investigator can interpret more reliably. However, we obviously performed ORR assessment, which, as described in the study method, was blinded but not independent. At the first disease restaging, performed within 12 to 16 weeks after treatment initiation, We described an ORR of 35.3% (95% CI 28.1-43.1) [resulting from 4 complete (2.4%) and 55 partial responses (32.9%)], 49 disease stabilizations (28.1%, 95% CI 21.5-35.6), and 61 disease progressions (36.5%, 95% CI 29.2-44.3). Baseline fT3/fT4 ratio did not differ between patients who achieved an ORR [median value 0.19 (IQR 0.17-0.22)] compared with all others [median value 0.18 (IQR 0.16-0.19); p = 0.930]. Conversely, post-treatment fT3/fT4 ratio was significantly higher in patients showing an ORR [median value 0.27 (IQR 0.22-0.30)] than in the remaining subgroup [median value 0.19 (IQR 0.17-0.25); p < 0.001]. Even with the methodological limitations already described, we have added the latter analysis in subsection “3.2 Dynamic Changes in Thyroid Function” of the article.

Time Point 2 is set at 12 weeks post-treatment. I am curious about the typical interval for efficacy evaluations. How many patients had already experienced disease progression before Time Point 2? Why was 12 weeks specifically chosen for Time Point 2? Would it be more appropriate to use the first efficacy evaluation time point for this analysis?

Time point 2 was set at 12 weeks post-treatment because the monitoring agency required the first disease reassessment between 12 and 16 weeks after the start of treatment (following the fourth cycle of therapy). Accordingly, time point 2 set at 12 weeks post-treatment corresponds to the first efficacy evaluation suggested by the Reviewer. In this regard, we fully agree that it is the most appropriate timing to retest thyroid function for the purposes of this analysis. We had already described this schedule roughly in the subsection “2.2 Data Collection and Assessment of Outcomes”. The same paragraph has been modified to make the schedule of disease re-evaluations clearer. Among the progressing patients, six (3.6%) had already experienced disease progression after the second or third course of treatment and thus before the expected time point for restaging.

I am also interested in the authors' perspective on the insufficient correlation between TSH levels and prognosis.

We were not surprised by the lack of a significant correlation between baseline TSH levels or its changes and survival outcomes. Two main reasons may account for this finding. From a general point of view, using the fT3/fT4 ratio instead of the value of the two individual hormones or TSH could be a better marker of peripheral deiodination activity and may even help stratify patients with euthyroid sick syndrome. In this condition, TSH is not usually used because its level usually remains within the normal range for several reasons (pituitary dysfunction, lower hypothalamic TRH production, and reduced TSH pulsatility) and is thus less reliable [Fragidis S, et al. Low T3 Syndrome and Long Term Mortality in Chronic Hemodialysis Patients. World J Nephrol (2015) 4(3):415–22. doi: 10.5527/wjn.v4.i3.415; Warner MH, Beckett GJ. Mechanism Behind the non-Thyroidal Illness Syndrome: An Update. J Endocrinol (2010) 205:1–13. doi: 10.1677/JOE-09-0412]. In addition, abnormal TSH level at baseline was an exclusion criterion for the purposes of the present study. Similarly, patients who developed clinically relevant thyroid dysfunction, defined by abnormal TSH levels at time point 2 as a result of immune checkpoint blockade, were excluded from the survival analysis. We can reasonably assume that such selection criteria may have contributed to the attenuation of longitudinal variability in TSH levels and the possible association with survival outcomes. We also added a specific comment on this in the “Discussion” section.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

No further comments.