Association Between Single-Nucleotide Polymorphism and Trastuzumab Deruxtecan-Induced Interstitial Lung Disease in Breast Cancer Using the Japonica Array NEO

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Dear Author,

Thank you for providing this interesting manuscript on clinical and genetic factors associated with trastuzumab deruxtecan (T‑DXd)–related interstitial lung disease (ILD). Although the topic is clinically important, especially in Japanese patients, there are several comments I would like to raise:

1. Sample size and statistical power
   The sample size (n=54; 15 ILD, 39 non‑ILD) is small for a genome‑wide association study and is a major limitation of this work. This is likely underpowered to detect modest genetic effects. I suggest clearly this study as an exploratory pilot analysis and adding a brief discussion on the need for larger, multi‑center cohorts to enable adequately powered genetic studies.

2. Very late ILD onset and causality
   ILD occurred 48~1,187 days after T‑DXd initiation (median 215 days, ≈7.1 months), with the latest case at 1,187 days (≈3.3 years), which is later and more variable than most published data. Please clarify how you judged very late ILD cases (especially >12 months) to be T‑DXd‑related, whether patients were still receiving T‑DXd at ILD onset, how alternative causes (e.g., infection, tumor progression, other drugs) were excluded, and why these events were included without sensitivity analyses.

3. Contradictory rs12625311 finding
   The candidate SNP analysis shows rs12625311 in TSHZ2 associated with ILD (p = 1.59 × 10⁻²), but with the opposite risk‑allele direction compared with the previous drug‑induced ILD report by Udagawa et al. I recommend (a) briefly summarizing the Udagawa study (drugs, population, reported effect of rs12625311) and (b) discussing potential reasons for the opposite direction (different drugs or mechanisms, study design, or context‑dependent TSHZ2 biology in T‑DXd–related ILD).

4. Null clinical associations and limited power
   None of the baseline clinical factors were significantly associated with ILD in your univariate analyses, despite prior reports of clinical risk factors for T‑DXd‑related or drug‑induced ILD. This may reflect limited power and the homogeneity of the cohort rather than a true lack of association. It would be helpful to explicitly discuss this issue.

5. Insufficient detail on ILD diagnosis
   In this study, ILD was diagnosed using CT findings and clinical course, with confirmation by pulmonologists or radiology reports and multidisciplinary discussion, but the specific criteria and process are not fully described. Please specify the diagnostic criteria (radiologic patterns, any pulmonary function test data, use of standardized ILD/causality tools) and clarify whether a uniform adjudication protocol was applied across all cases.

Author Response

Review1-1 Sample size and statistical power
The sample size (n=54; 15 ILD, 39 non‑ILD) is small for a genome‑wide association study and is a major limitation of this work. This is likely underpowered to detect modest genetic effects. I suggest clearly this study as an exploratory pilot analysis and adding a brief discussion on the need for larger, multi‑center cohorts to enable adequately powered genetic studies.

Response

We thank the reviewer for this valuable suggestion. We have revised the Discussion to clarify that this study should be considered an exploratory pilot analysis and to emphasize the need for larger, multi-center cohorts to enable adequately powered genetic studies.

 

 

Review1-2 Very late ILD onset and causality
ILD occurred 48~1,187 days after T‑DXd initiation (median 215 days, ≈7.1 months), with the latest case at 1,187 days (≈3.3 years), which is later and more variable than most published data. Please clarify how you judged very late ILD cases (especially >12 months) to be T‑DXd‑related, whether patients were still receiving T‑DXd at ILD onset, how alternative causes (e.g., infection, tumor progression, other drugs) were excluded, and why these events were included without sensitivity analyses.

Response
We thank the reviewer for raising this important point.
In our cohort, all ILD events occurred during ongoing T-DXd treatment, and no patients had received concomitant systemic anticancer therapy at the time of ILD onset. For each suspected case, alternative etiologies, including infection, tumor progression, or other drug-related lung injury, were carefully evaluated based on clinical course, radiologic findings, and multidisciplinary discussion involving oncologists, pulmonologists, and radiologists.
Because all ILD events occurred during active treatment and were assessed using the same diagnostic approach, we considered these cases appropriate for inclusion in the primary analysis, including those with later onset.
To improve clarity, we have revised the Methods section to explicitly state that the same diagnostic criteria and multidisciplinary evaluation were applied to all cases, including late-onset events. We have also revised the Results section to clarify that all ILD events occurred during ongoing T-DXd therapy and that no patients received concomitant systemic anticancer treatment at ILD onset.

Review1-3 Contradictory rs12625311 finding and potential biological mechanisms linking TSHZ2
The candidate SNP analysis shows rs12625311 in TSHZ2 associated with ILD (p = 1.59 × 10⁻²), but with the opposite risk‑allele direction compared with the previous drug‑induced ILD report by Udagawa et al. I recommend (a) briefly summarizing the Udagawa study (drugs, population, reported effect of rs12625311) and (b) discussing potential reasons for the opposite direction (different drugs or mechanisms, study design, or context‑dependent TSHZ2 biology in T‑DXd–related ILD).

Response
We thank the reviewer for this helpful suggestion.
We have added a brief summary of the previous drug‑induced ILD study, including the study population, implicated drugs, and the reported risk direction of the rs12625311.
We also expanded the Discussion to address the opposite direction observed in our cohort and discuss possible explanations, such as differences in drug mechanisms, study design, and context-dependent effects of TSHZ2 in T-DXd–related ILD.

 

Review1-4 Null clinical associations and limited power
None of the baseline clinical factors were significantly associated with ILD in your univariate analyses, despite prior reports of clinical risk factors for T‑DXd‑related or drug‑induced ILD. This may reflect limited power and the homogeneity of the cohort rather than a true lack of association. It would be helpful to explicitly discuss this issue.

Response
We thank the reviewer for this important comment. We have revised the Discussion to explicitly acknowledge that both the limited sample size and the relative clinical homogeneity of our cohort may have reduced our ability to detect significant associations with baseline clinical factors, despite prior reports suggesting such risk factors.

Review1-5 Insufficient detail on ILD diagnosis
In this study, ILD was diagnosed using CT findings and clinical course, with confirmation by pulmonologists or radiology reports and multidisciplinary discussion, but the specific criteria and process are not fully described. Please specify the diagnostic criteria (radiologic patterns, any pulmonary function test data, use of standardized ILD/causality tools) and clarify whether a uniform adjudication protocol was applied across all cases.

Response
We thank the reviewer for the helpful comment. We have revised the Methods section to clarify the radiologic criteria, diagnostic process, and exclusion of alternative causes. The multidisciplinary diagnostic approach is now more clearly described.

Reviewer 2 Report

Comments and Suggestions for Authors

The authors presented trastuzumab-deruxtecan ILD genetic possible mechanisms in Japanese cancer patients.

Please take into consideration my comments:

1. Rephrase Results Section, as there is no need to duplicate information, in tables and in text. Additionally, in this sections you do not discuss, just describe. So, for example, Udagawa et al. [16] study should be moved to Discussion or Introduction section, wherever the authors decide. Please organize better Results Section in this manner.
2. Because the research didn’t succeed in proving any relationship between Tdx-ILP and Japanese breast cancer patients, I consider the title misleading. The actual title suggests that using Japonica Array NEO you may identify ILD in breast cancer patients treated with Tdx, which is wrong.
3. Please emphasize better why this research is practice-changing for medical audience.

Author Response

Review2-1 Revise results section
Rephrase Results Section, as there is no need to duplicate information, in tables and in text. Additionally, in this sections you do not discuss, just describe. So, for example, Udagawa et al. [16] study should be moved to Discussion or Introduction section, wherever the authors decide. Please organize better Results Section in this manner.
Response
Thank you for this helpful suggestion. We have extensively revised the Results section to eliminate redundancy between the text and tables. Specifically, we removed detailed procedural descriptions (which were moved to the Methods section) and comparative interpretations of previous studies (which were moved to the discussion section). The revised section now focuses strictly on describing the analysis and the primary findings.

Review2 -2 Title
Because the research didn’t succeed in proving any relationship between T-DXd-ILP and Japanese breast cancer patients, I consider the title misleading. The actual title suggests that using Japonica Array NEO you may identify ILD in breast cancer patients treated with T-DXd, which is wrong.
Response
We thank the reviewer for this important comment.
We agree that the original title may have implied diagnostic identification rather than association analysis.
We have revised the title to more accurately reflect the exploratory and association-based nature of the study.

Review2 -3 emphasize better why this research is practice-changing for medical audience.
Response
We thank the reviewer for this important comment.
We have revised the Discussion to more clearly highlight the potential clinical implications of our findings.
Specifically, we emphasize that identification of host genetic factors associated with T-DXd–related ILD may contribute to future risk stratification, patient monitoring strategies, and individualized treatment decision-making.
We also clarify that our findings are hypothesis-generating and require validation before clinical implementation.

 

Reviewer 3 Report

Comments and Suggestions for Authors

This study investigated clinical and genetic factors associated with trastuzumab deruxtecan (T-DXd)–induced interstitial lung disease (ILD) in Japanese patients with breast cancer and found that baseline clinical characteristics alone could not predict ILD risk. Genome-wide analysis did not identify significant genetic variants, but a candidate SNP analysis suggested that host genetic factors may contribute to individual susceptibility. Overall, the findings indicate that T-DXd–induced ILD is multifactorial and that integrating genetic and clinical information may improve future risk stratification. However, while the topic is important and timely in the field of Cancer Research, several sections of the manuscript require further revision and clarification before it can be considered for publication in Cancers.

1. Since all participants were Japanese patients treated at a single institution, the generalizability of the findings to non-Japanese populations or broader clinical settings remains uncertain. The authors should further discuss how ethnic background and institutional factors may influence the applicability of the results.
2. Although ILD diagnosis was based on CT findings and multidisciplinary evaluation, additional details regarding the diagnostic criteria, interobserver agreement, and the exclusion of alternative causes of lung injury would improve clarity and reproducibility.
3. Blood samples were collected either before or during T-DXd treatment. The authors should clarify whether differences in sampling timing may introduce bias, particularly if treatment-related effects influenced sample quality or patient selection.
4. The functional relevance of the identified SNP within the TSHZ2 gene remains unclear. Additional discussion regarding potential biological mechanisms linking TSHZ2 to lung injury, immune responses, or drug-induced toxicity would strengthen the manuscript.
5. The identified SNP demonstrated an association in the opposite direction compared with previous reports. The authors should further elaborate on possible explanations, including differences in linkage disequilibrium structure, population characteristics, or treatment-specific mechanisms.
6. Given previously reported pharmacokinetic differences in T-DXd exposure between populations, the authors may consider discussing whether variations in drug exposure levels or dose intensity could have influenced ILD development in this cohort.
7. The current similarity index (percent match) is approximately 35%, which is relatively high. The manuscript should be revised to reduce textual similarity to below 20% by improving paraphrasing and restructuring overlapping sections where necessary.

Author Response

Review3-1

Since all participants were Japanese patients treated at a single institution, the generalizability of the findings to non-Japanese populations or broader clinical settings remains uncertain. The authors should further discuss how ethnic background and institutional factors may influence the applicability of the results
Response
We thank the reviewer for this important comment. We agree that the single-institution design and the inclusion of only Japanese patients may limit the generalizability of our findings to other populations and clinical settings. We have revised the Discussion section to explicitly acknowledge this limitation. We also emphasize that validation in larger multicenter cohorts will be necessary to confirm the broader applicability of these findings.

Review3-2

Although ILD diagnosis was based on CT findings and multidisciplinary evaluation, additional details regarding the diagnostic criteria, interobserver agreement, and the exclusion of alternative causes of lung injury would improve clarity and reproducibility.
Response
We thank the reviewer for the helpful comment. We have revised the Methods section to clarify the radiologic criteria, diagnostic process, and exclusion of alternative causes. The multidisciplinary diagnostic approach is now more clearly described.

Review3-3
Blood samples were collected either before or during T-DXd treatment. The authors should clarify whether differences in sampling timing may introduce bias, particularly if treatment-related effects influenced sample quality or patient selection
Response
We thank the reviewer for this important comment. Genomic DNA was extracted from plasma-depleted whole blood, and germline variants were analyzed to investigate host genetic factors. Germline variants are generally stable and not expected to be influenced by treatment exposure. Therefore, differences in sampling timing were unlikely to affect the genetic analyses. We have clarified this point in the Methods section.

Review3-4

The functional relevance of the identified SNP within the TSHZ2 gene remains unclear. Additional discussion regarding potential biological mechanisms linking TSHZ2 to lung injury, immune responses, or drug-induced toxicity would strengthen the manuscript.
Response
We thank the reviewer for this valuable suggestion.
We have revised and clarified the Discussion to more explicitly connect the previously described biological features of TSHZ2 with the context of drug-induced ILD.
Specifically, we reorganized the section to emphasize its role as a transcriptional regulator, its reported expression in immune cell subsets, and the potential regulatory impact of intronic variants on gene expression.
We also added a brief statement linking immune dysregulation to drug-induced ILD, while carefully avoiding overinterpretation given that no functional studies of this specific variant have been reported to date.

Review3-5

The identified SNP demonstrated an association in the opposite direction compared with previous reports. The authors should further elaborate on possible explanations, including differences in linkage disequilibrium structure, population characteristics, or treatment-specific mechanisms.
Response
We thank the reviewer for this important comment and suggestion. We have expanded the Discussion to address the opposite direction of association compared with previous reports.
We now discuss potential explanations, including differences in drug mechanisms, disease context, and treatment-specific effects.
Although the ethnic background was similar, variations in clinical characteristics, study design, or linkage disequilibrium structure may have contributed to the discrepant findings.

Review3-6
Given previously reported pharmacokinetic differences in T-DXd exposure between populations, the authors may consider discussing whether variations in drug exposure levels or dose intensity could have influenced ILD development in this cohort.
Response
We thank the reviewer for this insightful suggestion. We have clarified in the Discussion that we evaluated the association between weight-based dose intensity and ILD development in our cohort and found no significant association. We also added a brief statement acknowledging that direct pharmacokinetic measurements were not available in this study; therefore, the potential impact of interindividual differences in systemic drug exposure cannot be fully excluded.

Review3-7

The current similarity index (percent match) is approximately 35%, which is relatively high. The manuscript should be revised to reduce textual similarity to below 20% by improving paraphrasing and restructuring overlapping sections where necessary.
Response
Thank you for pointing out the similarity index. We take this issue seriously and have conducted a thorough revision of the entire manuscript. We have extensively rephrased and restructured the manuscript to enhance originality. Following these revisions, a check with the plagiarism detection tool Scribbr confirmed that the similarity index has been successfully reduced to 8%, which is well within the acceptable range.

Reviewer 4 Report

Comments and Suggestions for Authors

Dear Editor,

The authors conducted an important study evaluating clinical and genetic factors associated with ILD development in patients treated with T-DXd. They retrospectively analyzed 54 patients who had received T-DXd. ILD occurred in 15 patients (27.8%), with a median onset time of 215 days. None of the baseline clinical or treatment-related factors were found to be significantly associated with ILD development. Genome-wide analyses did not identify significant SNPs after correction for multiple testing. However, a targeted analysis focusing on SNPs previously linked to drug-induced ILD revealed one variant that was significantly associated with ILD in this cohort. The study emphasizes that integrating clinical assessment with genetic information may help further refine risk stratification for T-DXd–induced ILD in the future. It presents an original idea and is, overall, well written.

The introduction section is quite lengthy. It should be shortened to avoid distracting the reader.

The mortality rate associated with ILD development in patients treated with T-DXd should be specified.

The objective paragraph in the introduction section should be written more concisely and simply.

It should also be noted that the study of limitations is retrospective.

Sincerely

Comments on the Quality of English Language

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

Author Response

Review4-1and Review4-3

The introduction section is quite lengthy. It should be shortened to avoid distracting the reader.
The objective paragraph in the introduction section should be written more concisely and simply.
Response
We appreciate the reviewer’s constructive comments. In response, we have substantially revised and shortened the Introduction section to improve focus and readability. Redundant background information has been removed, and the narrative has been streamlined to emphasize the clinical significance of T-DXd–induced interstitial lung disease (ILD), the observed ethnic differences in incidence, and the remaining knowledge gap regarding potential host susceptibility factors.
In addition, the objective paragraph has been rewritten in a clearer and more concise manner to directly state the purpose of the study without unnecessary contextual detail. We believe these revisions enhance clarity and ensure that the Introduction more effectively leads to the research question.

Review4-2
The mortality rate associated with ILD development in patients treated with T-DXd should be specified.
Response
We appreciate the reviewer’s comment. Although we had stated that no Grade ≥3 events were observed, we have now explicitly added a statement clarifying that no fatal (Grade 5) ILD events occurred during the study period.

Review4-4
It should also be noted that the study of limitations is retrospective.
Response
We thank the reviewer for this important comment. We have revised the Discussion to explicitly acknowledge the retrospective design as a limitation and to describe its potential impact on clinical assessment and selection bias.

 

Review4-5

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

Response

We appreciate the reviewer’s comment. We have used the Editage English editing service, and we have attached the certification file.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

The authors improved the manuscript and answered all my comments.

Reviewer 4 Report

Comments and Suggestions for Authors

Dear Editor,

The manuscript is acceptable.

Sincerely