Review Reports

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Very interesting data, long-term study. Worth publishing.

Is prevalence in Turkey similar to reported 6-8/1000000/year? Only 113 patients treated over the period of 13 years.

What was the protocol for screening for metastatic disease? (2 patients died before receiving systemic therapy)

Author Response

- Comment 1) Is prevalence in Turkey similar to reported 6-8/1000000/year? Only 113 patients treated over the period of 13 years.

-Response 1) Thank you for this important and valuable comment.
The true incidence or prevalence of uveal melanoma in Turkey is currently unknown, as there is no national population-based cancer registry that reports uveal melanoma as a distinct entity. The most recent GLOBOCAN 2023 Türkiye data do not provide uveal melanoma specific incidence figures.

Our cohort consists of patients treated in volunteer medical oncology centers, and therefore does not represent a population-based sample. In Turkey, a considerable proportion of uveal melanoma patients, particularly those without metastatic disease, may be followed exclusively in ophthalmology clinics and may never be referred to medical oncology departments. Consequently, the number of patients included in our study should not be interpreted as reflecting national incidence or prevalence rates.

Furthermore, due to the retrospective nature of the study, some eligible patients may not have been captured because of incomplete records or missing follow-up data. Notably, the primary aim of this study was not to estimate incidence or prevalence, but rather to evaluate follow-up patterns, time to metastatic progression, and outcomes of local and systemic treatments in patients who were referred to oncology clinics.

We have clarified this point in the revised manuscript (Discussion / in line 408-410).

-Comment 2) What was the protocol for screening for metastatic disease? (2 patients died before receiving systemic therapy)

- Response 2) We thank the reviewer for this comment. Follow-up and surveillance of patients managed in the localized disease setting were conducted according to the routine clinical practice and institutional protocols of each participating center, reflecting real-world management in this retrospective multicenter study. This approach has been clarified in the Methods section (lines 146–148) and acknowledged as a limitation in the Limitations section (lines 406–416).

Reviewer 2 Report

Comments and Suggestions for Authors

 

Manuscript: Real-World Treatment Patterns and Survival in Uveal Melanoma: A Multicenter Cohort Study by the Turkish Oncology Group (TOG)

Manuscript ID: cancers-4051283

This manuscript presents a large, multicenter real-world cohort study evaluating treatment patterns and survival outcomes in uveal melanoma across Türkiye. Given the rarity of uveal melanoma and the limited availability of comprehensive real-world datasets—particularly from non-Western populations—this work represents a valuable and timely contribution to the literature. The study is well-organized, methodologically sound for a retrospective design, and clearly written. The analyses are generally appropriate, and the conclusions are well supported by the presented data.

Major Comments

1. Selection Bias and Treatment Allocation
   As an observational retrospective study, treatment selection (e.g., liver-directed therapy, dual immune checkpoint inhibition) was likely influenced by patient fitness, and disease burden. Although this is partially acknowledged, the Discussion would benefit from a more explicit clarification that the observed survival benefit associated with liver-directed therapy and dual ICI may reflect selection bias rather than causal inference. A short explanatory sentence would strengthen scientific rigor.
2. Heterogeneity of Liver-Directed Therapies
   Liver-directed therapies are grouped together in survival analyses, yet these modalities (TARE, RFA, resection, HAI) differ substantially in indication and efficacy. The authors should clarify this limitation more explicitly and caution against overgeneralization of the OS benefit attributed to liver-directed therapy as a single category.

Minor Comments

1. Language and Typographical Issues

   • Minor typographical errors should be corrected (e.g., “Unkonwn” → “Unknown” in Table 1; “Unilober/Bilober” → “Unilobar/Bilobar” in Table 2).

   • Consistent formatting of percentages (e.g., spacing before “%”) should be ensured throughout tables.

2. Clarification of Statistical Reporting

   • In Table 3, the presentation of p-values (“p=0.19 / 0.05”) could be clarified in a footnote to specify which value corresponds to ORR and which to DCR.

   • Ensure consistency between text and figure legends when reporting median PFS values.

3. Figures and Legends

   • Figures are clear and appropriate; however, legends could be slightly expanded to explicitly state the number of patients at risk where applicable

Comments on the Quality of English Language

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

Author Response

Major Comments

Comment 1: Selection Bias and Treatment Allocation
As an observational retrospective study, treatment selection (e.g., liver-directed therapy, dual immune checkpoint inhibition) was likely influenced by patient fitness, and disease burden. Although this is partially acknowledged, the Discussion would benefit from a more explicit clarification that the observed survival benefit associated with liver-directed therapy and dual ICI may reflect selection bias rather than causal inference. A short explanatory sentence would strengthen scientific rigor.

Response 1: Thank you for this meaningful and insightful comment. We fully agree that, given the retrospective and observational design of the study, treatment allocation was not randomized and may have been influenced by clinical factors such as performance status and disease burden. To address this concern and strengthen the scientific rigor of the manuscript, we have revised the Discussion section to explicitly acknowledge the potential for selection bias and clarify that the observed survival differences associated with liver-directed therapy should not be interpreted as evidence of a causal relationship (Discussion in line 344-352 and line 384-386).

Comment 2: Heterogeneity of Liver-Directed Therapies
Liver-directed therapies are grouped together in survival analyses, yet these modalities (TARE, RFA, resection, HAI) differ substantially in indication and efficacy. The authors should clarify this limitation more explicitly and caution against overgeneralization of the OS benefit attributed to liver-directed therapy as a single category.

Response 2: 

We thank the reviewer for this important observation. We agree that liver-directed therapies represent a heterogeneous group of interventions with differing indications, patient selection criteria, and efficacy profiles. In our study, these modalities were analyzed as a single category primarily due to limited sample sizes within individual treatment subgroups, which precluded meaningful modality-specific survival comparisons.

To address this limitation, we have revised the Discussion section (in line 348-352) to acknowledge the heterogeneity of liver-directed therapies explicitly and to caution against overgeneralizing the observed overall survival benefit as a uniform class effect. We now emphasize that the observed survival differences should be interpreted as reflecting selected patients receiving any liver-directed approach rather than as attributable to a specific modality.

Minor Comments

Comment 1: Language and Typographical Issues

Minor typographical errors should be corrected (e.g., “Unkonwn” → “Unknown” in Table 1; “Unilober/Bilober” → “Unilobar/Bilobar” in Table 2).

Consistent formatting of percentages (e.g., spacing before “%”) should be ensured throughout tables

Response 1: We thank the reviewer for this careful and insightful comment. We regret that these minor issues were overlooked in the initial version, and the necessary corrections have now been implemented.

Comments 2: Clarification of Statistical Reporting

In Table 3, the presentation of p-values (“p=0.19 / 0.05”) could be clarified in a footnote to specify which value corresponds to ORR and which to DCR.

Ensure consistency between text and figure legends when reporting median PFS values.

Response 2: We thank the reviewer for this careful and helpful comment. We have revised the footnote of Table 3 to clearly indicate which p-value corresponds to ORR and which corresponds to DCR. We also reviewed the manuscript to ensure consistency between the text and figure legends regarding the reported median PFS values.

Comment 3: Figures and Legends

Figures are clear and appropriate; however, legends could be slightly expanded to explicitly state the number of patients at risk where applicable

Response 3: We thank the reviewer for this helpful comment. We have expanded the figure legends to explicitly indicate that the number of patients at risk is shown below the corresponding survival curves.

Reviewer 3 Report

Comments and Suggestions for Authors

This work investigated the current clinical treatment status of uveal melanoma. However, from the perspectives of pharmaceutical sciences and clinical therapeutics, several important issues limit the interpretability and translational value of the study. Major concerns are as follows:

1. The manuscript reports clinical outcomes but does not adequately integrate the pharmacological mechanisms of the investigated therapies with observed clinical responses. A clearer discussion linking drug mechanism of action to treatment efficacy would strengthen the translational interpretation.
2. Critical pharmacotherapy details, including dosing regimens, treatment duration, dose modifications, and reasons for treatment discontinuation, are not sufficiently described. 
3. The analysis focuses primarily on efficacy endpoints, with limited reporting of adverse events or treatment-related toxicity. Safety is also an important indicator of drugability.
4. The manuscript reports multiple lines of therapy but does not sufficiently evaluate how treatment sequencing influences clinical outcomes. A more detailed analysis of sequential therapy would be valuable for guiding clinical decision-making.

Author Response

Comment 1: The manuscript reports clinical outcomes but does not adequately integrate the pharmacological mechanisms of the investigated therapies with observed clinical responses. A clearer discussion linking the drug mechanism of action to treatment efficacy would strengthen the translational interpretation.

Response 1: We thank the reviewer for this important and constructive comment. In response, we have expanded the Discussion section to integrate better the pharmacological mechanisms of action of the investigated therapies with the observed clinical outcomes. Specifically, we now discuss the molecular and immunobiological features of uveal melanoma and explain how these characteristics mechanistically relate to the efficacy of immune checkpoint inhibitors, dual immune checkpoint blockade, and tebentafusp, thereby strengthening the translational interpretation of our findings. We believe these additions improve the mechanistic and translational relevance of the manuscript.

Comment 2: Critical pharmacotherapy details, including dosing regimens, treatment duration, dose modifications, and reasons for treatment discontinuation, are not sufficiently described. 

Response 2: We thank the reviewer for this important comment. Detailed pharmacotherapy data, including dosing regimens, treatment duration, dose modifications, and reasons for treatment discontinuation, were not consistently available across participating centers due to the retrospective and multicenter nature of the study. In routine clinical practice, particularly over a long study period and across multiple institutions, such granular treatment details are not uniformly documented in oncology records, especially when patients receive care at different centers. In addition, standardized and comprehensive documentation of performance status and detailed toxicity profiles is variable in routine practice, which limited the availability and reliability of these data for inclusion in our analyses

We therefore deliberately avoided reporting incomplete or potentially misleading pharmacotherapy details. We have now explicitly acknowledged this limitation in the revised Discussion section (in line 412-414).

Comment 3: The analysis focuses primarily on efficacy endpoints, with limited reporting of adverse events or treatment-related toxicity. Safety is also an important indicator of drugability.

Response 3: We agree with the reviewer that safety and treatment-related toxicity are important components of therapeutic evaluation. However, due to the retrospective and multicenter design of the study, adverse event data were not consistently or reliably documented across participating centers. To preserve the scientific validity and interpretability of our analyses, we chose not to include incompletely documented safety data. This limitation has now been explicitly acknowledged in the revised manuscript (in line 408-418).

Comment 4: The manuscript reports multiple lines of therapy but does not sufficiently evaluate how treatment sequencing influences clinical outcomes. A more detailed analysis of sequential therapy would be valuable for guiding clinical decision-making.

Response 4: We thank the reviewer for this critical and constructive comment.  We have revised the manuscript to more explicitly address the impact of treatment sequencing on clinical outcomes. Specifically, we have provided a detailed description of the sequencing of liver-directed therapies, including their use as first-line locoregional treatment and the frequency of repeated applications (Results section in line 216-222).

In addition, we evaluated the clinical relevance of access to immunotherapy in the first-line setting and its association with overall survival. Accordingly, Figure 3 has been revised to illustrate overall survival according to first-line systemic treatment (chemotherapy versus immunotherapy). This finding has been addressed and discussed in the revised Discussion section (in line 369-374). These additions aim to better characterize treatment sequencing patterns and their potential influence on survival outcomes, thereby enhancing the clinical interpretability and decision-making relevance of the study.

Reviewer 4 Report

Comments and Suggestions for Authors

This is a well-conducted multicenter real-world study providing valuable national data on treatment patterns and outcomes in uveal melanoma, a rare malignancy with limited real-world evidence, particularly from non-Western populations. The manuscript is generally well written, and the results are clinically relevant. There are several minor comments that may help improve clarity and methodological transparency.

1. 
The multivariate Cox model includes limited covariates and may be underpowered given the number of metastatic events. Important confounders such as performance status, disease-free interval, and metastatic burden are missing. The exploratory nature of the multivariate results should be emphasized.

2. 
Additional clarification regarding patient selection for liver-directed therapies (e.g., TARE, RFA, surgery) would strengthen interpretation of the observed survival benefit. Please specify whether treatment decisions were influenced by factors such as performance status, hepatic tumor burden, liver function, or institutional practice.

3. 
Grouping all immunotherapies together introduces heterogeneity, especially given the different efficacy of single-agent ICI, dual ICI, and tebentafusp. This limitation should be explicitly acknowledged.

4. 
In Section 3.2, the sentence stating that “Any statistically significant differences were observed between patients with de novo and recurrent metastatic disease” appears to be a wording inconsistency and should be revised. Minor standardization of p-value notation and clearer presentation of p-values for ORR and DCR in Table 3 are also recommended.

Author Response

Comment 1: The multivariate Cox model includes limited covariates and may be underpowered given the number of metastatic events. Important confounders such as performance status, disease-free interval, and metastatic burden are missing. The exploratory nature of the multivariate results should be emphasized.

Response 1:  We thank the reviewer for this insightful comment and fully agree with the concerns raised. Given the limited number of metastatic events, the multivariate Cox regression model may indeed be underpowered. As a surrogate of disease burden, multiple liver metastases were included in the multivariate analysis. However, due to the retrospective and multicenter nature of the study, performance status was not consistently reported across participating centers. It therefore could not be reliably incorporated into the multivariate model. Disease-free survival was not included in the multivariate model due to the limited number of DFS events, which would have resulted in an unstable and underpowered analysis. Priority was therefore given to variables with stronger clinical relevance and sufficient event counts to allow reliable modeling.

The limited statistical power of the multivariate analysis and its exploratory nature have now been explicitly acknowledged in the Limitations section of the manuscript (lines 408–418).

Comment 2: Additional clarification regarding patient selection for liver-directed therapies (e.g., TARE, RFA, surgery) would strengthen interpretation of the observed survival benefit. Please specify whether treatment decisions were influenced by factors such as performance status, hepatic tumor burden, liver function, or institutional practice.

Response 2: We thank the reviewer for this important comment. Selection of liver-directed therapies in our study was not randomized and reflected routine multidisciplinary clinical decision-making. Treatment allocation was influenced by factors such as hepatic tumor burden, liver function, patient fitness, prior therapies, and institutional experience and availability.

We have now clarified this point in the revised manuscript and explicitly acknowledged that patients selected for liver-directed approaches may represent a clinically more favorable subgroup. Accordingly, we emphasize that the observed survival benefit should be interpreted cautiously and not as evidence of a causal treatment effect (Discussion in line 341-351).

Comment 3: Grouping all immunotherapies together introduces heterogeneity, especially given the different efficacy of single-agent ICI, dual ICI, and tebentafusp. This limitation should be explicitly acknowledged.

Response 3: We thank the reviewer for this important observation. We agree that grouping all immunotherapies together introduces heterogeneity due to the differing efficacy profiles of single-agent immune checkpoint inhibitors, dual immune checkpoint inhibition, and tebentafusp. This limitation has now been explicitly acknowledged in the revised manuscript (Discussion in line 372-375), and we have cautioned against overinterpretation of immunotherapy as a uniform treatment category.

Comment 4: In Section 3.2, the sentence stating that “Any statistically significant differences were observed between patients with de novo and recurrent metastatic disease” appears to be a wording inconsistency and should be revised. Minor standardization of p-value notation and clearer presentation of p-values for ORR and DCR in Table 3 are also recommended.

Response 4: We thank the reviewer for pointing out this wording inconsistency. The sentence in Section 3.2 has been revised for clarity. In addition, the explanatory footnote of Table 3 has been expanded to more clearly specify the presentation of p-values for ORR and DCR, thereby improving interpretability for readers.

Reviewer 5 Report

Comments and Suggestions for Authors

This manuscript presents a large, multicenter cohort analysis of uveal melanoma patients treated across multiple tertiary oncology centers in Turkey. The topic is timely and clinically relevant. However, several revisions and  clarifications are required.

Detailed comments:

Adding continuous line numbering would improve the review process. Besides, it is mandatory in Cancers.

The iThenitcate percent match (38%) is large. It must be decreased below the generally acceptable 30% level.

Please clarify whether all 19 centers contributed patients equally, or whether there was substantial variability in patient numbers per center. This information is relevant for assessing potential center-specific bias.

When reporting higher ORR and DCR for immunotherapy versus chemotherapy, please explicitly state that most comparisons did not reach statistical significance due to small sample sizes, to avoid overinterpretation.

Please clarify whether “clinically confirmed” uveal melanoma included imaging-only diagnoses or ophthalmologic assessment without histology, and how this may have affected staging accuracy.

It would be helpful to specify which variables were considered “clinically relevant” and thus included in the multivariate Cox model.

The sentence “Any statistically significant differences were observed…” should be rewritten to “No statistically significant differences were observed…”.

Please clarify whether treatment choice was physician-driven, center-specific, or influenced by national reimbursement policies.

 

Author Response

Comment 1:

Adding continuous line numbering would improve the review process. Besides, it is mandatory in Cancers.

Response:
Continuous line numbering has now been added throughout the revised manuscript in accordance with the journal requirements.

Comment 2:

The iThenticate percent match (38%) is large and should be reduced below 30%.

Response:
We thank the reviewer for this important comment. Due to the lack of direct access to iThenticate at our institution, the revised manuscript was independently re-evaluated using the intihal.net plagiarism detection system, which is widely used by Turkish academic institutions. Following careful rephrasing and reduction of overlapping text, the similarity index was found to be 20%, which is below the acceptable threshold for the journal. We therefore believe that the current version of the manuscript adequately addresses concerns regarding textual similarity.

Comment 3:

Please clarify whether all 19 centers contributed patients equally, or whether there was substantial variability in patient numbers per center.

Response:

Patient contribution was not uniform across centers. Only centers contributing at least two patients were included, with the number of patients per center ranging from 2 to 13. This clarification has been added to the Methods section (in line 152-153), and potential center-related bias is now acknowledged in the Limitations.

Comment 4:

When reporting higher ORR and DCR for immunotherapy versus chemotherapy, please explicitly state that most comparisons did not reach statistical significance due to small sample sizes.

Response:
We agree with this comment. The Discussion section has been revised to explicitly state that although ORR and DCR were numerically higher with immunotherapy, these comparisons did not reach statistical significance, likely due to small sample sizes, and should therefore be interpreted with caution (lines 367–369).

Comment 5:

Please clarify whether “clinically confirmed” uveal melanoma included imaging-only diagnoses or ophthalmologic assessment without histology, and how this may have affected staging accuracy.

Response: 

We thank the reviewer for this comment. As clarified in the Methods section, patients with clinically confirmed uveal melanoma were included even in the absence of histologic biopsy, in line with routine clinical practice. However, we were unable to establish a direct relationship between biopsy findings and staging, as histopathologic features are not incorporated into the AJCC Cancer Staging Manual, Eighth Edition (2017; corrected 4th printing, 2018).

The incompleteness of AJCC T and N classifications in our cohort primarily reflects the use of data derived from patients referred to medical oncology clinics, for whom detailed ophthalmologic staging records were not consistently accessible. As a result, staging information generated within ophthalmology clinics could not be fully retrieved. We have acknowledged this limitation in the Discussion (in lines 323-325).

Comment 6:

It would be helpful to specify which variables were considered “clinically relevant” and thus included in the multivariate Cox model.

Response

We thank the reviewer for this helpful comment. Variables included in the multivariate Cox regression model were selected a priori based on established clinical relevance in metastatic uveal melanoma, as supported by the literature, and on data availability across centers. These variables included treatment-related factors (such as liver-directed therapy and combination immune checkpoint inhibition) and disease burden (specifically liver metastatic burden), all of which have been consistently associated with prognosis in prior studies; the relevant references have now been incorporated into the manuscript (metot secition in lines 172-174).

Due to the retrospective nature of the study, reliable data on performance status and treatment-related toxicities were not consistently available and therefore could not be included in the multivariate model. In addition, the limited number of events further constrained the number of covariates that could be robustly modeled. These issues have been acknowledged in the Discussion section as study limitations (in lines 408-418), and the exploratory nature of the multivariate analysis has been explicitly emphasized.

Comments 7: The sentence “Any statistically significant differences were observed…” should be rewritten.

Response:
The sentence has been revised (in line 211).

Comment 8:

Please clarify whether treatment choice was physician-driven, center-specific, or influenced by national reimbursement policies.

Response:
Treatment selection was influenced by a combination of physician judgment, center-specific expertise and availability, and national reimbursement policies. This has now been clearly stated in the revised Discussion (in the limitation paragraph line 408-418) section.

Round 2

Reviewer 3 Report

Comments and Suggestions for Authors

The authors have addressed all of my concerns about this work. I suggest accepting this revision in its present form.

Reviewer 5 Report

Comments and Suggestions for Authors

The authors have revised and improved their work, current version can be accepted.