Outcomes and Predictors of Recurrence and Survival in Surgically Resected Localized Chromophobe Renal Cell Carcinoma: Results from the Canadian Kidney Cancer Information System (CKCis)

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

To the authors,

General Comments

This study analyzes a large prospective national CKCis cohort of patients with surgically resected, localized chromophobe renal cell carcinoma (chRCC). Given the rarity of chRCC, this histology-specific analysis with a large sample size is a major strength. The statistical approach is generally appropriate, the results are consistent with prior studies, and the identification of high-risk pathological features has clear clinical relevance. Some points would benefit from clarification to improve interpretability.

 

Major Comments

1. Exclusion Criteria and Generalizability
   Patients with bilateral, multifocal, and pN1 disease were excluded from the RFS analysis. The authors should more clearly state that the recurrence estimates mainly apply to sporadic, unilateral, pN0 chRCC and may not be generalizable to node-positive disease.
2. Proportional Hazards Assumption
   Cox models were used, but assessment of the proportional hazards assumption is not described. The authors should indicate whether this was evaluated or acknowledge it as a limitation.
3. Multicollinearity Between Tumor Size and pT Stage
   Both pathologic tumor size and pT stage were included in the OS model, raising potential multicollinearity concerns. The authors should report whether this was assessed, perform a sensitivity analysis, or acknowledge this limitation.
4. Post-Recurrence Management
   Post-recurrence treatments are well described, but their association with survival is not analyzed. The authors should state that such analyses were beyond the scope of this study and warrant future investigation.

 

Minor Comments

1. Sarcomatoid Features
   The low prevalence of sarcomatoid differentiation (1.5%) should be briefly discussed, particularly in relation to the lack of centralized pathology review.
2. Long-Term Outcome Interpretation
   With a median follow-up of 4.9 years, 10-year RFS and OS estimates should be interpreted cautiously. Reporting numbers at risk or adding a brief caveat would improve clarity.
3. Editorial Issues
   Minor typographical and formatting issues should be corrected, including wording and table alignment.

Author Response

Thank you for taking the time to review our paper. We are glad you see the value in our results.  All of your concerns and comments are addressed below.

Major comments:

1. Patients with bilateral, multifocal, and pN1 disease were excluded from the RFS analysis. The authors should more clearly state that the recurrence estimates mainly apply to sporadic, unilateral, pN0 chRCC and may not be generalizable to node-positive disease.

• Thank you for pointing this out. Under results, section 3.2, we have added the following on page 6, line 185-186 of the results section to remind readers we are referring to the pN0, solitary chRCC patients in the RFS modelling. We have also highlighted this in the discussion on page 9, line 253-256 as we summarize the findings from our RFS model by clarifying: … in the RFS model for pN0 solitary chRCC patients…... We have also modified the last sentence in the conclusions of the paper (page 11, line 2324-326) and abstract (page 2 lines 60-61) to remove the specific high risk features so there is no chance of misleading the readers regarding the patient population analyzed: There is a subgroup of chRCC patients with less favourable outcomes who should be the focus of future research that aims to prevent recurrence and RCC death.

2. Proportional Hazards Assumption
Cox models were used, but assessment of the proportional hazards assumption is not described. The authors should indicate whether this was evaluated or acknowledge it as a limitation.

• Thank you for asking us this question.  We did consider this, but we did not assess the proportional hazard assumptions due to the low event rate and thus the potential to produce unreliable results.  We have included this in the discussion on page 10, line 311-312 in the section about limitations to our study: Other limitations relate to the low number of events in our study.  This resulted in the inability to perform an assessment of the proportional hazard assumption.

3. Multicollinearity Between Tumor Size and pT Stage
Both pathologic tumor size and pT stage were included in the OS model, raising potential multicollinearity concerns. The authors should report whether this was assessed, perform a sensitivity analysis, or acknowledge this limitation.

• Thank you for asking us to clarify this. It is reassuring that both pathological tumor size and pT stage are also significant variables in other clear cell RCC models (including references 13,18,20). At your suggestion, however, we did do sensitivity analyses (removing pT stage and subsequently removing pathological size) and the results were consistent.  We have included that information in the results section, in Section 3.4, page 8, line 213-215 by adding the sentence:  Given both pathological size and pT stage were significant, sensitivity analyses were performed removing each of these variables and the results are consistent.

4. Post-Recurrence Management
Post-recurrence treatments are well described, but their association with survival is not analyzed. The authors should state that such analyses were beyond the scope of this study and warrant future investigation.

• Thank you for pointing this out. We have added the following sentence in the discussion, under limitations on page 11 row 314-316: Also, studying the impact of post recurrence treatments and their association with survival was beyond the scope of this study but is the focus of an ongoing CKCis research.

 Minor Comments

1. Sarcomatoid Features
The low prevalence of sarcomatoid differentiation (1.5%) should be briefly discussed, particularly in relation to the lack of centralized pathology review.

• Thank you for this comment. On further review, our incidence of sarcomatoid is 1.6% not 1.5% and this has been corrected in Table 1, in the results section page 5 line 165, and in the discussion page 9, line 262. We did go back and do a further review of the literature and found the incidence varied between 1.8% and 9% so instead of including the middle value of 5-6%, we have changed the results section to indicate the range.  This change is made in the discussion, page 9, line 263-265: The presence of sarcomatoid features in chRCC varies in the literature from 1.8% to 9%. Thus, our data is not dissimilar to the lowest published rates. This resulted in deleting references 23 and 24 and changing them to the ones currently included in the reference section.  We have also modified in the discussion, on page 10, line 310-311 that the lack of central review may have also influenced other pathological findings including sarcomatoid features:  A major limitation includes the the lack of centralized pathology review to confirm chRCC as well as other pathological findings like the presence of sarcomatoid features.  

2. Long-Term Outcome Interpretation
With a median follow-up of 4.9 years, 10-year RFS and OS estimates should be interpreted cautiously. Reporting numbers at risk or adding a brief caveat would improve clarity.

• Thank you very much for pointing this out. We have extended the sentence in the discussion at the end of the paragraph discussing RFS and OS on page 9, line 240-241: …although our 10 year outcomes must be interpreted with the knowledge that our median follow-up was 5 years. To make it clearer, we have also reconfigured the RFS and OS curves to ensure the data is presented as clearly as possible and the graphs now show the 95% CI.

Editorial Issues
Minor typographical and formatting issues should be corrected, including wording and table alignment.

• Thank you for this comment. We have attempted to correct all typographical and formatting issues throughout the manuscript.

Reviewer 2 Report

Comments and Suggestions for Authors

Dear authors.

 This study provides a valuable analysis of oncologic outcomes in chromophobe renal cell carcinoma (chRCC), a relatively rare subtype often underrepresented in large datasets. The use of a multi-institutional cohort of 790 patients from the CKCis database is a significant strength, offering robust real-world evidence regarding the prognosis of localized chRCC. The identification of key predictors such as pathologic T stage, sarcomatoid features, and positive surgical margins provides clinically meaningful insights for risk stratification and follow-up planning. Overall, the study is well-designed and addresses an important knowledge gap.

 To enhance clarity and clinical interpretability, I suggest a few revisions. Since pathologic T stage emerged as the strongest predictor of recurrence-free survival (RFS) and overall survival (OS), it would be helpful for the authors to illustrate stage-specific risk more clearly. This could be achieved by providing Kaplan–Meier curves stratified by pathologic stage or by adding a concise summary of recurrence and mortality events by stage. Including either option would significantly improve clinical interpretability.

 Additionally, there are minor inconsistencies in reported values across sections. For example, tumor necrosis is reported as 20.1% in the Results text but 19.6% in Table 1, and positive margins are listed as 7.7% in the text versus 7.3% in Table 1. These values should be verified and harmonized throughout the manuscript. It would also be helpful to clarify why the Charlson Comorbidity Index denominator (n=742) is lower than the full cohort size (n=790), possibly by noting missing data.

 Overall, this is a well-conducted and clinically relevant study, and I believe it will be suitable for publication after these minor clarifications are addressed.

 

Sincerely. 

Author Response

Thank you for taking the time to review our research.  We have addressed all of your concerns below and have paid special attention to your concerns regarding presenting the results more clearly. 

1. Since pathologic T stage emerged as the strongest predictor of recurrence-free survival (RFS) and overall survival (OS), it would be helpful for the authors to illustrate stage-specific risk more clearly. This could be achieved by providing Kaplan–Meier curves stratified by pathologic stage or by adding a concise summary of recurrence and mortality events by stage. Including either option would significantly improve clinical interpretability.

• Thank you for requesting this additional information. We have included these results by adding in Kaplan – Meier curves for RFS and OS stratified by pT stage as new figures (Figure 2B for RFS and Figure 3B for OS). We felt that visualizing the results was the easiest and clearest way to review the data.  

2. There are minor inconsistencies in reported values across sections. For example, tumor necrosis is reported as 20.1% in the Results text but 19.6% in Table 1, and positive margins are listed as 7.7% in the text versus 7.3% in Table 1. These values should be verified and harmonized throughout the manuscript. It would also be helpful to clarify why the Charlson Comorbidity Index denominator (n=742) is lower than the full cohort size (n=790), possibly by noting missing data.

• Thank you for pointing out these discrepancies. Table 1 has been updated to include the correct percentages which consider missing data.  And you are correct, the CCI denominator of n=742 (instead of n=790) is due to missing data. We have made this clearer with the addition of this information at the bottom of the table for CCI along with all the other variables. (page 5, row 156-158)

 

Reviewer 3 Report

Comments and Suggestions for Authors

 

This manuscript presents outcomes of 790 patients with surgically resected localized chromophobe renal cell carcinoma (chRCC) from the Canadian Kidney Cancer Information System cohort. Given the rarity of chRCC and the frequent aggregation of non-clear cell subtypes in prior reports, this represents one of the largest contemporary series focusing specifically on localized chRCC. The study provides clinically meaningful long-term data, and the favorable recurrence-free and overall survival rates are consistent with prior institutional and registry-based findings.

Major

1. Regarding the results, although the cohort size is substantial, the number of recurrence events remains relatively small, with only 45 recurrences observed. Given this limited number of events, further clarification on model construction would strengthen the manuscript. Specifically, it would be helpful to describe the number of events included in the multivariable recurrence model and how many covariates were entered, in order to address potential concerns regarding overfitting. The wide confidence intervals for some predictors, particularly for higher pT stage, suggest that model stability may be limited.
2. The exclusion of pN1 patients from the recurrence model due to instability is understandable from a statistical perspective. However, given that six of eight node-positive patients recurred, this subgroup appears biologically distinct and clinically high-risk. Even if formal modeling is limited by small numbers, the manuscript would benefit from emphasizing that pN1 disease likely represents an extremely aggressive subset within chRCC and deserves explicit clinical attention.
3. Because only 20 of 53 deaths were attributed to RCC and overall survival rather than cancer-specific survival was modeled, the issue of competing risks should be addressed. In a population with generally favorable prognosis and a non-negligible proportion of non-RCC deaths, a competing risk framework could provide additional insight. If such an analysis is not feasible, this limitation should at least be acknowledged.
4. The recurrence pattern also warrants further elaboration. Since the majority of recurrences were metastatic, additional information regarding metastatic sites and whether recurrence timing differed by pathological stage would enhance biological interpretation and clinical applicability.
5. The discussion section would benefit from a more cautious tone regarding adjuvant therapy development. While the manuscript appropriately identifies a higher-risk subgroup, the overall prognosis of localized chRCC in this cohort is excellent, with 5- and 10-year recurrence-free survival rates of 94% and 90%, respectively. Given this low absolute recurrence risk, the clinical justification for aggressive development of adjuvant systemic therapy is debatable. The discussion currently leans toward suggesting that high-risk chRCC should be the focus of future adjuvant trials. However, the authors should more explicitly acknowledge the potential for overtreatment, the likely high number needed to treat, and the uncertain systemic therapy sensitivity of this histologic subtype. A more balanced interpretation would strengthen the scientific rigor of the manuscript.
6. Furthermore, the excellent outcomes reported in this study raise a broader and potentially more impactful question: whether all surgically managed localized chRCC requires equally intensive intervention and surveillance. Rather than focusing primarily on escalation strategies, the discussion could also consider whether de-escalation approaches, including tailored follow-up or even active surveillance in selected low-risk tumors, may be appropriate. Given the favorable natural history demonstrated here, this perspective would add important nuance.

< !-- notionvc: 4b5899e6-9642-43d1-8e02-611f88f8a21d -->

Author Response

Thank you for taking the time to review our research. We greatly appreciate your suggestions to improve our paper.  We have also ensured the English/grammar/sentence structure is optimal.  Each of your comments are addressed in detail below.

1. Regarding the results, although the cohort size is substantial, the number of recurrence events remains relatively small, with only 45 recurrences observed. Given this limited number of events, further clarification on model construction would strengthen the manuscript. Specifically, it would be helpful to describe the number of events included in the multivariable recurrence model and how many covariates were entered, in order to address potential concerns regarding overfitting. The wide confidence intervals for some predictors, particularly for higher pT stage, suggest that model stability may be limited.

• Thank you for these comments. In terms of model construction, the a priori variables included in the model are as indicated in Table 2. We have tried to make this clearer by adding a sentence in the results, section 3.2, page 6, line 185-186: The a priori variables included in RFS model for pN0 solitary chRCC patients are shown in Table 2.  As you point out, the event number is low and we made a point to ensure we did not analyze too many covariates. We have also tried to make it clearer what variables were added a prior and why by adding information in the methods section, page 3, line 117-118 and page 4, lines 121-126: Variables included in the models were based on prior published RCC literature and variables with biological plausibility. Variables included: pathological size of tumor (in mm), pT stage, presence of any sarcomatoid features, presence of any necrosis, and presence of a positive margin.  For overall survival (OS) modelling, Charlson Comorbidity Index score (CCI) and age were also included. CCI increased in 1-point increments and age at diagnosis increased in 1 year increments. In terms of describing the number of events in more detail, as you suggest, we have created a new figure (now labelled Fig 1) to further clarify this (the addition of this figure was also suggested by another reviewer). Regarding your comment about the wide confidence intervals, we have included this as a limitation in the discussion section on page 10, line 311-314. Other limitations relate to the low number of events in our study…….It resulted in wide confidence intervals for some of the significant variables in our RFS and OS models.

2. The exclusion of pN1 patients from the recurrence model due to instability is understandable from a statistical perspective. However, given that six of eight node-positive patients recurred, this subgroup appears biologically distinct and clinically high-risk. Even if formal modeling is limited by small numbers, the manuscript would benefit from emphasizing that pN1 disease likely represents an extremely aggressive subset within chRCC and deserves explicit clinical attention.

• Thank you for suggesting this addition.  We have included a sentence on page 9 line 2156-257: This population of pN1 patients likely represents an extremely aggressive subset of chRCC.

3. Because only 20 of 53 deaths were attributed to RCC and overall survival rather than cancer-specific survival was modeled, the issue of competing risks should be addressed. In a population with generally favorable prognosis and a non-negligible proportion of non-RCC deaths, a competing risk framework could provide additional insight. If such an analysis is not feasible, this limitation should at least be acknowledged.

• Thank you for this comment.  During our analysis phase, we did consider a competing risk strategy but chose not to.  This is because of the small event rate but more importantly n=16 (30%) of the deaths were from unknown causes.  Even though these patients were unlikely to have died of RCC, we did not want to make erroneous assumptions.  Based on your helpful comments however, we did add in the sentence in the Results, page 9, line 241-242: It should also be noted that many the patients in our cohort died of causes unrelated to RCC. We also acknowledge this limitation in the discussion, page 11, line 317-318: Finally, as with many real-word databases, some data is missing including cause of death in some patients which prevented us from performing a competing risk analysis. 

3. The recurrence pattern also warrants further elaboration. Since the majority of recurrences were metastatic, additional information regarding metastatic sites and whether recurrence timing differed by pathological stage would enhance biological interpretation and clinical applicability.

• Thank you for this helpful suggestion.  We are currently working on a more detailed paper looking at the patients who recurred in this cohort but also the  18 patients who presented with synchronous metastatic disease ie. Focusing on a larger cohort of metastatic chRCC patients and will be sure to include details regarding sites of metastases.  In terms of providing the median time to recurrence by pathological stage, we cannot provide that. The median time to recurrence is not estimable as the recurrence free survival curves do not go below 50%.  Your comment made us realize that the median time to recurrence we reported in the paper (1.9 years) should not be included for the same reason. We mistakenly calculated the median time to recurrence for only those who recurred, which is misleading. Thus, we deleted this data from the Results, section 3.1, page 5, line 169-170 as well as from the abstract.  For the exact same reason, we should not have included the median overall survival of 8.5 years and thus deleted this data from the Results, section 3.3, page 7,  line 202. Thank you for allowing us to make these corrections.

4. The discussion section would benefit from a more cautious tone regarding adjuvant therapy development. While the manuscript appropriately identifies a higher-risk subgroup, the overall prognosis of localized chRCC in this cohort is excellent, with 5- and 10-year recurrence-free survival rates of 94% and 90%, respectively. Given this low absolute recurrence risk, the clinical justification for aggressive development of adjuvant systemic therapy is debatable. The discussion currently leans toward suggesting that high-risk chRCC should be the focus of future adjuvant trials. However, the authors should more explicitly acknowledge the potential for overtreatment, the likely high number needed to treat, and the uncertain systemic therapy sensitivity of this histologic subtype. A more balanced interpretation would strengthen the scientific rigor of the manuscript.

• Thank you so much for this valuable comment.  You are completely correct.  Thus we have changed the last sentences of that paragraph in the Discussion on page 10, lines 299-305: Despite our study identifying a subset of chRCC patients at the highest risk of recurrence, conducting a randomized trials to examine the role of adjuvant therapy in these patients is impractical at this time due to the low event rate and thus the large sample size that would be needed. Also, the optimal systemic therapy for chRCC is still unknown.  Currently there is no evidence-based role for adjuvant systemic therapy in chRCC.

5. Furthermore, the excellent outcomes reported in this study raise a broader and potentially more impactful question: whether all surgically managed localized chRCC requires equally intensive intervention and surveillance. Rather than focusing primarily on escalation strategies, the discussion could also consider whether de-escalation approaches, including tailored follow-up or even active surveillance in selected low-risk tumors, may be appropriate. Given the favorable natural history demonstrated here, this perspective would add important nuance.

• Thank you for this very insightful comment. In the discussion, we have thus removed the first sentence (page 10, lines 275-276) so that we are not focusing this paragraph on just on the highest risk patients. As per your suggestion, we have added a sentence at the end of that paragraph on page 10 , lines 287-289 to bring up the concept of de-escalation: Our results also suggest there may be a subset of resected chRCC patients who have a very low risk of recurrence or death in whom follow up may be de-escalated.  Further research on this very low risk group is warranted. We did not make any additional comments about active surveillance in this patient population as it is beyond the scope of our research question.  We are however reviewing the n=94 patients in CKCis who had localized RCC who were not treated with surgical resection (surveillance, RFA, cryotherapy) and hopefully will be able to provide more insight to your readers in a future publication.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

To the  authors,

It has been properly corrected. Thank you for your hard work.