Tumour Mutational Burden and Its Relationship with Clinical Outcomes in Locally Advanced and Recurrent/Metastatic Adenoid Cystic Carcinoma with and Without NOTCH Pathway Activation

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Overall, this manuscript addresses a relevant and interesting topic in ACC, particularly regarding the relationship between NOTCH pathway activation, TMB, and clinical behavior. The study includes a relatively large cohort for such a rare malignancy, which certainly adds value to the work. In general, the manuscript is clearly written, objective, and well-organized. Nevertheless, a few points deserve further discussion and clarification in order to strengthen the interpretation of the data and the conclusions presented.

First, a substantial proportion of tumor samples were obtained from metastatic sites, especially in the MSK cohort. Since TMB may vary between primary tumors and metastases, it would be important to discuss whether differences in TMB, or even in the frequency of NOTCH GoF mutations, were observed according to biopsy site. In addition, was any multivariable adjustment performed accounting for biopsy origin (primary versus metastatic tissue) or specific metastatic sites? This seems particularly relevant given the higher frequency of bone and CNS metastases in the MSK cohort.

Another point that would benefit from more in-depth discussion is the finding that patients with intermediate TMB values (3–5 mut/Mb) had worse survival, without a clear linear relationship between increasing TMB and poorer prognosis. At present, the discussion does not sufficiently explore the possible biological explanation for this somewhat paradoxical observation. Since this effect disappeared after excluding tumors with NOTCH1/2 GoF mutations, could this association be primarily driven by the enrichment of NOTCH-activated tumors within this subgroup rather than by TMB itself? This possibility seems biologically plausible and deserves more explicit consideration.

Additionally, therapeutic information appears incomplete across the cohorts. Given that overall survival is a major endpoint of the study, additional information regarding systemic treatments received by these patients would be important, particularly for those with more aggressive disease or NOTCH alterations, as treatment heterogeneity may substantially influence survival outcomes.

Finally, although the findings are interesting, some of the translational implications may currently be somewhat overstated considering the retrospective nature of the study, the heterogeneity between cohorts, and the absence of independent prospective validation. It may therefore be more appropriate to present the clinical applicability of TMB in ACC more cautiously at this stage.

Author Response

Please see the attachment.

From attachment:

Reviewer 1:

Overall, this manuscript addresses a relevant and interesting topic in ACC, particularly regarding the relationship between NOTCH pathway activation, TMB, and clinical behavior. The study includes a relatively large cohort for such a rare malignancy, which certainly adds value to the work. In general, the manuscript is clearly written, objective, and well-organized. Nevertheless, a few points deserve further discussion and clarification in order to strengthen the interpretation of the data and the conclusions presented.

First, a substantial proportion of tumor samples were obtained from metastatic sites, especially in the MSK cohort. Since TMB may vary between primary tumors and metastases, it would be important to discuss whether differences in TMB, or even in the frequency of NOTCH GoF mutations, were observed according to biopsy site. In addition, was any multivariable adjustment performed accounting for biopsy origin (primary versus metastatic tissue) or specific metastatic sites? This seems particularly relevant given the higher frequency of bone and CNS metastases in the MSK cohort.

Another point that would benefit from more in-depth discussion is the finding that patients with intermediate TMB values (3–5 mut/Mb) had worse survival, without a clear linear relationship between increasing TMB and poorer prognosis. At present, the discussion does not sufficiently explore the possible biological explanation for this somewhat paradoxical observation. Since this effect disappeared after excluding tumors with NOTCH1/2 GoF mutations, could this association be primarily driven by the enrichment of NOTCH-activated tumors within this subgroup rather than by TMB itself? This possibility seems biologically plausible and deserves more explicit consideration.

Additionally, therapeutic information appears incomplete across the cohorts. Given that overall survival is a major endpoint of the study, additional information regarding systemic treatments received by these patients would be important, particularly for those with more aggressive disease or NOTCH alterations, as treatment heterogeneity may substantially influence survival outcomes.

Finally, although the findings are interesting, some of the translational implications may currently be somewhat overstated considering the retrospective nature of the study, the heterogeneity between cohorts, and the absence of independent prospective validation. It may therefore be more appropriate to present the clinical applicability of TMB in ACC more cautiously at this stage.

Thank you for your comments.

We were also concerned there may be differences between biopsies obtained from metastatic and primary sites. We tested whether there was a significant difference between TMB values obtained from metastatic sites vs primary sites using Wilcoxon’s test. There were no significant differences detected in either cohort. To make this clearer in the text, we have amended this sentence in the results (page 9, paragraph 3, lines 266-268) to now read the following:
“As a result, we investigated the associations between TMB and biopsy site and no significant differences were seen in either the NHS cohort (p=0.1) or the MSK cohort (p=0.6).”

We have also tested whether NOTCH1 mutations were significantly different in frequency between the primary and metastatic biopsy sites. No significant difference was found for either cohort (NHS p=0.16, MSK p=0.81). The following has been added to the results (page 7, paragraph 1, lines 206-208) to clarify this:
“There were no significant differences in the rate of NOTCH GoF mutations between metastatic or primary biopsy sites in either the NHS cohort (p=0.16) or MSK cohort (p=0.81).”

Multivariate analysis was not done to account for this as it was not found to be significantly different between groups. No analysis was done regarding subsite of biopsy.

We agree that activating NOTCH mutations are likely driving the shorter OS in the 3-5 mut/Mb category in the MSK cohort. We have added the following line to the discussion (page 10, paragraph 2, lines 307-310) to make this clearer:
“The NOTCH mutations may be driving the shorter survival rather than the TMB in this subcohort, which is supported by the subcohort no longer having a significantly-shorter survival once removing patients with NOTCH GoF mutations.”

The paucity of data on therapeutics received by patients within the cohorts is a limitation of the dataset. We have amended the following line in the limitations in the discussion (page 11, paragraph 5, lines 358-362) to better acknowledge this:
“In addition, although clinical data ascertaining patient demographics and disease spread is relatively comprehensive, due to the retrospective nature of this dataset treatment details are incomplete and therefore we were not able to analyse any impact of treatment on survival outcomes. As such, further prospective studies in larger well controlled cohorts remain a top priority for future research.”

We agree that the strength of the findings are limited by the retrospective and heterogenous nature of the study. We have modified the following sentence in the conclusion (page 12, paragraph 2, lines 377-379) to better reflect this uncertainty:
“This may be helpful in directing clinical management including timing of interval surveillance imaging and clinician-patient discussions regarding earlier initiation of systemic therapies and warrants validation in a prospective study.”

Reviewer 2 Report

Comments and Suggestions for Authors

Is it possible to add the RNAseq results and present correlation of the fusions seen in the reminder of the NOTCH negative patients. Since major drivers for the Adenoid Cystic Carcinoma are fusions, it would be clinically important to review and add the data to cohort.

I suggest adding more molecular detail about the NOTCH1 alterations. Are all these NOTCH1 gain of function mutations present in the PEST domain of the gene? Are they similar to those seen other disease such as oral cavity squamous cell carcinoma or more similar to those seen in Melanoma or CLL?  

Author Response

Please see the attachment.

From attachment:

Reviewer 2:

Is it possible to add the RNAseq results and present correlation of the fusions seen in the reminder of the NOTCH negative patients. Since major drivers for the Adenoid Cystic Carcinoma are fusions, it would be clinically important to review and add the data to cohort.

I suggest adding more molecular detail about the NOTCH1 alterations. Are all these NOTCH1 gain of function mutations present in the PEST domain of the gene? Are they similar to those seen other disease such as oral cavity squamous cell carcinoma or more similar to those seen in Melanoma or CLL?  

Thank you for your comments.

 We agree that the central genetic change in ACC is considered to be MYB or MYBL1 fusions, and we are aware that the Persson paper shows there may be differences in outcome by type of fusion on FISH. However, performing MYB FISH analysis is not standard of care for ACC at present and only DNA analysis was included in this study. Therefore, we are not able to provide this analysis. We have added the following to the discussion (page 11, paragraph 6, lines 363-367) to acknowledge this:
“ACC is known to be predominantly driven by MYB or MYBL1 fusions, most commonly MYB::NFIB fusions. Recent data suggests that the type of MYB fusion detected by fluorescent in-situ hybridization (FISH) testing may influence survival outcomes in ACC [53]. However, MYB fusion FISH is not currently standard of care for ACC patients and insufficient data were available to include in the analysis of this study.”

We have included details on the NOTCH testing in the methods section. NOTCH1 alterations were included if they were predicted to be gain of function in the PEST or NRR regions as is commonly accepted in ACC. However, for clarity we have added this to the discussion (page 10, paragraph 3, lines 311-313):
“ACC with NOTCH GoF mutations in the PEST or NRR domains represent a distinct subtype of ACC associated with attenuated progression-free and overall survival times [21,40,43,44].”

Reviewer 3 Report

Comments and Suggestions for Authors

Dear Authors

Identification of the different genetic subtypes of ACC has a significant  impact on the tumor overall surveillance and treatment strategies. The presented study is properly designed. However,  implementation of the obtained results for the clinical policy is limited. The main drawbacks of the presented study is a lack of the detailed clinical features of the used tumor specimens. It makes the introduction of the presented results into clinical practice more difficult. Especially, that the reviewed study is based on the retrospective analysis.  For improvement of the obtained results presentation, p values should be completed into Tab 1 and Tab 2 to compare both cohorts. In the discussion subsection the authors should attempt to combine specific types of genetic mutations with susceptibility to specific chemotherapeutics.  

Author Response

Please see the attachment.

From attachment:

Reviewer 3:

Dear Authors

Identification of the different genetic subtypes of ACC has a significant  impact on the tumor overall surveillance and treatment strategies. The presented study is properly designed. However,  implementation of the obtained results for the clinical policy is limited. The main drawbacks of the presented study is a lack of the detailed clinical features of the used tumor specimens. It makes the introduction of the presented results into clinical practice more difficult. Especially, that the reviewed study is based on the retrospective analysis.  For improvement of the obtained results presentation, p values should be completed into Tab 1 and Tab 2 to compare both cohorts. In the discussion subsection the authors should attempt to combine specific types of genetic mutations with susceptibility to specific chemotherapeutics. 

Thank you for your comments.

We agree that further information on the two cohorts would be advantageous and this is a limitation of the retrospective nature of the study. We have added the following to the discussion (page 11, paragraph 5, lines 358-362):
“In addition, although clinical data ascertaining patient demographics and disease spread is relatively comprehensive, due to the retrospective nature of this dataset treatment details are incomplete and therefore we were not able to analyse any impact of treatment on survival outcomes. As such, further prospective studies in larger well controlled cohorts remain a top priority for future research.”

We have added P values to Table 1 and Table 2 to improve the presentation. We have adjusted the results section (page 5, paragraph 2, lines 156-162) to reflect this:
“In cases where the biopsy location was known, 47/95 (50%) tumour samples were taken from metastatic sites in the NHS group, whereas 100/139 (72%) specimens in the MSK group were procured from metastases, a significantly higher proportion. There was a higher proportion of tumour samples from patients with bone and CNS metastases in the MSK cohort compared with the NHS cohort and the NHS cohort had significantly more patients with metastatic disease in the lungs only, but the overall distribution of metastatic disease was not significantly different.”

Unfortunately there has only been limited clinical research into targeting specific genetic alterations in ACC. The most heavily-investigated has been the NOTCH pathway, with the ACCURACY trial investigating AL101. Otherwise, only MYB fusions have promising data with the recent publications of MYB degrader therapies. We have incorporated this into the discussion (page 11, paragraph 7 and page 12, paragraph 1, lines 368-373):
“There has been limited investigation into targeting specific gene alterations in ACC. NOTCH GoF has been evaluated as a candidate drug target in the ACCURACY study, although there was no clear signal of efficacy [54]. MYB fusions are also of therapeutic relevance with ongoing studies of MYB degrader therapies. These have recently posted results suggesting a signal of efficacy [55,56]. No other genetic targets have been formally tested.”

Round 2

Reviewer 3 Report

Comments and Suggestions for Authors

I accept the revised manuscript in current form.