Brain Metastases among Cancer Patients Diagnosed from 2010–2017 in Canada: Incidence Proportion at Diagnosis and Estimated Lifetime Incidence

Round 1

Reviewer 1 Report

Liu et al describe estimates of concurrent (also called "synchronous") and lifetime occurrence of brain metastases, using data from Canadian Cancer Registries for concurrent BM data and extrapolating to lifetime BM incidence from the literature. 

References from the literature were appropriate, but seem to have excluded what appears to be a relevant study: Ascha et al (2019) also report incidence proportions of synchronous and lifetime brain metastases using United States cancer registry data linked to Medicare claims for three primary cancer sites (breast, lung, and skin cancers). Please explain why this study was not included as part of the response to reviewers, or include the citation in the manuscript if you agree it is appropriate. 

The authors did a good job highlighting the importance of missing data to conclusions drawn from this work, particularly around missingness of data available from Quebec. The note in Table 2, "**The % missing were >10% before rounding procedures." is useful in the context of the main text, but lacks description in the table itself (e.g. in a caption). Please consider adding a caption to explain (if possible) or removing those notes in favor of presenting that information in main text. 

The authors note in discussion that "Melanoma cells are suggested to adopt phenotypic switching mechanism to resist the effect of 310 drugs and gain a greater ability to metastasize under targeted therapy [30]". This may be true, but some therapies (i.e. immunotherapies) may lead to "psuedoprogression" and potential false positive BM diagnosis. Please note the possibility of measurement/diagnostic error for patients on such therapy. 

Author Response

We thank the reviewer for their critiques and questions. The responses have been addressed in a point form format below:

Liu et al describe estimates of concurrent (also called "synchronous") and lifetime occurrence of brain metastases, using data from Canadian Cancer Registries for concurrent BM data and extrapolating to lifetime BM incidence from the literature. 

1. References from the literature were appropriate, but seem to have excluded what appears to be a relevant study: Ascha et al (2019) also report incidence proportions of synchronous and lifetime brain metastases using United States cancer registry data linked to Medicare claims for three primary cancer sites (breast, lung, and skin cancers). Please explain why this study was not included as part of the response to reviewers, or include the citation in the manuscript if you agree it is appropriate. 

We did not use results from Ascha et al (2019) when estimate lifetime brain metastases because Ascha et al restricts its study population to individuals aged 65+. As we aimed to estimate concurrent and lifetime incidence for the entire Canadian cancer afflicted population, the comparison of the entire Canadian population to a limited population of the elderly in the US would have been inappropriate.

2. The authors did a good job highlighting the importance of missing data to conclusions drawn from this work, particularly around missingness of data available from Quebec. The note in Table 2, "**The % missing were >10% before rounding procedures." is useful in the context of the main text, but lacks description in the table itself (e.g. in a caption). Please consider adding a caption to explain (if possible) or removing those notes in favor of presenting that information in main text. 

Thank you for your suggestions. We added the following text to the footer of Table 2: "Provinces with greater than 10% missing data were excluded from the IPR estimates. Brain metastases at diagnoses were completely missing in Quebec."

3. The authors note in discussion that "Melanoma cells are suggested to adopt phenotypic switching mechanism to resist the effect of 310 drugs and gain a greater ability to metastasize under targeted therapy [30]". This may be true, but some therapies (i.e. immunotherapies) may lead to "psuedoprogression" and potential false positive BM diagnosis. Please note the possibility of measurement/diagnostic error for patients on such therapy. 

This is a good point. We included “pseudoprogression following immunotherapies” in the discussion. (Line 316): "Among targeted therapies used, immunotherapy administration of anti-CTLA-4 antibodies has been described lead to the uncommon occurrence of pseudoprogression of melanoma, potentially causing new lesions [33]. However, subsequent tumor repression is observed with this phenomenon. This may lead to false positives of BM diagnoses leading to overestimation of BM incidence cases with the increasing prevalence of immunotherapy usage."

Reviewer 2 Report

This is a study based on the Canada Cancer Registry in regard to brain metastases at diagnosis and through the course of disease. This covers a period of 2010-2017. The literature appears to be relatively limited for Canada during the time frame.

This is a very well written paper that is also devoid of any grammatical issues. The 

The paper is challenged by large gaps in the reporting rates of the different provinces. In many cases no information is available.

The level/extent of follow-up is difficult to know in this study as data is retrospective and slow to be entered. This can definitely effect results and should be explained.

The number and rate of brain metastases appears to be considerably lower than previously published results from other countries. One particular slightly unrelated example is melanoma rates which may reflect isolated lesions vs metastatic?

Author Response

We thank the reviewer for their critiques and questions. The responses have been listed in a point form format below:

This is a study based on the Canada Cancer Registry in regard to brain metastases at diagnosis and through the course of disease. This covers a period of 2010-2017. The literature appears to be relatively limited for Canada during the time frame.

This is a very well written paper that is also devoid of any grammatical issues. The paper is challenged by large gaps in the reporting rates of the different provinces. In many cases no information is available.

1. The level/extent of follow-up is difficult to know in this study as data is retrospective and slow to be entered. This can definitely effect results and should be explained.

The CCR is updated on a yearly basis with Canadian provinces submitting their cancer registry data to Statistics Canada. In 2021, data up to 2017 was released with the CCR website stating “delays in the reporting of new cases to Statistics Canada typically results in undercounts of cases which are more pronounced in the most recently reported diagnosis year. Generally, the reporting delay ranges between 2% and 3% nationally”. We would like to point out that cancer patients were only passively followed up after their cancer diagnosis by the surveillance system -- their vital status is updated through data linkage to the national vital status database. Cancer progression and recurrence data were neither collected nor reported by the provincial cancer registries.

We add the following sentence to the methods section to clarify: (Line 68) “Due to delays in reporting, undercounting of cases is most prominent in the last reported diagnosis year of 2017. The under-reporting is estimated to be 2% to 3% for all cancers combined and could contribute to the underestimation of both concurrent and lifetime BM counts [16]”.

2. The number and rate of brain metastases appears to be considerably lower than previously published results from other countries. One particular slightly unrelated example is melanoma rates which may reflect isolated lesions vs metastatic?

We compared our concurrent incidence proportions with the US rates reported in Cagney et. al. (2017)¹⁸ using SEER data. Their study population restricted to patients 18 years and older and invasive cancer only while our study population has no restriction on age at diagnosis and include in situ cancer diagnosis.

For cancer sites such as esophagus, colorectal, renal, and melanoma, the concurrent incidence proportions are consistent between the two countries, considering the 95%CIs of our estimates. For example, the Canadian IP for melanoma was 0.73 (95% CI 0.61-0.84) and the US IP was 0.65, which is within our 95%CI.

For breast and lung cancers, the Canadian concurrent BM estimates were indeed considerably lower. Canada has organized breast cancer screening programs in each provinces, which could contribute to a smaller proportion of stage 4 breast cancer in Canada than in US. For example, in Alberta where the authors are based, the stage 4 breast cancer account for 3.3% of all breast cancer diagnosis in 2004-2010 while Cagney reported 5.4% of all breast cancer were stage 4. This may explain a lower IP in breast cancer patients in Canada. As for lung cancer, we don’t have an explanation which can be investigated in a future study.