Abstract
Brain Cancer Canada Travel Award Recipient Brain metastases (BMs) exceed primary CNS tumours and comprise the majority workload in neuro-oncology. However, estimates of post-diagnosis BMs incidence are lacking. We address this gap and provide a population-based 10-year incidence rate of BMs following cancer diagnosis. We identified cancer patients diagnosed between 10/01/2012 and 09/30/2013 in Alberta, who were free from brain tumours at diagnosis. Brain/head-related radiology reports following cancer diagnosis for all patients were retrieved from a provincial radiology database up to 10/31/2022. Stratified by their number of follow-up reports, we drew patient samples from each stratum. All reports from sampled patients were manually labeled for BM (yes/no). We estimated BM incidence with three methods: (1) weighting using sampling weights, (2) simulating BM labels based on BM probability and distributions of the first BM occurrences, and (3) predicting BM labels using a natural language processing model. Of 11,800 eligible cancer patients, a total of 11,826 follow-up reports were identified from 4,647 patients having at least one report. 1,262 reports from 249 sampled patients were labeled. The estimated 10-year incidence rate (per 100,000 person-years) was 437.7, 436.6 (95%CI: 422.7–452.6), and 498.2 (95%CI: 447.8–552.3) from the three methods, respectively. The estimated cumulative incidence of BM at 10-year post-diagnosis was 4.16% (95% CI: 3.74–4.65%, method 2) and 4.43 (95%CI: 3.97–4.90%, method 3). We provide a first estimate of the incidence rate and cumulative incidence of BMs following cancer diagnosis, offering valuable insights for healthcare resource planning.