Development and Validation of the Predicting Risk of Ischemic Stroke in Malignancy Estimation Tool

BACKGROUND: The risk of ischemic stroke is highest during the first year following a new diagnosis of cancer, but no tools exist to identify patients at highest risk. METHODS: Using linked clinical and administrative databases, we conducted a population-based retrospective cohort study of adults in Ontario, Canada, with newly diagnosed cancer from 2010 to 2021. Patients were randomly selected for prediction model derivation (60%) or validation (40%). The final model predicting ischemic stroke within 1 year following cancer diagnosis was derived using multivariable Fine-Gray regression with candidate predictors selected via backward elimination. Subdistribution-adjusted hazard ratios and 95% CIs were calculated, where all-cause mortality was treated as a competing event. Performance of the prediction model was assessed in the validation cohort based on the C statistic and calibration plots for discrimination and calibration, respectively. RESULTS: There were 698 566 eligible patients, with 418 911 in the derivation cohort and 279 576 in the validation cohort. The overall rate of stroke per 1000 person-years was 6.7 (95% CI, 6.4-6.9). The final model included 22 predictors, including age, sex, demographic factors, cancer characteristics, and treatment characteristics. Discrimination was good, with a C statistic of 0.73. The model was well calibrated, with points following the desired 45-degree line. CONCLUSIONS: We derived and validated the PRIME (Predicting Risk of Ischemic Stroke in Malignancy Estimation) tool with good discrimination for ischemic stroke in patients with a new cancer diagnosis. The model was built into an online calculator (https://study.ohri.ca/PRIME/) and has the potential to stratify patients with cancer based on their risk of stroke within a year following their diagnosis.