Development and validation of a risk assessment tool for ischemic stroke in cancer patients

Cancer patients have an increased risk of ischemic stroke. However, there is no reliable risk prediction model. We aimed to create a risk prediction model for ischemic stroke in cancer patients within the first 2 years after cancer diagnosis. A cohort of 26,717 cancer patients from the University of Osaka Hospital was used to develop the model. A cohort of 31,881 patients from Osaka International Cancer Institute was used for external validation. Data in the development cohort generated an AHANDS score: Age ≥ 75 years, Hypertension, Atrial fibrillation, a high Neutrophil-to-lymphocyte ratio, an elevated D-dimer level, and cancer Stage IV or distant metastasis at the time of cancer diagnosis. Discrimination was assessed using the c-statistic. We compared the AHANDS and Khorana scores in the development cohort. Ischemic stroke occurred within 2 years of diagnosis of cancer in 163 patients (0.61%) in the development cohort and in 99 (0.31%) in the validation cohort. The AHANDS score performed better than the Khorana score (c-statistic 0.703 vs. 0.54; p < 0.0001). A similar result was obtained in the external validation cohort (c-statistic 0.75). This score is intended to guide future stroke prevention strategies in patients with cancer.