Abstract Body: Introduction: Corticospinal tract (CST) lesion load (LL) based on a CST template is commonly used as a predictor of motor outcomes after stroke, but the optimal approach is still unclear. We compared 4 LL metrics and 3 CST templates to explore which approach explains the most variance in motor outcome, hypothesizing that strongest results would be found using elderly subjects to create the CST template. Methods: Data from the ENIGMA dataset were studied, with n=221 after exclusion for data quality; missing data; and bilateral, brainstem, or cerebellar stroke (median=62 years). The four lesion load metrics were 1) grid split of the CST into 16ths using (Fig 1A) and assessing the number of sections with > 5% injury, 2) radial split of the CST into 16ths (Fig 1B) and assessing the number of sections with >5% injury, 3) computing the LL per slice weighted by the inverse diameter of the CST (WLL, Fig 1C) then taking the AUC across all slices, and 4) taking the WLL value from the maximally affected slice (Max-WLL). The three tract templates were: 1) a previously published template derived from 12 healthy participants (31-68 years), 2) a template derived from 1052 participants from the Human Connectome Project and 3) a template created using a subset of 326 subject ages 50-80 years from the HCP Aging (HCPA) study. Bayesian ordinal regression models were used to look at relationship between motor outcome and the LL for each method, controlling for M1 LL, corpus callosum LL, age, lesion side, and lesion volume. Results. Heatmaps show the most frequently injured CST slice (Fig 2A) and lesion location (Fig 2B). Modeling with Max-WLL of the HCPA template was the optimal approach, explaining 40% (95%CI: 29%, 48%) of variance in motor score, with Max-WLL accounting for 52.4% of the model’s explanatory value (95%CI: 32%, 63%) (Fig 3C). The effect of LL was nonlinear (Fig 3B). The R² for the other methods was 35% (Grid Split), 35% (Radial Split) and 39% (WLL). Results remained significant whether the motor outcome was impairment or function. Discussion. Results are consistent with recent work showing Max-WLL is optimal for explaining motor outcomes, and with the fact that CST diameter and integrity changes with age. Findings indicate that an age-matched template better captures CST anatomy in elderly patients, and that focusing on the brain slice with maximal CST injury provide the optimal approach to calculating CST injury for understanding motor outcomes after stroke.