Objective: To assess the performance of the Brainomix 360 software (Brainomix) in evaluating ischemic core volumes on NCCT compared with CT perfusion (CTP) in patients with acute ischemic stroke.

Methods: This retrospective multicenter study included patients with anterior circulation large vessel occlusions who underwent pretreatment NCCT and CTP, successful reperfusion (modified Thrombolysis in Cerebral Infarction ≥2b), and post-treatment MRI from three stroke centers. Ischemic core volumes were calculated on NCCT using the ML algorithm deployed by Brainomix 360 and on CTP using Olea software (Olea Medical). Comparative analysis was performed between core volumes estimated on NCCT, CTP, and MRI-derived final infarct volume (FIV).

Results: A total of 111 patients were included. Estimated ischemic core volumes (mean ± SD, mL) were 20.4 ± 19.0 on NCCT and 19.9 ± 18.6 on CTP, with no significant difference (P = 0.82). A moderate (r = 0.40) but significant (P < 0.001) correlation was found between NCCT and CTP core estimates. The mean difference between FIV and core volume estimates on NCCT and CTP was 29.9 ± 34.6 mL and 29.6 ± 35.0 mL, respectively (P = 0.94). Correlations between FIV and core volume estimates were similar for NCCT (r = 0.30, P = 0.001) and CTP (r = 0.36, P < 0.001).

Conclusions: Machine learning-based ischemic core volume estimates on NCCT using Brainomix 360 are comparable in magnitude and correlation to those obtained from concurrent CTP and MR-assessed FIV.