We appreciate the effort by Delora et al to help identify sources of false positives and false negatives in AI software.1 We were surprised to see that the software performance reported differed substantially from previous reports and from the Food and Drug Administration (FDA) clearance documents for both the Rapid AI and Viz AI software. In previous studies, the average sensitivity and specificity for the Rapid AI large vessel occlusion (LVO) software has been 94% and 90%, respectively.2–4 Results of prior Viz LVO studies show an average sensitivity of 92% and specificity of 87%.5–7 FDA 510K clearances cite 96% sensitivity and 98% specificity for Rapid LVO and 88% …