Methinks NCCT Stroke

Methinks NCCT Stroke is artificial intelligence software that analyzes non-contrast head CT images to detect and notify clinicians of suspected intracranial hemorrhage (ICH) and large vessel occlusion (LVO) affecting the ICA, MCA-M1, and MCA-M2 arteries. It is intended to assist hospital networks and trained physicians in workflow triage by prioritizing cases for further evaluation, not as a primary diagnostic tool.

Both the subject device and predicate (Rapid NCCT Stroke) use AI/ML neural networks to analyze NCCT images in parallel with standard clinical care and provide notifications via PACS. The subject device identifies an additional LVO location (MCA-M2) compared to the predicate (which covers ICA and MCA-M1 only), uses a single algorithm with two outputs rather than two cascaded functions, and operates in the cloud rather than on-premise or cloud options.

EN ISO 14971:2019 (risk management), ISO 62304:2015 (software development), and 21 CFR Part 820.30 (design controls, verification, and validation). Cybersecurity was addressed per Section 524B of the FD&C Act including vulnerability assessment, Software Bill of Materials (SBOM), and penetration testing.

Both devices share identical intended use (triage and notification for suspected ICH and LVO on NCCT), identical product codes and regulations (QAS, 21 CFR 892.2080), and comparable technological architecture (AI/ML processing of NCCT in parallel to standard care). The subject device's addition of MCA-M2 detection and single-algorithm architecture do not raise new safety or effectiveness questions because they represent incremental enhancements within the same clinical workflow paradigm. Performance data demonstrate the device exceeds pre-specified sensitivity and specificity goals (ICH: 94.7% sensitivity, 99.5% specificity; LVO: 76.4% sensitivity, 91.1% specificity) and does not introduce new risks compared to the predicate.