JLK-NCCT

JLK-NCCT is radiological computer-aided triage and notification software that analyzes non-contrast head CT (NCCT) images using artificial intelligence to detect suspected intracranial hemorrhage (ICH) and large vessel occlusion (LVO) in the internal carotid artery, middle cerebral artery M1, and M2. It sends mobile notifications with compressed preview images to alert clinicians for triage and workflow prioritization, without modifying original images or replacing standard diagnostic methods.

Both JLK-NCCT and predicate device (Rapid NCCT Stroke, K222884) use AI/ML algorithms to process DICOM-format NCCT images and generate automated notifications. Key differences: JLK-NCCT expands LVO detection to include MCA-M2 (predicate only covers ICA and MCA-M1), uses mobile-only notifications (predicate supports PACS, email, and mobile), and employs different AI models (JLK-ICH and HAS models versus hypodensity and HVS models). Both serve as notification-only tools operating parallel to standard care.

The document references ISO 13485 (Design Controls, Nonconforming Product, Corrective and Preventative Action) and DICOM standards for image processing. FDA guidance documents cited include 'Content of Premarket Submissions for Device Software Functions' (June 14, 2023). The special controls standard referenced is 21 CFR 892.2080.

JLK-NCCT is substantially equivalent because it has the same intended use as the predicate (triage and notification for ICH and LVO), similar technological architecture (AI/ML-based NCCT analysis with mobile notifications), and comparable clinical performance (78.5% sensitivity, 90.3% specificity for LVO detection; 1.67-minute notification time). The differences in AI algorithms and expanded MCA-M2 detection do not introduce new safety or effectiveness questions, and the ICH component is identical to a previously FDA-cleared device (K243363).