JLK-ICH

The JLK-ICH is a radiological computer-aided triage and notification software designed to rapidly identify intracranial hemorrhage (ICH) from non-contrast computed tomography (NCCT) images of the head. This device employs advanced artificial intelligence and machine learning algorithms to operate as a notification-only, parallel workflow tool, demonstrating high sensitivity and specificity in its detection and achieving substantial equivalence to a legally marketed predicate device.

JLK, Inc.'s JLK-ICH is a radiological computer-aided triage and notification software designed to analyze non-contrast computed tomography (NCCT) images of the head for findings suggestive of intracranial hemorrhage (ICH). Functioning as a notification-only, parallel workflow tool, it aims to assist hospital networks and clinicians in prioritizing patient images for review, thus streamlining critical care. A comprehensive assessment determined the device is substantially equivalent to the legally marketed predicate device, Viz ICH, allowing it to be marketed for its stated indications for use under applicable regulations. The JLK-ICH system leverages advanced artificial intelligence and machine learning algorithms, hosted on servers and accessible via a mobile application, to process NCCT scans and deliver timely notifications. Its performance validation testing, including a standalone evaluation on an independent dataset, demonstrated robust results with a sensitivity of 97.3% and specificity of 97.9% for ICH detection. Additionally, its average NCCT-to-notification time of 0.19 minutes is faster than its predicate, confirming its safety and effectiveness without raising new concerns compared to the predicate device.

AI Regulatory Concerns

Powered by

Jev reads the full FDA letter and returns calibrated concern probabilities per regulatory category. These are AI estimates, not regulatory advice.

• Q1. What specific types of intracranial hemorrhage (e.g., epidural, subdural, intraparenchymal) did the validation testing encompass, and were there performance differences across these types?; How does the rapid notification time of 0.19 minutes translate into improved clinical outcomes, such as reduced time to treatment or improved patient prognosis, in real-world settings?; Given its 'notification-only' and 'parallel workflow' design, what are the best practices for integrating JLK-ICH into diverse hospital IT infrastructures and existing clinical decision pathways to maximize its utility?; Could the AI/ML algorithms be trained to identify other critical findings on NCCT head scans, such as mass effect, midline shift, or hydrocephalus, to further enhance its value?; What post-market surveillance or real-world performance data will be collected to monitor the long-term safety, effectiveness, and consistency of the JLK-ICH system's performance across varied clinical environments?