JBS-LVO

JBS-LVO is an artificial intelligence-powered software designed to expedite the triage of acute stroke patients by rapidly analyzing CT angiography images for suspected large vessel occlusions and notifying specialists through a mobile application. Its methodology relies on a convolutional neural network trained on diverse radiological data, demonstrating high sensitivity and specificity in LVO detection and aiming to improve patient care coordination via an integrated critical pathway management feature.

JBS-LVO is a radiological computer-aided triage and notification software designed to assist in the rapid identification of suspected large vessel occlusions (LVOs) in acute stroke patients. It employs a locked artificial intelligence algorithm, specifically a convolutional neural network (CNN), to analyze computed tomography angiography (CTA) images of the brain, focusing on the terminal internal carotid artery (ICA) and MCA-M1 segment. The system, which processes DICOM standard images, operates in parallel with standard care workflows, sending notifications of suspected LVOs to neurovascular specialists via a mobile application. This application allows for previewing compressed, informational images but is not intended for primary diagnosis; clinicians must review non-compressed images on a diagnostic viewer. The algorithm was trained on a diverse dataset of over 2,000 imaging studies sourced from various CT manufacturers and is designed to integrate into hospital networks, though it has a limitation of not processing scans containing metallic artifacts. Performance validation of JBS-LVO was conducted through a retrospective study using an independent dataset, with ground truth established by US board-certified neuroradiologists. The study demonstrated high accuracy for LVO detection, with a sensitivity of 91.8% (95% CI: 85.8% - 95.8%), specificity of 92.8% (95% CI: 87.2% - 96.5%), and an area under the curve (AUC) of 95.6% (95% CI: 93.0% - 98.1%). The system also proved efficient, with a mean time-to-notification from CTA acquisition to specialist alert for LVO-positive cases of 2.95 minutes (95% CI: 2.89 - 3.02 minutes), meeting its target of under 3.5 minutes. A distinctive feature of JBS-LVO is its 'Critical Pathway' (CP) functionality integrated into the mobile application, which provides real-time updates, activity logs, and push notifications to enhance care coordination and communication among medical staff involved in the patient's treatment pathway.

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. How does the 'Critical Pathway' feature specifically impact inter-clinician communication and patient management timelines compared to workflows without this feature?; What is the performance of the JBS-LVO algorithm in patient populations with subtle LVO findings or anatomical variations in the ICA and MCA-M1 segments?; Given the limitation of not processing scans with metallic artifacts, what proportion of acute stroke CTA scans are typically excluded by this device, and are there plans to address this limitation?; How does the algorithm's performance vary across different CT scanner manufacturers and imaging protocols beyond the initial training dataset diversity mentioned?; What are the human factors considerations and specific training requirements for clinicians to effectively integrate JBS-LVO notifications into their decision-making process without over-reliance or misinterpretation of informational previews?