Annalise Enterprise CTB Triage-OH

The Annalise Enterprise CTB Triage -OH is an AI-powered software designed to enhance radiology workflow by identifying potential obstructive hydrocephalus in adult non-contrast brain CT scans. It utilizes a deep-learning convolutional neural network to analyze images and alert clinicians for prioritized review, demonstrating high sensitivity and specificity in performance testing.

The Annalise Enterprise CTB Triage -OH is a software tool designed to assist clinicians in managing non-contrast brain CT studies for adult patients aged 22 years and older. Its primary function is to aid in the triage and prioritization of studies that exhibit features suggestive of obstructive hydrocephalus. The system analyzes these CT scans using an artificial intelligence algorithm and provides study-level notifications to existing order and imaging management systems, such as PACS/RIS. This facilitates earlier review of potentially critical cases by reordering the worklist, though it is explicitly stated that the device never downgrades a study's existing priority. It is intended for use by trained clinicians qualified to interpret brain CTs and operates in parallel with standard clinical workflows, serving as an aid rather than a standalone diagnostic tool or a means to rule out findings. The core technology is an artificial intelligence algorithm, specifically a convolutional neural network developed using deep-learning techniques. This network was trained on an extensive dataset of over 200,000 CT brain studies, which were annotated by radiologists and included images from a variety of scanner manufacturers. The device's performance was validated through a retrospective standalone study comparing its findings against a ground truth established by board-certified neuroradiologists. For obstructive hydrocephalus, the device demonstrated high sensitivity and specificity across different CT slice thicknesses; for instance, with a slice thickness of ≤ 1.5 mm at one operating point, sensitivity was 97.3% and specificity was 94.0%. A separate bench study also assessed triage effectiveness, reporting an average turn-around time of approximately 81.6 seconds for identifying and prioritizing cases.

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 algorithm's performance vary with different severities or subtle presentations of obstructive hydrocephalus within the 'positive' cases?; What was the inter-rater reliability among the neuroradiologists establishing the ground truth, and how were discrepancies resolved beyond requiring a third truther?; Were there specific types of co-existing pathologies or image artifacts that most frequently impacted the algorithm's accuracy for obstructive hydrocephalus?; How does the device's performance generalize to CT scanners not heavily represented in the pivotal study or to patient populations with different demographic distributions than the test set?; What is the computational time required for the AI to analyze a typical CT brain study, and how does this integrate into the overall reported triage turn-around time of 81.6 seconds?