BriefCase-Triage

BriefCase-Triage is designed to assist medical specialists by expediting the identification of suspected Central Pulmonary Embolism in CTPA images for adults, thereby optimizing workflow triage. The system utilizes a cloud-based deep learning artificial intelligence algorithm to analyze DICOM images, flagging potential positive cases and delivering rapid, non-diagnostic notifications with preview images to prioritize radiologist review.

BriefCase-Triage is a radiological computer-aided triage and notification software designed for analyzing Computed Tomography Pulmonary Angiography (CTPA) images in adults and transitional adolescents (18 years and older). Its primary purpose is to assist hospital networks and trained medical specialists in optimizing workflow triage by flagging and communicating suspected positive cases of Central Pulmonary Embolism (Central PE). The system operates in parallel with standard image interpretation, providing notifications for suspected Central PE findings to facilitate quicker attention to potentially critical cases; these notifications include compressed preview images intended for informational and prioritization purposes only, not for diagnosis, and the device does not alter the original medical images. The software employs a deep learning artificial intelligence algorithm, trained on manually labeled CTPA images, which runs on a Linux-based server in a cloud environment. It processes filtered DICOM images chronologically and, upon detecting a suspected Central PE, sends a pop-up notification with a non-diagnostic, low-quality preview image to a desktop application for review by appropriately trained medical specialists. A retrospective, multicenter clinical study involving 328 unique patient cases from six US-based sites demonstrated the device's efficacy, achieving a sensitivity of 89.2% and specificity of 94.5% for Central PE detection; furthermore, the mean time-to-notification for positive cases was 29.3 seconds, showcasing a significant improvement over the predicate device and highlighting its potential to expedite the review of time-sensitive findings.

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 (sensitivity, specificity, PPV, NPV) vary when encountering rare or atypical presentations of Central PE not extensively represented in the training dataset?; What is the real-world impact of the 29.3-second mean time-to-notification on clinical decision-making speed and patient outcomes, beyond comparison to a predicate device?; Given that the preview image is "compressed, low-quality, grayscale" and "not for diagnostic use," what studies have been conducted to assess the potential for misinterpretation or cognitive bias introduced by these previews before the clinician views the full images?; How does the algorithm's performance specifically adjust or account for variations in image quality or acquisition protocols across different CT scanners (Philips, GE, Siemens, Toshiba) and imaging sites?; What mechanisms are in place for ongoing monitoring and retraining of the AI algorithm to maintain performance as imaging technologies evolve and new Central PE characteristics emerge in the patient population?