CXRSecond

CXRSecond is an AI-assisted web app that provides a “second read” for chest X-rays (CXR) to help clinicians catch common, high-impact findings such as pneumothorax, pleural effusion, consolidation, cardiomegaly, and misplaced lines/tubes. It ingests DICOM images, runs FDA-cleared-model-style triage logic (positioning, quality checks, then finding probabilities), and outputs a structured summary with heatmaps and a confidence/limitations statement. It’s designed for urgent care centers, small hospitals, and teleradiology groups that don’t have 24/7 subspecialty coverage. The app focuses on workflow: it flags studies that need immediate attention, generates a draft impression, and logs clinician feedback to monitor drift and performance by site. It does not pretend to “diagnose”; it prioritizes, explains, and documents. Realistically, the hard part is regulatory, liability, and integration—not the model.