Radiology— FDA advisory panel RA

Radiology covers the imaging chain end to end: X-ray systems, CT, MRI, diagnostic ultrasound, mammography hardware, image-intensified fluoroscopy, PACS and image-management software, and — the panel's defining recent dynamic — software-only devices, including computer-assisted detection and triage products. Most of the medical-AI wave clears through this panel: an algorithm that flags intracranial hemorrhage on CT or prioritizes a worklist files as a radiological device, typically citing a software predicate.

This is my own working territory — my day job is deploying clinical imaging AI inside hospitals — and the practical advice I'd give a reader of this panel's data is to distinguish hardware iterations (a new detector generation, largely predictable) from the software product codes, where clearance volume has grown fastest and where reading the actual 510(k) summary matters most: two clearances under the same software code can describe products with very different clinical claims.