The American College of Radiology (ACR) Council has approved a joint ACR-SIIM (Society for Imaging Informatics in Medicine) Practice Parameter for Imaging Artificial Intelligence (AI).
The practice parameter, developed with SIIM, applies to physicians, technologists, medical physicists, informatics and IT teams, data scientists, and administrators who deploy or use AI in imaging workflows. It covers AI tool selection, predeployment evaluation, ongoing performance monitoring, and patient privacy protection. Facilities that implement AI in accordance with the parameter can earn the ACR Recognized Center for Healthcare-AI (ARCH-AI) designation, the ACR said.
The ACR Data Science Institute (DSI) meanwhile has published details of Assess-AI, described as the world's first AI quality registry for medical imaging, in the Journal of the American College of Radiology.
Assess-AI, the newest ACR National Radiology Data Registry, supports post-deployment AI governance by measuring concordance between clinical AI outputs and radiology report-derived data. The service integrates deidentified data via ACR Connect with centralized analytics and national benchmarking, allowing facilities to compare their AI tool performance against aggregated results from other sites using AI for identical use cases, the ACR said.
Assess-AI currently supports multiple imaging AI use cases including intracranial hemorrhage, pulmonary embolism, pneumothorax, large vessel occlusion, bone age, breast density, and pleural effusion, among others.
![A normal mammogram confirmed by three-year radiologic follow-up illustrates reader-marked regions of interest (ROIs) during (A) unaided (round 1) and (B) artificial intelligence (AI)–assisted (round 2) reading. Each colored dot represents an ROI for recall by a human reader. Readers could mark more than one ROI per case, represented by multiple dots of the same color. During AI-assisted reading, the AI system displayed three visible prompts: two with suspicion of malignancy scores of 35% (left mediolateral oblique [L MLO] and craniocaudal [L CC]) and one with a suspicion of malignancy score of 10% (right craniocaudal [R CC]), shown as polygonal overlays. Without AI, six of 10 readers (60%) marked a false-positive ROI. With AI assistance, this fell to two of 10 (20%). R MLO = right mediolateral oblique.](https://img.auntminnie.com/mindful/smg/workspaces/default/uploads/2026/07/2026-07-14-radiology-mammogram-ai-auto-bias.H0bYO8QlWs.jpg?auto=format%2Ccompress&dpr=2&fit=crop&h=167&q=70&w=250)
