a2z Radiology AI Receives EU MDR Certification for AI That Detects 182 Findings Across Head, Chest, Abdomen and Pelvis CT
a2z Radiology AI Receives EU MDR Certification for AI That Detects 182 Findings Across Head, Chest, Abdomen and Pelvis CT
a2z-Drafter is the first AI product certified under the EU Medical Device Regulation to support both detection and report drafting across head, chest, abdomen and pelvis CT.
BOSTON--(BUSINESS WIRE)--a2z Radiology AI Inc. today announced that a2z-Drafter has been certified under the EU Medical Device Regulation (MDR). The software supports triage, detection and report drafting on CT scans of the head, chest, abdomen and/or pelvis.
“Radiology teams have had to piece together separate AI tools for each condition and body region,” said Samir Rajpurkar, co-founder and CEO of a2z Radiology AI. “a2z-Drafter is one product that works on head, chest and abdomen-pelvis CT scans and gives the radiologist a structured starting point for the report. This certification lets us bring that to radiology teams in the European Union.”
a2z-Drafter detects 182 validated radiological findings on CT studies of the head, chest, abdomen and/or pelvis, with supplementary attributes for selected findings. The findings span acute conditions such as subarachnoid hemorrhage, midline shift, pneumothorax and bowel obstruction, findings that need prompt workup such as obstructing ureteral stones and aortic aneurysms, and findings for routine follow-up. The software provides structured output to assist with triage and prioritization of studies for review and with report drafting. Each finding carries a priority category that a provider’s worklist can use to help prioritize studies, and a provider’s reporting system can use the list of findings and attributes to prepare draft report content for the radiologist to review.
The software is designed to be integrated into existing PACS, RIS and worklist systems. The interpreting physician independently reviews the complete study and remains responsible for the diagnosis and clinical decisions. The software analyzes CT images acquired with or without contrast in adults aged 18 and older.
“A single CT scan can reveal an acute finding that needs urgent attention alongside several that require routine follow-up,” said Jeff Brown, MD, a radiologist and chief medical officer of a2z Radiology AI. “a2z-Drafter assigns each detected finding a priority category within the radiologist’s existing workflow. The radiologist still reviews the entire study and is responsible for the final interpretation.”
Pranav Rajpurkar, co-founder of a2z Radiology AI and Associate Professor of Biomedical Informatics at Harvard Medical School, placed the certification in the context of the field’s research. “Generalist AI, one product that covers many findings across the body, has been a research goal for years,” he said. “With this certification, it is a regulated clinical product. Each finding was evaluated on its own against acceptance criteria set in advance, across the head, chest, abdomen and pelvis.”
a2z-Drafter is certified as a Class IIb medical device under the EU Medical Device Regulation by notified body Scarlet NB B.V. (NB 3022). Radiology providers in the European Union can contact a2z at info@a2zradiology.ai to learn more. In the United States, a2z-Drafter is not FDA-cleared and is not commercially available.
About a2z Radiology AI
a2z Radiology AI builds AI to support care teams in medical image interpretation. Its triage software for abdomen-pelvis CT is FDA-cleared in the United States, and a2z-Drafter is certified under the EU Medical Device Regulation. The name a2z reflects its founding ambition to support comprehensive medical image interpretation, from A to Z.
The company was co-founded by Pranav Rajpurkar, an Associate Professor of Biomedical Informatics at Harvard Medical School, and Samir Rajpurkar, a technology entrepreneur. a2z is backed by Khosla Ventures and SeaX Ventures.
For more information, visit a2zradiology.ai.
Contacts
Media contact: a2z Radiology AI, info@a2zradiology.ai
