Health ministry selects six consortia for AI medical devices

The health ministry said it selected six AX-Sprint consortia to speed the clinical adoption of AI digital medical devices.

According to the Korea Health Industry Development Institute, the initiative was launched after the government found that support for building the clinical evidence and real-world data needed to commercialize AI medical products was inadequate.

Seong Chang-hyeon, director general of the health ministry's health industry policy bureau, said, "The program will be a key catalyst that helps AI digital medical devices move beyond technology development and contribute to public health in actual clinics." Jeong Yeong-hun, planning director at the Korea Health Industry Development Institute, said the government, developers and medical institutions "must work closely together to create tangible clinical results and industrial outcomes."

The program supports partnerships between developers and medical institutions for clinical validation, real-world data collection and marketing. Applicants were limited to digital medical devices that had received Ministry of Food and Drug Safety clearance by April 24, while products still undergoing approval were excluded.

One brain-disease project will deploy AI that analyzes brain perfusion and cerebral infarctions at neurology departments in 22 hospitals to test its effects on acute-stroke outcomes and its economic value. Another will validate MRI-based AI that assists Parkinson's diagnoses at three hospitals, examining diagnostic accuracy and whether it reduces unnecessary PET scans.

Two cancer-diagnosis projects will apply chest X-ray and breast-cancer AI at 10 hospitals, collecting 250,000 real-world cases as they pursue health-insurance reimbursement. A separate consortium will introduce real-time gastric-endoscopy analysis at seven major hospitals to test diagnostic accuracy for precancerous gastric lesions and reduce unnecessary biopsies.

Cardiovascular projects will validate AI that analyzes biological signals to predict arrhythmia risks at 15 university hospitals. A second project will evaluate the safety and effectiveness of retinal-image-based AI for cardiovascular risk assessments and plans to secure a reimbursement fee for evaluating chronic-disease patients and patients before surgery.