국내 의료 인공지능 기술의 선진입 의료기술 트랙 도입 현황과 특성 분석

Abstract

Objectives: The number of artificial intelligence (AI)-based medical devices approved in Korea has been increasing. However, these technologies have not yet been incorporated into the National Health Insurance system and are introduced into clinical practice via the Early Access Pathway, en compassing the “New Health Technology with Deferred Assessment” and “Innovative Health Tech nology” programs. This study analyzed the current status and characteristics of AI-based medical de vices utilizing this pathway to inform future healthcare policy development. Methods: A database was constructed using publicly available notifications and approval information issued by the Minis try of Health and Welfare and the Ministry of Food and Drug Safety. AI technologies approved un der the Early Access Pathway were identified and analyzed for device classification, clinical data re view, data modality, clinical function, time interval between approval and early-access designation, and institutional and regional distribution. Results: Thirty-two AI technologies were identified, all domestically developed. Most were Class II devices (56.3%), and 53.1% required clinical data review. Imaging data (68.8%) and diagnostic support functions (59.4%) predominated. The median time from device approval to early-access designation was 424.0 days (n=32), which decreased to 194.5 days (n=18) for approvals after the 2022 policy refinement. Adoption was concentrated in secondary medical institutions (67.2%) and the Seoul metropolitan area (54.7%). Conclusion: AI-based medi cal technologies remain in the early stages of clinical adoption, primarily as imaging-based diagnos tic support tools. Although policy streamlining has reduced time to clinical implementation, regional disparities persist. Targeted strategies to promote balanced development and adoption are essential for a sustainable AI policy framework.