DIGITIMES observed that AI medical devices approved by the Taiwan Food and Drug Administration (TFDA) are predominantly applied in radiological imaging, with deep learning as the core technology. These products are mainly positioned as diagnostic support tools.
Taiwanese-made AI medical devices are primarily developed by startups and remain in early development stages, with their numbers significantly fewer than imported products. To address three major challenges – AI medical implementation, certification, and reimbursement – the Ministry established three major AI centers for smart healthcare in Taiwan.
The goal is to accelerate AI medical industrialization through institutional frameworks and hospital data sharing.
DL image interpretation now Taiwan's major AI medical application
Rule-based systems rely on expertsfor medication and order checks
Statistical models require large samples and can assess disease risks
ML suitable for risk prediction and disease classification, learning from big data
DL processes unstructured data, ideal for image and sound anomaly interpretation
Generative AI focuses on chatbot consultations and medical record assistance
Chart 1: Medical applications, development features, and limitations of AI tech evolution
Chart 2: Number of TFDA-approved AI/ML medical devices and specialty distribution
Cardiovascular, chest imaging and bone analysis are key areas for MIT devices
MIT AI medical devices mostly from startups, with hospitals providing verification
Chart 4: Number of players of MIT AI/ML medical devices approved by TFDA by type
Taiwan's major smart healthcare AI centers help address challenges
