The NMPA is now seeking comments to the draft guidance “Artificial Intelligence Medical Software Classification ”, due by May 3, 2021.
AI technologies have the potential to transform health care by deriving new and important insights from the vast amount of data. They require an updated paradigm of medical device regulation to ensure that the safety and effectiveness of the software are maintained. This guideline, which has been drafted based on latest development status, considerations from international regulatory agencies as well as the risks and technical characteristics of AI algorithms, will provide meaningful guidance to the whole industry and regulators.
According to the guidance, there are roughly four application directions in medical-related fields based on products’ intended use:
Decision support software
Software that uses artificial intelligence algorithms to assist medical decision-making, such as software for lesion identification and drug calculation, should be managed as medical devices.
Image/data processing software
Software that uses artificial intelligence algorithms for medical imaging and data processing, such as image segmentation and fusion, data calculation, etc., should be managed as medical devices.
Analysis and mining products
Software which uses artificial intelligence algorithms to analyze and mine medical-related data for drug development, medical research, hospital non-medical information management, etc., and if its intended use does not have the medical purpose described in the definition of medical device, it will not be managed as a medical device.
Medical assistant products
Based on electronic medical records, software that is only used for non-medical device purposes such as hospital guidance and health consultation will not be considered as a medical device.
The application of artificial intelligence technology in the medical field is at the beginning stage. The risk of AI in clinical use has not yet been fully and in-depth evaluated. Therefore, in principle, Software as Medical Device (SaMD) that utilize AI technologies will be managed as the Class III, the highest risk device in China. In the future, based on the risk of artificial intelligence medical software in clinical use and its technical evaluation, the management category can be adjusted under appropriate conditions.
Please click HERE for our technical review on AI-aided Software Guideline. The article was published on BioWorld, a Hong Kong-based biotech magazine.