Pharmaceutical News
TFDA announces the Guidelines on the Review and Approval of AI/ML-based CADe Software as Medical Device
2021/07/17

In order to ensure the safety, effectiveness and quality of artificial intelligence and machine learning (AI/ML) based computer aided detection (CADe) systems, the TFDA has referenced relevant regulations in the US, Japan and South Korea and published the guidelines on the review and approval of such products with examples for illustration.  This document is formulated based on the “Guidelines on the Review and Approval of AI/ML-based Medical Software”.  The TFDA expressed that the new guidelines should be able to help medical software developers to evaluate their products and prepare for the regulatory submission.

The TFDA pointed out that all AI/ML-based CADe software as medical device should comply with relevant regulations governing medical devices.  The product descriptions and the validation of safety and effectiveness mentioned in the Guidelines are derived from currently available references.  However, as the technology advances quickly, the TFDA reviewers may request the submission of additional safety and effectiveness data based on the features of the product.  The Guidelines will be updated from time to time, said the TFDA.

As for the scope of the application, the Guidelines describe the key points for considerations when reviewing the applications of Class 2 and Class 3 AI/ML-based CADe software, whether it is integrated into a medical device or is a stand-alone software as a medical device (SaMD).  Such products are designed to automatically identify suspicious lesions or abnormalities on medical images, and to use the AI/ML technology to mark its location, as well as to analyse medical images, data and/or other medical test results, such as the images of ultrasound, X-ray, magnetic resonance imaging (MRI), computed tomography (CT), positron emission tomography (PET), endoscopy, etc.

【2021-07-08 / TFDA News 】