Computer systems can learn and adapt by implementing human intelligence, which is known as
artificial intelligence (AI). When it used in combination with medical software, it creates an artificial
intelligence (AI)-based medical device that uses in data and algorithms to help with activities like
diagnosis and treatment suggestions, potentially increasing the accuracy and effectiveness of medical
care, and the use of Artificial Intelligence/Machine Learning (AI/ML) into medical devices, especially
Software as a Medical Device (SaMD), this enclosed important concepts like AI/ML, medical devices,
SaMD, 510(k) notifications, FDA, CDRH, PMA, TPLC, IMDRF, SPS, and ACP. Through launch the
issues raised by AI/ML in SaMD, the structure required to strike a compromise between patient safety
and innovation. The subject is to establish the SaMD Pre-Specifications (SPS) and Algorithm Change
Protocol (ACP) as crucial elements for open changes, ongoing monitoring, and post-market
surveillance. Primary goal is to extend the framework by providing guidance for change control plans
using AI/ML methods, particularly those with software learning over time. The Premarket Approval (PMA) pathway and the 510(k) procedure were covered, taking into account the Quality System
Regulation and the IMDRF Quality Management System. Clinical evaluation ensures safety &
effectiveness of medical devices. The approach also emphasized the need of using real-world data,
sticking to an iterative update cycle, and adopting Good Machine Learning Practices. Basically, the
suggested regulatory framework aimed to offer a structured strategy to deal with the difficulties of
AI/ML-based SaMD alterations within the regulatory landscape while encouraging innovation and
ensuring patient safety.
Keywords: AI / ML, Clinical Evaluation, Medical device, SaMD, 510k notification, FDA,
CDRH, PMA, TPLC, IMDRF, SPS, ACP
Publication date: 01/08/2025
https://ijbpas.com/pdf/2025/August/MS_IJBPAS_2025_9309.pdf
Download PDF
https://doi.org/10.31032/IJBPAS/2025/14.8.9309
