The Future of AI-Assisted Medical Devices in Precision Medicine: A Systematic Review

Abstract

Background: The integration of Artificial Intelligence (AI) into medical devices has accelerated exponentially between 2020 and 2025, fundamentally altering the landscape of diagnostic medicine. This period is defined by the transition from theoretical algorithms to regulatory-approved, clinically deployed Software as a Medical Device (SaMD), particularly in image-centric specialties.
Objectives: This systematic review aims to (1) quantify and characterize regulatory trends for AI medical devices (AIMDs) in the US and EU; (2) evaluate the clinical efficacy and workflow impact of AI technologies in Ophthalmology, Oncology, and Musculoskeletal (MSK) disorders, with a specific focus on AI-assisted Point-of-Care Ultrasound (POCUS); and (3) assess the role of these technologies in democratizing access to expert-level diagnostics.
Methods: A PRISMA 2020–compliant literature search was conducted across PubMed/MEDLINE, Embase, Cochrane Library, and IEEE Xplore for peer-reviewed studies published between January 1, 2020, and December 31, 2025. Grey literature from FDA and EU regulatory databases was included to capture approval trends. Risk of bias was assessed using QUADAS-AI and ROBIS tools.
Results: The search identified 1,240 records; 67 pivotal studies and systematic reviews were included. Regulatory data reveal >1,000 FDA-authorized AI devices by 2025, with radiology and ophthalmology dominating. In Ophthalmology, autonomous AI for diabetic retinopathy and glaucoma has demonstrated sensitivity comparable to retina specialists (>90%), enabling widespread tele-screening. In Oncology, AI-assisted breast and prostate ultrasound has significantly improved novice diagnostic accuracy (AUC gains >0.10) and reduced unnecessary biopsies through enhanced specificity. In MSK, AI models for fracture detection and real-time POCUS guidance for nerve blocks have standardized procedure quality and reduced inter-operator variability.
Conclusions: AI medical devices have shifted from "assistive" to "autonomous" and "augmentative" roles, effectively democratizing diagnostic capacity. High-quality evidence supports their deployment to bridge workforce gaps, though challenges regarding regulatory harmonization and algorithmic bias persist.