Page 13 - ITPS-8-3

INNOSC Theranostics and
Pharmacological Sciences AI in medical device safety

contact while enabling timely diagnosis and treatment, A digital twin is a virtual replica of a medical device
thus enhancing patient safety. or system, including its interactions with patients and
Despite the promise of AI to improve patient outcomes, the environment, which is powered by real-time data
several issues and inconsistencies remain. A systematic review collection and AI-driven analysis. This virtual model
of 53 studies on AI in patient safety revealed variability in mirrors the physical device’s performance and the
AI reporting and the absence of standardized benchmarks. patient’s response, enabling continuous monitoring of
47
This variability highlights the need for thorough validation safety and performance. The digital twin can simulate
of AI systems in real-world clinical settings to ensure their various scenarios and predict potential risks or adverse
reliability in predicting safety outcomes. events before they occur, providing an advanced layer of
surveillance in materiovigilance systems. As AI enhances
To take everything into consideration, AI-powered the digital twin, it becomes increasingly sophisticated,
decision support tools have demonstrated their ability to learning from ongoing patient interactions and device
improve medication administration, patient stratification, usage, thus enabling proactive safety measures. This
and error detection, thereby bolstering patient safety. virtual model tracks the evolution of device behavior,
47
Both in hospital and home settings, the integration of AI assesses the long-term impact of AI-driven algorithms,
into patient monitoring systems has improved real-time and predicts device performance in diverse patient
monitoring, increased predictive accuracy, and accelerated populations. 51,52
response times. However, addressing concerns related
27
to data privacy, algorithm transparency, and integration 5. Ethical and regulatory considerations for
into clinical workflows requires further investigation and AI in materiovigilance
validation. As AI continues to evolve, it holds the potential
48
to revolutionize materiovigilance and significantly improve The use of AI-powered solutions in materiovigilance raises
patient outcomes. significant ethical and legal concerns. While AI applications
in this field offer increased accuracy and efficiency in
AI is transforming materiovigilance, particularly in detecting and recording adverse events, they also introduce
enhancing the safety and customization of medical devices. complex issues related to privacy, accountability, and
For individuals with physical disabilities, AI-optimized, equity. 53,54 The question of whether AI fits within existing
3D-printed assistive devices are improving satisfaction legal frameworks or whether a new category should be
and mobility, fostering greater independence. A prime established to address its unique features and implications
example is the FDA-approved Nevro HFX iQ system, an remains a subject of ongoing debate. The integration of
55
AI-powered spinal cord stimulator that personalizes pain AI in materiovigilance necessitates the development of
treatment by continuously adjusting neurostimulation robust governance frameworks to address these ethical
based on real-time patient feedback, thereby enhancing dilemmas and guide decision-making. These frameworks
pain management. In addition, the system’s adaptive should consider the distinct aspects of medical device
algorithms detect early signs of infection or complications, regulation while prioritizing accountability, transparency,
allowing timely interventions. This AI-driven approach in and privacy protection. 56
materiovigilance exemplifies how advanced monitoring
can significantly improve patient safety and outcomes. 49 Several regulations have been proposed in various
countries to address these concerns, such as the AI Act
AI-powered bionic limbs by Össur further illustrate and the Medical Device Regulation in the European
the impact of AI in materiovigilance, particularly for Union. The AI Act represents the European Union’s first
optimizing prosthetic limbs. These advanced prosthetics comprehensive regulation of AI, classifying AI systems
use ML to adapt in real time to each user’s gait and according to their risk levels. It establishes stringent safety
movements, enhancing comfort, stability, and balance, even and ethical standards for AI applications deemed to be
on challenging terrains. By continuously adjusting to the of higher risk. This regulation came into effect in August
user’s movements, they provide a more natural experience, 2024 across the 27 European Union member states, with
improve mobility, and reduce the risk of falls, thereby full enforcement scheduled for completion by August
significantly enhancing patient safety and satisfaction. 2027. Similarly, the Medical Device Regulation, which
57
This example underscores AI’s role in personalizing has been in effect since May 2021, governs the safety
device performance to meet individual needs, positively and efficacy of medical devices within the European
impacting patient outcomes. 50 Union. It requires manufacturers to comply with rigorous
Another emerging approach to monitoring standards for design, clinical assessment, and post-market
medical device safety is the concept of the digital twin. surveillance. Together, these frameworks ensure the secure

Volume 8 Issue 3 (2025) 7 doi: 10.36922/itps.6204