Dies ist eine Übersichtsseite mit Metadaten zu dieser wissenschaftlichen Arbeit. Der vollständige Artikel ist beim Verlag verfügbar.
Regulatory landscape of applications of AI in measurement endeavors for medical diagnostics
0
Zitationen
1
Autoren
2025
Jahr
Abstract
Abstract Image-based medical diagnostics, particularly in pathology, presents unique challenges due to the complexity and variability of high-resolution medical images. Traditionally dependent on the subjective expertise of experienced clinicians, this field is undergoing a transformation through the integration of artificial intelligence (AI) and digital processing pipelines. To ensure quality-assured diagnostics beyond human-centered methods, a metrological framework is essential—one that accounts for diagnostic accuracy as well as the inherent difficulty of image interpretation. Recent advances highlight the relevance of entropy-based measures to quantify the complexity of interpreting ordinal and nominal image properties, offering a robust tool for assessing diagnostic uncertainty. Applications of this approach have emerged in AI-driven breast cancer diagnostics and in cognitive assessment tasks within the NeuroMET project, where entropy has been used to evaluate diagnostic performance and reliability. Despite these technological advances, regulatory and standardization challenges—such as compliance with the medical device regulation and the AI Act—continue to limit widespread adoption. On the other hand, emerging solutions and proposed frameworks—such as the International Data Spaces initiative—offer promising opportunities to address current challenges in the application of AI to medical diagnostics. This work reviews key regulatory developments, addressed also by activities of ISO/TC215 in AI standardization (Health informatics), highlighting both persistent barriers and areas requiring further refinement. In addition, the manuscript explores the intersection of digital pathology, AI, and modern metrology, emphasizing the need for unified metrics and standardized protocols to bridge subjective clinical expertise with automated diagnostic systems across domains including oncology and neurology.
Ähnliche Arbeiten
Biostatistical Analysis
1996 · 35.445 Zit.
UCI Machine Learning Repository
2007 · 24.290 Zit.
An introduction to ROC analysis
2005 · 20.596 Zit.
The use of the area under the ROC curve in the evaluation of machine learning algorithms
1997 · 7.102 Zit.
A method of comparing the areas under receiver operating characteristic curves derived from the same cases.
1983 · 7.061 Zit.