@article{ATM156062,
author = {Lucía Osoro and Baptiste Vasey and Peter McCulloch and Felix Broghammer and Stephen Gilbert and Juan Carlos Rejon-Parrilla and Dipak Kalra and Rubén Casado-Arroyo},
title = {Beyond the algorithm: health technology assessment frameworks for AI in cardiology under the European Union Health Technology Assessment Regulation: a systematic review},
journal = {Annals of Translational Medicine},
volume = {14},
number = {3},
year = {2026},
keywords = {},
abstract = {Background: Artificial intelligence (AI) is increasingly used in cardiovascular care to support diagnosis, monitoring and clinical decision-making. However, its dynamic and adaptive nature challenges conventional health technology assessment (HTA) frameworks, which are typically designed for static interventions. This review aims to assess how existing literature supports HTA-relevant evaluation of AI-based cardiovascular technologies and examine their alignment with the evidentiary requirements outlined in the European Union Health Technology Assessment Regulation (EU HTAR).Methods: A structured literature search was conducted in PubMed, Scopus, and ScienceDirect databases for studies published between January 2020 and December 2025. After screening 223 records, 33 full texts were reviewed and six met inclusion criteria. A narrative synthesis was performed, alongside a comparative analysis of three HTA frameworks.Results: Six studies were included in the final synthesis, covering cardiovascular AI applications such as stroke outcome prediction, atrial fibrillation screening and wearable-based monitoring. Supported by 17 documents. Four studies incorporated real-world data, though lifecycle adaptability and post-deployment evaluation were rarely addressed. Most focused on clinical or economic performance, without referencing formal HTA frameworks. Alignment with the EU HTAR was indirect and stakeholder engagement, particularly with cardiologists, was inconsistently reported. These findings indicate increasing clinical adoption of cardiovascular AI but limited integration with structured HTA processes or regulatory foresight.Conclusions: While AI tools in cardiology show increasing promise, current HTA practices do not yet fully align with the regulatory and methodological expectations of the EU HTAR. Adapted evaluation models are needed to support the effective, evidence-based adoption of AI technologies in cardiovascular care.},
issn = {2305-5847}, url = {https://atm.amegroups.org/article/view/156062}
}