Why the study?
Current diagnostic procedures for CAD are invasive, costly, and risky, while CFD-based approaches have extensive computational requirements that render them impractical for local hospital use. This review aims to understand the current state and limitations of AI-based methodologies for non-invasive FFR computation.
Do AI-based methodologies provide accurate and fast non-invasive fractional flow reserve (FFR) predictions in patients with coronary artery disease?
Do AI-based methodologies provide accurate and fast non-invasive fractional flow reserve (FFR) predictions in patients with coronary artery disease?
Answers are grounded in this paper and cite their sources.
AI-based approaches offer a rapid and accurate non-invasive alternative to invasive FFR and computationally heavy CFD methods for evaluating coronary artery stenosis, pending solutions to data quality and interpretability challenges.
PURPOSE: Given the global prevalence of coronary artery disease (CAD) and atherosclerosis, there is a critical need for accurate non-invasive diagnostic approaches. The current diagnostic procedures in the hospital are invasive, costly, and may pose potential risks to the patients. While computational fluid dynamics (CFD) based diagnostic approaches can have high accuracy, their extensive computational requirements render them impractical for local hospital use. Artificial Intelligence (AI)-based methodologies for predicting fractional flow reserve (FFR) in coronary arteries with stenosis have gained attention for their nearly instantaneous and accurate results, holding significant potential for impacting healthcare outcomes. The objective of this study is to present the current uses of AI for non-invasive FFR computation and understand the current state of this technology, address current limitations and suggest potential directions for future research. METHODS: This work provides a comprehensive review of AI-based approaches for the non-invasive measurement of the FFR. This research emphasizes key factors including the different possible algorithms, their inputs, the sample of patients, and statistical indexes such as accuracy, sensitivity and specificity of the predictions. RESULTS: The reviewed literature shows that AI methodologies lead to low computational times and high accuracy values and, therefore, are promising alternatives for accurate FFR prediction in atherosclerotic coronary arteries. The accuracy of the results was highly linked to the inputs used in the algorithms, and there is no consensus as to the most relevant inputs. Despite the evident advantages, the widespread implementation of ML approaches in hospitals still faces challenges such as lack of data, image quality, and result interpretability and explainability. CONCLUSION: AI emerges as a promising tool for non-invasive FFR computation, leading to high accuracy in CAD assessment. Addressing the identified challenges is crucial for enhancing diagnostic capabilities and patient care for CAD in hospitals.
No takes yet. Share an insight, caveat, or question.
Fernandes et al. (2026) conducted a review in Coronary artery disease (CAD) and atherosclerosis. Artificial Intelligence (AI)-based methodologies for predicting fractional flow reserve (FFR) vs. Invasive fractional flow reserve (FFR) or computational fluid dynamics (CFD)-based diagnostic approaches was evaluated on Diagnostic accuracy, sensitivity, specificity, and computational time of non-invasive FFR prediction.
Synapse has enriched 10 related papers. Consider them for comparative context: