Review in coronary functional evaluation methods: applications of computational fluid dynamics and artificial intelligence in the context of coronary artery disease

Q: What are the key findings of this study?

Computational fluid dynamics and artificial intelligence methodologies enable rapid, non-invasive estimation of coronary functional indices with high diagnostic accuracy compared to invasive FFR.

Q: What is the clinical evidence from this study?

Study design: Systematic Review. Population: Coronary artery disease. Intervention: Non-invasive coronary functional evaluation using computational fluid dynamics (CFD), fluid-solid interaction (FSI), and artificial intelligence/machine learning vs. Invasive coronary functional indices (e.g., fractional flow reserve [FFR], coronary flow reserve [CFR]) and anatomical stenosis assessment. Primary outcome: Diagnostic accuracy for detecting hemodynamically significant coronary ischemia.

Q: What does this research mean for the field?

Computational fluid dynamics and artificial intelligence methodologies enable rapid, non-invasive estimation of coronary functional indices with diagnostic accuracy comparable to invasive fractional flow reserve. Novelty: ClaimNovelty.SYNTHESIS. Consensus alignment: ConsensusAlignment.SUPPORTS_CONSENSUS.

Haojia Qu; Peiqi Lou; Yibin Pan; Debao Guan; Tianqi Wang; Hao Xu; Xiaorong Chen; Xinyang Ge

doi:10.3389/fphys.2026.1875617

October 8, 2026Frontiers in PhysiologyOpen Access

Review in coronary functional evaluation methods: applications of computational fluid dynamics and artificial intelligence in the context of coronary artery disease

View Full Paper

Ask AI

Bookmark

Share

Key result

AI and computational modeling accurately estimate coronary functional indices non-invasively compared to invasive FFR.

Why the study?

Anatomic stenosis severity frequently dissociates from the hemodynamic significance of a lesion, offering limited guidance for optimal revascularization. Synthesizing invasive and emerging computational and artificial intelligence methodologies is needed to elucidate the appropriate clinical application of each functional index.

Do computational fluid dynamics and artificial intelligence-based non-invasive coronary functional indices accurately detect hemodynamically significant ischemia compared with invasive physiological measurements in patients with coronary artery disease?

Key Points

To systematically review the clinical utility of coronary functional indices and examine emerging non-invasive diagnostic strategies that integrate computational fluid dynamics and artificial intelligence.
Conducted a systematic literature search across PubMed, Google Scholar, Web of Science, and Baidu Scholar from 1973 to 2026 in English and Chinese.
Included retrospective and prospective studies evaluating the diagnostic accuracy of invasive and non-invasive coronary functional indices; excluded case reports, comments, conference abstracts, and studies lacking complete diagnostic data.
Anatomical assessment frequently dissociates from hemodynamic significance, with evidence showing fractional flow reserve (FFR) guidance reduced stent placement compared with angiography alone (1.9 ± 1.3 vs. 2.7 ± 1.2 stents per patient, P<0.001) and revealed 37% of anatomical lesions did not require intervention.
Incorporating FFR into diagnostic decision-making reclassified approximately half of treatment decisions initially determined by anatomical imaging alone.
Computational fluid dynamics and artificial intelligence models enable non-invasive FFR derivation from standard angiography and coronary computed tomography angiography, matching conventional invasive accuracy without requiring pressure wires or pharmacologic hyperemia.

Study Design

Type

Systematic Review

Structured PICO

Do computational fluid dynamics and artificial intelligence-based non-invasive coronary functional indices accurately detect hemodynamically significant ischemia compared with invasive physiological measurements in patients with coronary artery disease?

P

Population

Retrospective and prospective studies published between 1973 and 2026 evaluating the diagnostic accuracy of invasive and non-invasive coronary functional indices in patients with coronary artery disease.

I

Intervention

Non-invasive coronary functional assessment methodologies integrating medical imaging (CCTA, ICA, IVUS, OCT) with biomechanical modeling (3-D CFD, FSI, reduced-order 0-D/1-D models) and artificial intelligence (machine learning and deep learning).

C

Comparator

Invasive physiological assessment (primarily pressure wire-derived fractional flow reserve [FFR] and coronary flow reserve [CFR]) and traditional anatomical stenosis severity assessment.

O

Outcome

Diagnostic accuracy (sensitivity, specificity, overall accuracy, AUC) and computational efficiency for detecting hemodynamically significant coronary stenosis/ischemia (typically defined against invasive FFR ≤0.80).surrogate

This systematic review synthesizes the evolution from invasive coronary physiology to CFD- and AI-driven non-invasive functional indices, highlighting their potential to provide rapid, accurate ischemia evaluation while identifying key methodological barriers to routine clinical adoption.

Limitations

Substantial methodological heterogeneity across included studies in imaging modality, model dimensionality, boundary condition specification, reference standard, sample size, and validation strategy
Inconsistent reporting of performance metrics across studies, rendering direct numerical comparisons challenging
Many computational studies rely on simplified or population-averaged boundary conditions rather than patient-specific in vivo physiological parameters
Several machine learning and imaging-based studies are limited by retrospective, single-center designs, modest sample sizes, or lack of independent external validation
Substantial heterogeneity across included studies in imaging modality, model dimensionality, boundary condition specification, reference standard, sample size, and validation strategy
Inconsistent reporting of diagnostic performance metrics across studies prevents direct numerical comparisons
Reliance on simplified or population-averaged boundary conditions in many CFD models rather than patient-specific physiological parameters
High computational cost and numerical instability in 3-D fluid-solid interaction (FSI) models limit real-time clinical application
Many machine learning and deep learning studies are constrained by retrospective designs, modest sample sizes, synthetic training data, or lack of external multi-center validation

Ask about this paper

Answers are grounded in this paper and cite their sources.

Implication

Supports non-invasive functional assessment in CAD; extends invasive FFR validation to CFD and AI methods.

Abstract

Coronary artery disease, driven by the progressive accumulation of atherosclerotic plaque within the epicardial coronary arteries, remains a leading cause of myocardial ischemia and global morbidity. A fundamental challenge in the clinical management of the disease has been the reliance on anatomic stenosis severity alone, a metric that frequently dissociates from the hemodynamic significance of a lesion and thus offers limited guidance for optimal revascularization. The advent of invasive measurements marked a transformative paradigm shift, enabling the development of a diverse functional indices for the evaluation of coronary physiology. Recently, the development of computational science and artificial intelligence has accelerated the translation of these physiological assessments into the non-invasive domain, primarily through high-fidelity numerical simulations and efficient machine learning algorithms. This review systematically examines the evolution of coronary functional assessment across several pivotal frontiers. It begins by contextualizing the diagnostic landscape prior to the widespread adoption of functional indices, where clinical decisions were predominantly informed by imaging-derived anatomical parameters. Subsequently, it traces the historical evolution of invasive coronary functional indices, critically evaluating the comparative effectiveness of pressure-derived metrics against flow-derived metrics in the detection of hemodynamically significant ischemia. Furthermore, a critical evaluation of emerging artificial intelligence methodologies for the non-invasive evaluation of functional significance were employed. By synthesizing these technological methodologies, this review aims to elucidate the appropriate clinical application of each functional index, thereby providing a road-map for clinicians to achieve precise, individualized ischemia evaluation and guide therapeutic decision-making in the contemporary era of precision cardiovascular medicine.

Discussion

No takes yet. Share an insight, caveat, or question.

Cite This Study

Qu et al. (2026) conducted a systematic review in Coronary artery disease. Non-invasive coronary functional evaluation using computational fluid dynamics (CFD), fluid-solid interaction (FSI), and artificial intelligence/machine learning vs. Invasive coronary functional indices (e.g., fractional flow reserve [FFR], coronary flow reserve [CFR]) and anatomical stenosis assessment was evaluated on Diagnostic accuracy for detecting hemodynamically significant coronary ischemia. Computational fluid dynamics and artificial intelligence methodologies enable rapid, non-invasive estimation of coronary functional indices with high diagnostic accuracy compared to invasive FFR.

synapsesocial.com/papers/6aee6b6871a75948062208d5 https://doi.org/10.3389/fphys.2026.1875617

View Full Paper

Ask AI

Bookmark

Share

Review in coronary functional evaluation methods: applications of computational fluid dynamics and artificial intelligence in the context of coronary artery disease

Key Points

Study Design

Structured PICO

Limitations

Ask about this paper

Implication

Abstract

Discussion

Cite This Study

Review in coronary functional evaluation methods: applications of computational fluid dynamics and artificial intelligence in the context of coronary artery disease

Key Points

Study Design

Structured PICO

Limitations

Ask about this paper

Implication

Abstract

Discussion

Cite This Study

Keep reading

Keep reading