Validation of physiological principles of Non-Invasive fractional flow reserve derived from CT coronary angiography

Jasmine Chan; Abdul R Ihdayhid; Harsh V Thakkar; Michael Michail; Michael Leung; Udit Thakur; Andrea Comella; Sean Tan; James D Cameron; Nitesh Nerlekar; Sujith Seneviratne; Stephen Nicholls; Brian Ko; Habib Samady; Adam J Brown

doi:10.1016/j.jcct.2025.05.240

Validation of physiological principles of Non-Invasive fractional flow reserve derived from CT coronary angiography

J Cardiovasc Comput Tomogr. 2025 Jun 19:S1934-5925(25)00344-2. doi: 10.1016/j.jcct.2025.05.240. Online ahead of print.

Authors

Affiliations

¹ Victorian Heart Institute and Victorian Heart Hospital, Monash University, Monash Health, 631 Blackburn Road, Clayton, Victoria, 3168, Australia.
² Harry Perkins Institute of Medical Research, PO Box 7214, Shenton Park, 6008, Perth, WA, Australia.
³ Victorian Heart Institute and Victorian Heart Hospital, Monash University, Monash Health, 631 Blackburn Road, Clayton, Victoria, 3168, Australia; Institute of Cardiovascular Science, University College London, Gower Street, London WC1E 6BT, United Kingdom.
⁴ Northeast Georgia Medical Centre, 743 Spring Street NE, Gainesville, GA, 30501, USA.
⁵ Victorian Heart Institute and Victorian Heart Hospital, Monash University, Monash Health, 631 Blackburn Road, Clayton, Victoria, 3168, Australia. Electronic address: adam.brown@monash.edu.

PMID: 40541555
DOI: 10.1016/j.jcct.2025.05.240

Abstract

Background: Coronary computed tomography angiography (CTA) simulation of fractional flow reserve (FFR) is derived from allometric and morphometric scaling principles, allowing inference of physiological parameters from anatomical measures like left-ventricular (LV) mass and coronary luminal dimensions. Validity of these assumptions in humans remains uncertain, with supporting data derived from animal models.

Methods: Twenty-two patients with non-obstructive coronary artery disease underwent proximal-LAD intravascular ultrasound (IVUS) and Combowire assessment at rest and hyperemia. Coronary volumetric flow (Q, cm³/sec) was derived from average baseline peak-velocity (cm/sec) x IVUS cross-sectional-area (cm²). Baseline microvascular resistance (BMVR, mmHg/cm²) was calculated: distal coronary pressure (mmHg) - right atrial pressure (mmHg) divided by Q. Patients underwent same-day CTA to provide quantitative measures including LV mass (g), cumulative coronary luminal volume (mm³) and vessel length (mm). Relationships between quantitative CTA-derived metrics and invasive physiology were evaluated using Pearson's correlation.

Results: Mean FFR was 0.94 ± 0.06; median coronary flow reserve velocity was 2.54 [IQR 2.1-3.1]. Baseline Q and BMVR were 2.30 ± 1.0 cm³/s and 44.6 ± 21.6 mmHg/cm², respectively. Average LV-mass was 148.6 ± 30.6g, coronary luminal volume 1038.6 ± 485.2 mm³ and vessel length 184.5 ± 66.8 mm. LV mass correlated strongest with coronary flow (r = 0.87, p < 0.001) followed by vessel length (r = 0.75, p < 0.0001) and coronary luminal volume (r = 0.73, p < 0.001). The scaling coefficient (1.87) significantly differed from experimental data. CT-derived metrics demonstrated strong negative correlation with BMVR (LV mass -0.70, coronary luminal volume -0.70, vessel length -0.76; P < 0.0001 respectively).

Conclusion: These findings support deriving coronary flow and microvascular resistance from CTA anatomical metrics. Revised scaling coefficients and hyperemic modelling could enhance CTA-derived FFR diagnostic performance.

Keywords: CT-Derived fractional flow reserve (CT-FFR); Coronary computed tomography angiography (CTA); Coronary volumetric flow; Fractional flow reserve (FFR); Microvascular resistance.