Validity and Diagnostic Performance of Computing Fractional Flow Reserve From 2-Dimensional Coronary Angiography Images

Authors

Vahid Mohammadi
Massoud Ghasemi
Reza Rahmani
Maryam Mehrpooya
Hamidreza Babakhani
Akbar Shafiee
Mohammad Sadeghian

Publication Date

1-1-2023

Journal

The Texas Heart Institute Journal

DOI

10.14503/THIJ-20-7410

PMID

36720243

Publication Date(s)

January 2023

Language

English

PMCID

PMC9969768

PubMedCentral® Posted Date

1-31-2023

PubMedCentral® Full Text Version

Post-Print

Published Open-Access

yes

Keywords

Male, Humans, Female, Coronary Angiography, Fractional Flow Reserve, Myocardial, Retrospective Studies, Coronary Stenosis, Hemodynamics, Predictive Value of Tests, Coronary Vessels, Severity of Illness Index

Copyright

This work is licensed under a Creative Commons Attribution-NonCommercial-No Derivative Works 4.0 International License.

Abstract

BACKGROUND: Measurement of fractional flow reserve (FFR) is the gold standard for determining the physiologic significance of coronary artery stenosis, but newer software programs can calculate the FFR from 2-dimensional angiography images.

METHODS: A retrospective analysis was conducted using the records of patients with intermediate coronary stenoses who had undergone adenosine FFR (aFFR). To calculate the computed FFR, a software program used simulated coronary blood flow using computational geometry constructed using at least 2 patient-specific angiographic images. Two cardiologists reviewed the angiograms and determined the computational FFR independently. Intraobserver variability was measured using κ analysis and the intraclass correlation coefficient. The correlation coefficient and Bland-Altman plots were used to assess the agreement between the calculated FFR and the aFFR.

RESULTS: A total of 146 patients were included, with 95 men and 51 women, with a mean (SD) age of 61.1 (9.5) y. The mean (SD) aFFR was 0.847 (0.072), and 41 patients (27.0%) had an aFFR of 0.80 or less. There was a strong intraobserver correlation between the computational FFRs (r = 0.808; P < .001; κ = 0.806; P < .001). There was also a strong correlation between aFFR and computational FFR (r = 0.820; P < .001) and good agreement on the Bland-Altman plot. The computational FFR had a high sensitivity (95.1%) and specificity (90.1%) for detecting an aFFR of 0.80 or less.

CONCLUSION: A novel software program provides a feasible method of calculating FFR from coronary angiography images without resorting to pharmacologically induced hyperemia.