Noninvasive computational FFR predicts haemodynamic effects of stenting

A novel noninvasive fractional flow reserve (FFR) calculation method uses data from computed tomography coronary angiography (CTCA) to accurately predict the haemodynamic effects of stenting, according to a recent Singapore study.

“Using noninvasive FFR prediction, we can determine the importance of each lesion’s role in the haemodynamics of tandem lesions and decide the optimal treatment strategy for tandem lesions,” the researchers said, pointing to the potential value of such an approach in treatment planning.

The study included 25 patients who had undergone CTCA and were scheduled for invasive coronary angiography and FFR-guided percutaneous coronary intervention (PCI). Noninvasive FFR calculation was based on CTCA images with reduced-order computational fluid dynamics (CFD) and resistance boundary conditions. Its ability to predict coronary haemodynamics before and after stenting was compared against the gold standard invasive FFR.

Before stenting, the mean invasive FFR in patients was 0.67±0.13, while noninvasive FFR was 0.69±0.12. This corresponded to a nonsignificant mean difference of 0.024 (95 percent limit of agreement (LoA), –0.11 to 0.15). [Front Bioeng Biotechnol 2021;9:39667]

Similarly, the mean invasive and noninvasive FFRs after stenting were not significantly different from each other (0.84±0.10 vs 0.86±0.08, respectively; mean difference, –0.025, 95 percent LoA, –0.21 to 0.16).

Pearson Correlation analysis and Bland-Altman plots further validated good agreement between the invasive and noninvasive FFR measurement methods, both before (r, 0.86; p<0.001) and after (r, 0.50; p<0.001) stenting. Moreover, CFD simulations were successfully completed in all patients, each taking from 0.4–1.5 hours long.

For predicting residual ischaemia after stenting, noninvasive FFR had a diagnostic accuracy of 87 percent, with sensitivity and specificity of 80 percent and 88 percent, respectively. In addition, the positive and negative predictive values were 57 percent and 96 percent, respectively.

Importantly, the noninvasive FFR method for predicting outcomes after stenting showed high reproducibility between the two experienced observers who conducted the analyses. The interobserver intraclass coefficient was 0.976, while the coefficient of variation was 2.34 percent.

“Computationally simulated FFR is highly reproducible, readily obtainable from standard CTCA images with reduced-order CFD and novel boundary conditions,” the researchers said. “This finding is promising for noninvasive detection of hemodynamically significant coronary stenosis and holds potential as a virtual coronary stent implantation planning tool to predict the hemodynamic effects of stenting in CAD patients.”

Important limitations of the study included its small study sample and the predominance of Asian patients, which may limit its generalizability and applicability in the clinics.

“A larger cohort size will be needed to corroborate these initial findings. Thus, we propose these findings be considered as a ‘proof-of-concept’,” the researchers said.

“Another clinical trial on a much larger cohort of patients undergoing PCI to further validate the current methodologies is prerequisite before applying this tool in routine clinical practice,” they added.