# Fractal analysis of dynamic stress myocardial CT perfusion decouples diagnostic accuracy for obstructive coronary artery disease from remote flow

## TL;DR

Fractal analysis improves detection of heart artery blockages in patients with poor blood flow, outperforming traditional methods.

## Contribution

Fractal analysis decouples diagnostic accuracy for obstructive CAD from remote flow in dynamic stress CTP.

## Key findings

- Fractal analysis outperformed relative MBF in detecting obstructive CAD (94% sensitivity, 92% specificity).
- Fractal analysis showed higher accuracy in intermediate and low remote flow subgroups compared to relative MBF.
- Microvascular ischemia prevalence increased with lower remote flow (9%, 37%, 70%).

## Abstract

In patients with reduced remote flow, dynamic stress CT perfusion imaging (CTP) has limited utility for detecting obstructive coronary artery disease (CAD). We compared fractal analysis, a descriptor of macro- and microvascular ischemia patterns, with relative myocardial blood flow (MBF) for detecting obstructive CAD stratified by remote flow.

This secondary analysis of the prospective multi-center AMPLIFiED trial included patients who underwent invasive coronary angiography (ICA) with fractional flow reserve (FFR) and dynamic stress CTP. Obstructive CAD was defined invasively (stenosis ≥ 90% or FFR < 0.8). We assessed diagnostic accuracy of fractal analysis and relative MBF in patient groups subdivided into high, intermediate, and low remote flow.

In 148 patients (30% female; 416 vessels), obstructive CAD was present in 71/148 patients (48%), while signs of microvascular ischemia were found in 26/148 patients (prevalence in high, intermediate, low remote flow: 9%, 37%, 70%). Fractal analysis outperformed relative MBF in detecting obstructive CAD (patient-based sensitivity and specificity: 94% and 92% versus 80% and 64%, p < 0.005). In vessel-based subgroup analysis, both methods performed comparably in high remote flow (area under the receiver-operating curve [AUC]: 0.94 versus 0.88, p = 0.45), while fractal analysis outperformed relative MBF in intermediate and low remote flow (AUC: 0.93 versus 0.77, p = 0.004; 0.92 versus 0.6, p < 0.001).

Fractal analysis eliminates reliance on remote flow for CTP imaging, which improves diagnostic accuracy for obstructive CAD particularly in patients with reduced remote flow or microvascular ischemia.

This study reports about data from the prospective, multi-center AMPLIFiED trial (UMIN000016353).