Core Laboratory Comparison of Angiography-Derived Indexes of Microvascular Resistance vs Pressure-Wire IMR and Absolute Resistance.

BACKGROUND

Coronary microvascular dysfunction should be assessed in patients with angina and nonobstructive coronary arteries. However, established indexes, such as the pressure wire-derived index of microvascular resistance (PW-IMR) and continuous thermodilution-derived absolute hyperemic microvascular resistance (Rmicro) remain underused, driving development of the wire-free angiography-derived index of microvascular resistance (angio-IMR).

OBJECTIVES

The aim of this study was to conduct a core laboratory comparison of 5 angio-IMR methods.

METHODS

Five angio-IMR methods derived from 3 software platforms (QAngio XA 3D, FlashAngio, and AngioPlus Core) were compared against PW-IMR in 270 patients (274 vessels) and Rmicro in 42 patients (109 vessels) in a blinded fashion. Diagnostic performance was assessed on a per-vessel basis using receiver-operating characteristic (ROC) curves, Bland-Altman plots, and confusion matrices.

RESULTS

The median PW-IMR was 19.02 (Q1-Q3: 12.79-28.08), with 81 of 274 vessels (29.6%) having PW-IMR ≥25. The median Rmicro was 387.50 Wood units (WU) (Q1-Q3: 312.00-501.00 WU), with 33 of 109 vessels (30.3%) having Rmicro ≥475 WU. The correlations among angio-IMR, PW-IMR, and Rmicro were poor across all methods. Three of the 5 methodologies overestimated PW-IMR (mean bias -20.72 to -26.36), with all exhibiting large random errors. ROC analysis using angio-IMR for discriminating PW-IMR ≥25 showed areas under the ROC curve ranging from 0.530 to 0.576 (P > 0.05 for all), while 3 methods (methods A, B, and D) showed significantly better areas under the ROC curve to discriminate Rmicro ≥ 475 WU. When using the prespecified angio-IMR ≥25 threshold, only method E demonstrated acceptable accuracies of 61.4% and 62.0% to identify PW-IMR ≥25 and Rmicro ≥475 WU, respectively.

CONCLUSIONS

All 5 angio-IMR methods demonstrated poor diagnostic accuracy compared with PW-IMR and Rmicro, highlighting significant limitations in their clinical utility.