Impact of Absolute Myocardial Blood Flow Quantification on the Diagnostic Performance of PET-Based Perfusion Scans Using 82Rubidium.

ABSTRACT

BACKGROUND: Guidelines propose the inclusion of quantitative measurements from 82Rubidium positron emission tomography (RbPET) to discriminate obstructive coronary artery disease (CAD). However, the effect on diagnostic accuracy is unknown. The aim was to investigate the optimal RbPET reading algorithm for improved identification of obstructive CAD. METHODS: Prospectively enrolled patients (N=400) underwent RbPET and invasive coronary angiography with fractional flow reserve and quantitative coronary angiography. Quantitative measurements (myocardial blood flow (MBF), MBF reserve, transient ischemic dilatation) by RbPET were step-wisely added to a qualitative assessment by the summed stress score based on their diagnostic accuracy of obstructive CAD by invasive coronary angiography-fractional flow reserve. Prespecified cutoffs were summed stress score ≥4, hyperemic MBF 2.00 mL/g per min, and MBF reserve 1.80, respectively. Hemodynamically obstructive CAD was defined as >90% diameter stenosis or invasive coronary angiography-fractional flow reserve ≤0.80, and sensitivity analyses included a clinically relevant reference of anatomically severe CAD (>70% diameter stenosis by invasive coronary angiography-quantitative coronary angiography). RESULTS: Hemodynamically obstructive CAD was present in 170/400 (42.5%) patients. Stand-alone summed stress score showed a sensitivity and specificity of 57% and 93%, respectively, while hyperemic MBF showed similar sensitivity (61%, P=0.57) but lower specificity (85%, P=0.008). With increased discrimination by receiver-operating characteristic curves (0.78 versus 0.85; P<0.001), combining summed stress score, MBF and MBF reserve showed the highest sensitivity of 77% but lower specificity of 74% (P<0.001 for both comparisons). Against anatomically severe CAD, all measures independently yielded high discrimination ≥0.90 with increased sensitivity and lower specificity by additional quantification. CONCLUSIONS: The inclusion of quantitative measurements to a RbPET read increases in the identification of obstructive CAD. REGISTRATION URL https//www.clinicaltrials.gov; Unique identifier NCT03481712.