Incremental risk stratification with quantitative coronary vasodilator function assessment by PET in patients with known or suspected CAD

Objectives Incremental risk stratification with quantitative coronary vasodilator function assessment by PET in patients with known or suspected CAD.

Methods 2705 consecutive patients referred for rest/stress Rb-82 PET were followed for a median of 1.4 years (IQR: 0.6-2.5). Myocardial perfusion images (MPI) were assessed using semi-quantitative visual analysis to determine the extent and severity of perfusion abnormalities. Left ventricular ejection fraction (LVEF) was quantified at rest and during peak stress. Rest and stress myocardial blood flow (MBF) were calculated using factor analysis and a 2-compartment kinetic model, and were used to compute CFR (stress/rest MBF). All cause mortality was ascertained using the National & Social Security Death Indices.

Results For patients with normal, mild-moderate, and severely abnormal scans based on summed stress scores (SSS) on MPI, the observed annualized mortality were 3.6, 7.2, and 13.1%, respectively (overall 290 deaths). After correction for age, sex, hypertension, dyslipidemia, family history of CAD, tobacco use, known CAD, chest pain, dyspnea, SSS, resting LVEF, and stress induced change in LVEF, reduced global CFR correlated with increased mortality (HR 1.39 per 0.5 unit decrease, p<0.0001). Cox proportional hazards analysis revealed that global CFR added incremental prognostic value beyond clinical and gated MPI variables (incremental χ2=30.2, p<0.001). Tertiles of increasing CFR showed a marked trend of higher stress MBF (1.36±0.65 vs. 1.99±0.86 vs. 2.65±1.04, p<0.0001) and only a subtle trend of lower resting MBF (1.18±0.54 vs. 1.15±0.49 vs. 1.02±0.39, p<0.0001), indicating that reductions in CFR were due to impaired peak vasodilator response.

Conclusions Quantitative CFR is a powerful, independent predictor of mortality in patients with known or suspected CAD and provides incremental risk stratification over clinical and gated MPI variables