Abstract 1770: Left Ventricular Filling Curve Variables are Superior to Resting Lung to Heart Ratios in Predicting Left Ventricular End-diastolic Pressure in Patients with Preserved Systolic Function Undergoing Gated Myocardial Perfusion Imaging

Abstract

Introduction: There is a 4–5 fold increase in morbidity and mortality when diastolic dysfunction is present in patients with coronary artery disease (CAD). Thallium-201 lung/heart ratio (TLHR) at stress has been proven to capture this increased risk. Most laboratories however use technetium-99m agents for stress myocardial perfusion imaging (MPI) which prevents access to stress TLHR. Method: This is a retrospective study of 52 consecutive patients (age 58±11 years, 73% had CAD) who underwent MPI and cardiac catheterization within 15 days, without interval change in medication or intervention. All patients had ejection fraction ≥45% and normal sinus rhythm. Left ventricular end-diastolic pressure (LVEDP) at cardiac catheterization was recorded. From MPI, peak filling rate (PFR), time to peak filling rate (TPFR), filling rate during first 1/3 of diastole (1/3FR) were obtained using 4D-MSPECT (University of Michigan) and rest TLHR was calculated manually. Results: PFR, TPFR and 1/3FR correlated significantly with LVEDP (r= −0.53, 0.45, −0.45 and p =0.00005, 0.0008, 0.0009 respectively). Rest TLHR did not correlate with PFR, TPFR and 1/3FR (r=0.11, 0.06, 0.09 and p =0.47, 0.68, 0.55 respectively) or with LVEDP (r=0.10, p =0.49). These variables had excellent negative predictive value (NPV) of ≥90% but modest positive predictive value (PPV) of ≥54%. Since PFR did not correlate with 1/3FR (r=0.18, p =0.19), we combined these 2 variables. This combination had a NPV of 84% and improved PPV to 86% and specificity to 94%. Conclusions: Our study validates diastolic filling variables from 4D-MSPECT showing excellent correlation with LVEDP at cardiac catheterization ( p ≤0.0009). PFR, TPFR and 1/3FR are superior to rest TLHR and they have excellent sensitivity (≥82%) and specificity (≥63%) for detecting LVEDP ≥18mmHg. Hence, a valid methodology of combining the risk of perfusion defects with diastolic impairment can be achieved by adding PFR, TPFR and 1/3FR to MPI results. Receiver Operated Characteristic Curve Analysis