Cardiac PET perfusion and metabolism to characterize local electrophysiological tissue properties for ventricular tachycardia ablation

Objectives The goal of this study is to evaluate the use of cardiac PET perfusion and viability images for predicting local electrophysiological (EP) tissue properties for patients undergoing ablation procedures for left ventricular tachycardia.

Methods Patients with ischemia who had both Rb-82 perfusion (rest) and F-18 fluorodeoxyglucose (FDG) viability PET before an EP radiofrequency ablation procedure were included in this retrospective study (n=18, 18 male, age 68±10 yr). The EP procedure included tissue substrate characterization with a catheter-based electroanatomic mapping system. Standard bipolar voltage EP tissue categories were scar (<0.5 mV), border zone (BZ) (0.5-1.5 mV) and normal (>1.5 mV). Polar maps of Rb-82 perfusion and FDG uptake were generated, registered and fused with discrete EP voltage data points using our custom software tool (CardioViewer). Using PET and EP data at spatially matched positions, histograms of PET values for different EP tissue categories were computed along with means and standard deviations. Rb-82 and FDG values were used in an ROC analysis to predict EP-derived scar (< 0.5 mV) and abnormal (< 1.5 mV) tissue classes.

Results There were 5625 matched PET-EP points for Rb-82 and 5619 for FDG. Mean normalized Rb-82 perfusion % was 52±23 for scar, 63±23 for BZ and 77±21 for normal (p< 0.001 between all categories). Mean normalized FDG uptake % was 40±26 for scar, 52±25 for BZ and 66±25 for normal (p< 0.001 between all categories). For prediction of EP scar tissue class the area under the ROC curve (AUC) was 0.72±0.01 for Rb-82 and 0.71±0.01 for FDG (p < 0.02). For prediction of EP abnormal tissue class the AUC curve was 0.74±0.01 for Rb-82 and 0.71±0.01 for FDG (p < 0.001).

Conclusions Local EP tissue category can be predicted from PET Rb-82 perfusion and FDG viability data using a novel point-specific approach. Prospective application of such predictions may aid the electrophysiologist in EP ablation procedures for left ventricular tachycardia.

Research Support This work was supported by the Proposed Research Initiated by Students and Mentors (PRISM) program of the University of Maryland School of Medicine and a Bradley-Alavi Fellowship from the Society of Nuclear Medicine and Medical Imaging.