Automatic Global and Regional Phase Analysis from Gated Myocardial Perfusion SPECT Imaging: Application to the Characterization of Ventricular Contraction in Patients with Left Bundle Branch Block

Abstract

Although many patients with heart failure benefit from cardiac resynchronization therapy (CRT), predicting which patients will respond to CRT remains challenging. Recent evidence suggests that the analysis of mechanical dyssynchrony using gated myocardial perfusion SPECT (MPS) may be an effective tool. The aim of this study was to evaluate global and regional gated MPS dyssynchrony measurements by comparing parameters obtained from patients with a low likelihood (LLk) of conduction abnormalities and coronary artery disease and patients with left bundle branch block (LBBB). A total of 86 consecutive patients with LLk and 72 consecutive patients with LBBB, all without prior myocardial infarction or sternotomy, were studied using gated MPS. Global (histogram SD [sigma], bandwidth [beta], and entropy [epsilon]) and regional (wall- and segment-based differences of means [Deltamicro(W) and Deltamicro(S), respectively] or modes [DeltaM(W) and DeltaM(S), respectively]) dyssynchrony measures were calculated by Fourier harmonic phase-angle analysis of local myocardial count variations over the cardiac cycle for each patient, and then unpaired t tests were used to determine which parameters were sex-specific and how well they discriminated between the LLk and LBBB populations. Receiver-operating-characteristic analysis was also performed to calculate the area under the curve (AUC), sensitivity (Ss), specificity (Sp), and optimal threshold (Th). Global parameters were found to be sex-specific, whereas regional differences were sex-independent. All parameters studied showed statistically significant differences between the groups (all global, P < 0.05; all regional, P < 0.0001). Receiver-operating-characteristic analysis yielded higher AUC, Ss, and Sp for epsilon and regional parameters (epsilon: AUC = 0.95/0.96, Ss = 94%/88%, Sp = 89%/91%, and Th = 53.9%/60.6% for women/men; Deltamicro(W): AUC = 0.93, Ss = 88%, Sp = 86%, and Th = 10.5 degrees ; Deltamicro(S): AUC = 0.94, Ss = 90%, Sp = 94%, and Th = 9.2 degrees ; DeltaM(W): AUC = 0.95, Ss = 90%, Sp = 94%, and Th = 15 degrees ; and DeltaM(S): AUC = 0.95, Ss = 88%, Sp = 90%, and Th = 10.5 degrees ) than for global parameters (sigma: AUC = 0.75/0.67, Ss = 81%/66%, Sp = 63%/64%, and Th = 16.5 degrees /22.2 degrees for women/men; beta: AUC = 0.80/0.72, Ss = 71%/71%, Sp = 79%/64%, and Th = 69 degrees /81 degrees for women/men). The computed parameters all discriminate effectively between LLk and LBBB populations. Measurements that are less dependent on the shape of the phase-angle distribution histogram provided higher sensitivity and specificity for this purpose. Further study is needed to evaluate these parameters for the purpose of predicting response to CRT.