Clinical value of gated SPECT-derived left ventricular strain and reserve: derivation of quantitative cutoffs.

Abstract

This study aimed to assess interobserver reproducibility of left ventricular global longitudinal (LS), circumferential (CS), and radial (RS) myocardial strain measurements, to determine cutoff values including coronary territory, ventricular wall, and ventricular region-specific measurements at rest and under stress, and to evaluate strain reserve using gSPECT-MPI. We prospectively analyzed 127 patients undergoing gSPECT-MPI with myocardial strain assessment using SyncTool™ (Emory Cardiac Toolbox™). The reference group included 65 apparently healthy subjects with normal gSPECT findings (age 61.6 ± 11 years; 42.7% men), and the case group with 62 patients with coronary artery disease (age 66.8 ± 11 years; 80% men). Interobserver variability for global LS, CS and RS was good to excellent at rest and during stress (global LS: rest ICC = 0.86, stress ICC = 0.81; global CS: rest ICC = 0.85, stress ICC = 0.87; global RS: rest ICC = 0.74, stress ICC = 0.80). Optimal cutoff values at rest and stress were: LS-rest >-18.75, LS-stress >-22.75, CS-rest >-36.55, CS-stress >-38.15, RS-rest < 34.65, RS-stress < 45.15. Reserve cutoff values were: LS absolute difference (LS-AD) >-0.19, CS-AD >-2.45, RS-AD < 0.19, LS relative change (LS-RCH) < 0.52, CS-RCH <-16.86, RS-RCH <-3.06, LS ratio (LS-R) < 1.01, CS-R < 0.83 and RS-R < 0.97. The optimal model (stress global LS > - 22.75, stress global CS > - 38.15, LS-RCH < 0.52, and RS-RCH < - 3.06) demonstrated good discrimination for case-group classification (AUC, 0.85; sensitivity, 79%; specificity, 70%). This study establishes reproducible and clinically relevant cutoff values for gSPECT-MPI-derived myocardial strain and strain reserve, filling an important gap in current literature.