Comparison between a count-based and geometrical approach for the assessment of left ventricular dyssynchrony using myocardial perfusion scintigraphy

OBJECTIVE: There are two distinct approaches for the assessment of left ventricular (LV) dyssynchrony by myocardial perfusion scintigraphy (MPS). The aim of this study was to compare the performance of the count-based and geometrical approach in clinical data using gated single photon emission computed tomography MPS. MATERIAL AND METHODS: Group 1 consisted of 113 patients (49 men, 64 women) with normal perfusion [summed rest score (SRS)≤3], normal LV ejection fraction (≥55%), and normal QRS duration (QRSd<120 ms). Group 2 consisted of 89 heart failure patients (79 men, 10 women) with no restriction for SRS, LV ejection fraction ≤35%, and QRSd ≥120 ms. All MPS parameters were obtained from the software Corridor4DM. Dyssynchrony parameters used were time to peak contraction, SD, and bandwidth (BW). RESULTS: SD and BW were estimated higher (difference group 1: SD 3.0±2.3 and BW 11.3±9.3, P-values <0.001; difference group 2: SD 2.4±4.3 and BW 1.3±17.0, P-value <0.001 and 0.479 respectively) using the count-based approach in comparison with the geometrical method. A significant and good correlation was found between these two methods (R=0.763, 0.902, 0.896 for time to peak contraction, SD, and BW respectively, P-values ≤0.001). SD and BW in both approaches were equally good parameters for differentiating heart failure patients (area under the curve: 0.995-0.998), although using different cut-off values. CONCLUSION: The count-based approach generally provides a wider phase distribution and subsequently greater SD and BW estimates compared with the geometrical algorithm. These differences result in clinically relevant deviations in normal and cut-off values that have to be recognized when evaluating patients.