Quantitative myocardial perfusion-SPECT: algorithm-specific influence of reorientation on calculation of summed stress score

Abstract

Purpose: Myocardial perfusion SPECT (MPS) with software-assisted determination of summed stress score (SSS) is of established importance for diagnosis/therapy planning in coronary artery disease. Differences in contour finding suggest algorithm-specific influence on quantification if heart axes are chosen incorrectly. Thus, this study quantified the influence of heart-axis tilt on SSS calculation using Quantitative Perfusion SPECT and 4D-MSPECT.

Patients and methods: Stress MPS of 50 men acquired on a triple-head gamma camera were correctly reoriented by experienced technologists (R0) and then tilted by 5 degrees/10 degrees/15 degrees/20 degrees/30 degrees/45 degrees along both long axes (R5-R45). SPECT images were quantified for SSS using QPS and 4D-MSPECT. SSS values for R0 and R5-R45 were analyzed using correlation analysis. Weighted kappa values (κ) were calculated to measure agreement regarding perfusion abnormality severity (4-step SSS rating: 0-3, 4-8, 9-13, and ≥14).

Results: For QPS SSS correlation, R0 vs. tilted datasets remained very high (R > 0.97) up to 20 degrees, but degraded for higher tilts (R = 0.895/0.780 for 30 degrees/45 degrees). 4D-MSPECT showed comparable SSS correlation only up to 10 degrees (R > 0.95) and strong deterioration thereafter (R = 0.863-0.347 for 15-45 degrees). Deviation in severity class from R0 increased from 6/50 (R5; κ = 0.914) to 25/50 (R45; κ = 0.252) using QPS and from 7/50 (R5; κ = 0.899) to 33/50 (R45; κ = 0.065) using 4D-MSPECT.

Conclusion: For tilted MPS datasets, considerable differences in SSS calculation emerge using QPS and 4D-MSPECT. QPS showed more stable results than 4D-MSPECT.