Semi-quantitative segmental perfusion scoring in myocardial perfusion SPECT: visual vs. automated analysis

Introduction: It is recommended that the physician apply at least a semi-quantitative segmental scoring system in myocardial perfusion SPECT.  We aimed to assess the agreement between automated semi-quantitative analysis using QPS (quantitative Perfusion SPECT) software and visual approach for calculation of summed stress  score (SSS), summed rest score (SRS) and summed difference score (SDS). Methods: We retrospectively studied 1782 consecutive patients who had undergone two-day stress-rest Tc99m-MIBI myocardial perfusion SPECT. Based on 17-segment 5-scale scoring system, SSS, SRS and SDS were calculated visually and using QPS software. Results: There was good correlation between visual analysis and QPS software in calculation of SSS and SRS and a fair correlation for SDS. However, there was statistically significant difference between two methods. By Bland-Altman analyses mean value of the differences (estimated bias) differs significantly from 0 on the basis of 1-sample t-test. Based on bias, Precision and 95% limits of agreement, discrepancies between measurements indicate no agreement equally through the range of measurements, so there is a proportional bias. Based on SSS, SRS and SDS ≤3 and SSS, SRS and SDS >3, there was fair concordance between the visual assessment and automated QPS calculation. Kappa statistics was 0.41 (P<0.001), 0.69 (P<0.001) and 0.25 (P<0.001) for SSS, SRS and SDS respectively. Conclusion: Although semi-quantification sores by visual and automated analysis is correlated, the agreement assessed by Bland-Altman analysis is not high especially in more extensive perfusion defects. Semi-quantitative automated analysis should be used as a supplement to the visual assessment.