Bhabha Atomic Research Centre

Myocardial perfusion imaging (MPI) aids in evaluating the left ventricular function and therefore helps in risk stratification and prognosis. The purpose of this study was to evaluate the clinical impact of three cardiac software packages (Emory Cardiac Toolbox, Cardiogam, and Myovation Evolution) at 1-year follow-up. This prospective investigation involved 370 patients who underwent myocardial perfusion scintigraphy. Participants were stratified into three groups based on left ventricular ejection fraction determined by two-dimensional echocardiography. Among these, 294 patients were available for follow-up at 1 year. Semiquantitative parameters (summed rest score [SRS], summed stress score [SSS], summed difference score [SDS]) and quantitative parameters (defect extent [DE], transient ischemic dilatation [TID], ejection fraction [EF], end-diastolic volume [EDV], end-systolic volume [ESV], and stroke volume [SV]) derived from each software platform (ECTb, Cgam, and ME) were used to evaluate their diagnostic performance in predicting major adverse cardiac events (MACE) using receiver operating characteristic curve analysis in patients during 1-year follow-up. Among the 294 participants available for follow-up, 6% ( n = 17) developed MACE at 1 year and 88% ( n = 15) belonged to the high-risk group. Follow-up analysis indicated that all three software packages could predict 1-year cardiac events with varying cut-off values, sensitivities, and specificities. Cut-off values for SRS (ECTb) was ≥4 (area under the curve [AUC], 0.82), while for Cgam it was ≥11 (AUC 0.8), DE (REST) (ECTb) was ≥6% (AUC 0.82), and Cgam value ≥4.18% (AUC 0.81), EF (REST) was ≤44% (AUC 0.82), ≤68% (AUC 0.84) and ≤35% (AUC 0.83) for ECTb, Cgam, and ME, respectively, EDV (REST) was ≥137 mL for both ECTb (AUC 0.83) and Cgam (AUC 0.85) while for ME it was ≥116 mL (AUC 0.8), ESV (REST) was ≥91 mL (AUC 0.83), ≥55 mL (AUC 0.86) and ≥78 mL (AUC 0.82) for ECTb, Cgam, and ME, respectively. Majority of the MACE occurred in the group with high-risk characteristics. All three software packages demonstrated the ability to predict 1-year cardiac events, albeit with varying parameter cut-off values, sensitivities, and specificities.