Operator dependent variability in quantitative analysis of myocardial perfusion images

Area of Science:

• Nuclear Cardiology
• Medical Imaging Analysis

Background:

• Myocardial perfusion imaging (MPI) is crucial for diagnosing coronary artery disease.
• Accurate quantification of perfusion defects and left ventricular function is essential for patient management.

Purpose of the Study:

• To assess inter-operator variability in quantifying myocardial perfusion defects and left ventricular ejection fraction (LVEF) using two common software programs.
• To determine the impact of operator variability on clinical interpretation of MPI studies.

Main Methods:

• Eight experienced technologists processed MPI studies from 13 patients using Cedars Emory (CEqual) and Cedars-Sinai quantitative gated single-photon emission tomography programs.
• Manual steps included defining left ventricle axes and selecting apex/base for CEqual analysis.
• Quantification of perfusion defect size, reversibility, and LVEF was performed.

Main Results:

• Significant inter-operator variability was observed in quantifying reversible apical perfusion defects across all vascular territories.
• A >10% variability in reversibility was noted in at least one territory for 9 of 13 patients.
• Inter-operator variability for LVEF quantification was consistently low (<5%) across all patients.

Conclusions:

• High variability in quantifying reversible perfusion defects may lead to misinterpretation and false clinical conclusions.
• Low variability in LVEF quantification suggests its reliability across different operators.
• Standardization of manual steps in MPI analysis is needed to improve consistency.