Preliminary evaluation of a fuzzy logic-based automatic quantitative analysis in myocardial SPECT

Area of Science:

• Nuclear Medicine
• Cardiology
• Medical Imaging Analysis

Background:

• Myocardial perfusion SPECT is crucial for diagnosing coronary artery disease (CAD).
• Accurate quantitative analysis of perfusion tracer distribution is essential for reliable CAD assessment.
• Existing software may have limitations in objectivity and accuracy.

Purpose of the Study:

• To validate a novel quantitative myocardial perfusion Single-Photon Emission Computed Tomography (SPECT) software.
• To assess the software's accuracy in detecting and localizing coronary artery occlusions.
• To evaluate the reproducibility of the software's quantitative analysis.

Main Methods:

• The software utilizes fuzzy logic and a truncated bullet model for image processing and defect filling.
• The left ventricular myocardium is partitioned into 18 isovolumetric sectors.
• Validation involved 343 patients, with reproducibility assessed in 49 and accuracy in 48 patients undergoing coronary angiography.

Main Results:

• High intra- and interoperator reproducibility with regression coefficients of 0.97.
• Overall sensitivity and specificity for detecting occluded coronary arteries were 90% and 80%, respectively.
• High sensitivity (92-92.5%) was observed for LAD and RCA occlusions, with good specificity (75-90%).

Conclusions:

• The new quantitative myocardial perfusion SPECT software provides objective analysis of tracer distribution.
• The software is an accurate tool for detecting and localizing coronary artery occlusions.
• This tool can aid in the clinical diagnosis and management of CAD.