Novel indices representing heterogeneous distributions of myocardial perfusion imaging

Area of Science:

• Nuclear Medicine
• Cardiovascular Imaging
• Quantitative Analysis

Background:

• Myocardial perfusion imaging (MPI) can show heterogeneous tracer distribution in cardiac diseases, even with normal perfusion.
• Quantitative methods for assessing this heterogeneity are lacking.
• Histogram analysis offers a potential solution for quantifying MPI heterogeneity.

Purpose of the Study:

• To quantitatively evaluate heterogeneity in MPI using histogram analysis.
• To assess standard deviation (SD), 95% bandwidth (BW95%), and entropy as indices of heterogeneity.
• To determine if these indices can differentiate between heterogeneous and non-heterogeneous MPI patterns.

Main Methods:

• Resting 99mTc-MIBI SPECT images from 20 healthy subjects and 29 cardiac patients were analyzed.
• Visual assessment by nuclear medicine specialists classified images into non-heterogeneity or heterogeneity groups.
• Histogram analysis calculated SD, BW95%, and entropy from %uptake data on polar maps.
• Receiver operating characteristic (ROC) curve analysis evaluated the diagnostic performance of these indices.

Main Results:

• 38 out of 49 cases (78%) were classified as heterogeneous based on visual assessment.
• All cardiac patients were in the heterogeneity group, while the non-heterogeneity group comprised only healthy subjects.
• Heterogeneity indices (SD, BW95%, entropy) were significantly higher in the heterogeneity group (p < 0.05).
• Area under the curve (AUC) values exceeded 0.90 for all indices, indicating high diagnostic accuracy.

Conclusions:

• Standard deviation (SD), 95% bandwidth (BW95%), and entropy derived from histogram analysis provide quantitative measures of MPI heterogeneity.
• These novel indices demonstrate potential for identifying subtle myocardial changes in patients with preserved perfusion.
• Quantitative heterogeneity assessment in MPI can aid in early cardiac disease detection.