Model-based quantification of myocardial perfusion images from SPECT

Area of Science:

• Medical Imaging
• Cardiology
• Quantitative Analysis

Background:

• Myocardial perfusion imaging is crucial for diagnosing cardiac conditions.
• Current bulls-eye algorithms for tomogram analysis have limitations in reproducibility and 3D information retention.
• Accurate quantification of myocardial perfusion defects is essential for patient management.

Purpose of the Study:

• To propose an automated system for quantitative analysis of myocardial perfusion tomograms.
• To improve the accuracy and reproducibility of myocardial perfusion defect detection and quantification.
• To introduce a novel mass bulls-eye map for enhanced defect quantification.

Main Methods:

• Automated delineation of the left ventricle using a flexible computer model to determine myocardial mass and shape.
• Computation of polar maps (bulls-eyes) using radial slices, preserving 3D gradient information.
• Generation of a mass bulls-eye map alongside the traditional count rate bulls-eye for defect quantification.

Main Results:

• The proposed system automatically delineates the left ventricle, capturing mass and shape information.
• Utilizing radial slices retains 3D gradient information, particularly at the ventricle base and apex.
• The mass bulls-eye map allows for precise quantification of perfusion defects identified in the count rate bulls-eye.

Conclusions:

• The developed system offers a more reproducible and accurate method for quantitative analysis of myocardial perfusion tomograms.
• The novel approach using radial slices and a mass bulls-eye map enhances the detection and quantification of perfusion defects.
• This automated system has the potential to improve diagnostic capabilities in cardiology.