Radiomics Signature of Aging Myocardium in Cardiac Photon-Counting Computed Tomography

Area of Science:

• Cardiology
• Medical Imaging
• Radiology

Background:

• Cardiovascular diseases are a leading cause of death, particularly in older adults.
• Aging cardiovascular structures require deeper understanding for risk prediction.
• Photon-counting computed tomography (PCCT) enhances cardiac imaging for detailed analysis.

Purpose of the Study:

• To investigate the potential of radiomics texture analysis in differentiating age-related changes in the left ventricular myocardium.
• To compare the efficacy of myocardial texture features against conventional parameters like epicardial adipose tissue (EAT) for age stratification.

Main Methods:

• Retrospective analysis of 90 patients undergoing ECG-gated contrast-enhanced cardiac CT using dual-source PCCT.
• Segmentation of the left ventricular myocardium and extraction of radiomics features using pyradiomics.
• Comparison of texture features between two age groups (50-60 and 70-80 years) using statistical analysis and a random forest classifier.

Main Results:

• Radiomics texture analysis successfully differentiated between the age groups, achieving a test accuracy of 0.74.
• Wavelet-HLH_glszm_GrayLevelNonUniformity emerged as a key feature for age differentiation.
• No significant differences were observed in EAT density, thickness, or other clinical parameters between the age groups.

Conclusions:

• Myocardial radiomics texture features are superior to conventional parameters like EAT for identifying age-related myocardial changes.
• This imaging biomarker holds potential for predicting cardiovascular risk by visualizing myocardial aging.
• PCCT-based radiomics offers novel insights into myocardial structural alterations during aging.