Analysis of Myocardial Textures in Relation to Nicotine Abuse Using Radiomics in Cardiac PCCT

Area of Science:

• Cardiovascular Imaging
• Radiomics
• Medical Physics

Background:

• Photon-counting computed tomography (PCCT) and radiomics offer advanced myocardial tissue characterization.
• Conventional imaging may miss early signs of cardiovascular changes.
• This study explores radiomic signatures of nicotine status in patients without coronary artery disease.

Purpose of the Study:

• To investigate if myocardial radiomic features from PCCT are associated with nicotine status.
• To assess the potential of cardiac radiomics for early cardiovascular risk stratification.

Main Methods:

• Retrospective analysis of 104 patients without coronary calcification using cardiac PCCT.
• Extraction and reduction of 105 radiomic features from myocardial tissue.
• Training and evaluation of machine learning models (logistic regression, random forest, gradient boosting) for smoker/non-smoker discrimination.

Main Results:

• No significant difference in myocardial septal thickness between smokers and non-smokers.
• Radiomic features moderately discriminated between smoking groups (ROC-AUC 0.66).
• Higher-order texture and shape-based features were most relevant.

Conclusions:

• PCCT-derived radiomics can identify subtle myocardial changes linked to smoking status.
• Cardiac radiomics shows promise as a non-invasive biomarker for early cardiovascular risk assessment.
• Further multicenter studies are needed for clinical translation.