Abstract
Recent advances in atherosclerotic plaque detection have shown that not only does lipid core size and depth play important roles in plaque rupture and thrombi formation, but lipid composition, especially cholesterol deposition, is equally important in determining lesion vulnerability. Here, we demonstrate a spectral analysis assisted photoacoustic imaging approach to differentiate and map lipid compositions within an artery wall. The approach is based on the classification of spectral curves obtained from the sliding windows along time-of-flight photoacoustic signals via a numerical k-means clustering method. The evaluation result on a vessel-mimicking phantom containing cholesterol and olive oil shows accuracy and efficiency of this method, suggesting the potential to apply this approach in assessment of atherosclerotic plaques.