Adventitia Segmentation on Superficial Femoral Artery Optical Coherence Tomography Images

Atherosclerotic disease on peripheral arteries, commonly known as peripheral artery disease represents accumulation of cholesterol in peripheral arteries. It causes the formation of different types of depositions called plaques which cause thickening of the arterial walls resulting in reduced blood flow to lower extremities, in case of femoral arteries, upper extremities, in case of brachial artery or brain in case of carotid artery. Current research shows that approximately 8.5 million Americans, aged over 40 years, are affected by peripheral artery disease while one fourth of them falls into a severe category. For these reasons, an early detection of the disease is important. In order to detect the disease, arteries need to be imaged properly. In the recent years, an increase of optical coherence imaging has occurred due to the ease of use and the ability to detect tissue morphology which is extremely important for the detection of the disease. From the tissue morphology, it is of most importance to observe the tissue between lumen and intimal layer of the artery, since the majority of the plaques form in that region, but it is also important to observe the region between intima and adventitia (the outer wall of the artery) since the plaque can penetrate the intimal layer as well. In this paper, a deterministic approach for the detection of the adventitia is described based on the previously detected intimal layer of the artery. The proposed method is evaluated on optical coherence tomography images from 8 specimens of porcine femoral arteries. The results show that the detection of adventitia is possible on this image modality with Dice coefficient 0.9805 and Hausdorff distance 0.1 mm.