Analysis of atherosclerotic lesions in the human body

A very important problem in modern medicine is the identification of atherosclerotic lesions within human tubular vessels. Current methods such as OCT and intravascular ultrasonography are very expensive and do not provide conclusive information about the formation of atherosclerotic plaques, especially within the human thoracic aorta. This paper presents a new solution to the identification of atherosclerotic lesions based on a specially constructed endoscope that uses CCD cameras. The endoscope analyses the condition of the walls of the blood vessels (side analysis) and represents a step towards an in vivo investigation of part of the human body. The remaining problem is the exact identification of the places where changes occur in the human tubular vessels. Research using the fluorescence phenomenon gives a summary result from an area whose location can be determined with an accuracy of millimetres. For pipes with sufficiently large internal diameter CCD cameras can be used to produce images of the examined tissues. It is possible for doctors to use the images so obtained to identify diseased areas with high accuracy. An alternative approach for identifying atherosclerosis is to use a neural network method to analyse the received pictures. Popular methods such as Machine Learning and Deep Learning can also be used to identify features in medical images. Creating a ‘learning’ database of endoscopic images in which a doctor identifies regions of healthy and atherosclerotic tissue is time consuming. However, after creating the database, the image identification algorithm works much faster than known numerical methods. It can significantly contribute to improving the effectiveness of atherosclerosis diagnostics. This paper presents initial results that confirm the effectiveness of a Machine Learning approach in identifying atherosclerotic lesions from the analysis of endoscope images obtained with a black and white camera following fluorescence stimulation and with a colour CCD camera using white light illumination. These findings are important since histological tests are not possible in in vivo investigations.