Structural Equation Modeling for Stroke Risk Assessment of the Common Carotid Artery based on Texture Analysis

Abstract

The intima media thickness (IMT) of the common carotid artery (CCA) as well as texture features extracted from the intima media complex (IMC) of the CCA may be used to evaluate the prevalent clinical cardiovascular disease (CVD) and the risk of stroke. This study investigated the association between the IMT and the texture features of the IMC of the CCA and the prevalent clinical CVD using structural equation modeling (SEM). Six hundred twelve (612) longitudinal-section ultrasound images of the left and right CCA were obtained from 306 subjects (158 men and 148 women), 42 of which had clinical CVD. Each of these images was intensity normalized and despeckled. Forty (40) texture features were extracted from the IMC through a semi-automated segmentation system. To this end, we proposed a new approach on how to analyze the above data and provide an evaluation of the stroke risk. SEM is an elegant procedure employed in this study for a conceptual model of relationships between 7 different factors (unobserved constructs and observable variables) extracted from the IMC ultrasound images. The main findings of this study are: (i) Out of seven IMC texture feature groups (factors) investigated, six of them showed good fit to the conceptual model. (ii) Six hypothesized paths in the conceptual model for the impact of each texture feature group on the IMT were tested. It turned out that five of the chosen factors have a significant impact on IMT. (iii) However, as the fits of both the measurement and the structural models are not so good, the obtained results need to be improved taking certain measures in relation to the SEM. Future work will investigate the relationships between IMC texture features and IMT in relation to CVD and the carotid side.