Background: This study investigated the feasibility of using a computer-assisted method to evaluate and differentiate the carotid plaque characteristics in radiation-induced and non-radiation-induced carotid atherosclerosis. Methods: This study included 107 post-radiotherapy (post-RT) nasopharyngeal carcinoma (NPC) patients and 110 subjects with cardiovascular risk factors (CVRFs). Each participant had a carotid ultrasound examination, and carotid plaques and carotid intima-media thickness (CIMT) were evaluated with grey scale ultrasound. The carotid plaque characteristics were evaluated for grey-scale median (GSM) and detailed plaque texture analysis (DPTA) using specific computer software. In DPTA, five different intra-plaque components were colour-coded according to different grey scale ranges. A multivariate linear regression model was used to evaluate the correlation of risk factors and carotid plaque characteristics. Results: Post-RT NPC patients have significantly higher CIMT (748 +/- 15.1 mu m, P=0.001), more patients had a plaque formation (80.4%, P<0.001) and more plaque locations (2.3 +/- 0.2, P<0.001) than CVRF subjects (680.4 +/- 10.0 mu m, 38.2% and 0.5 +/- 0.1 respectively). Among the five intra-plaque components, radiationinduced carotid plaques had significantly larger area of calcification (4.8%+/- 7.7%, P=0.012), but lesser area of lipid (42.1%+/- 16.9%, P=0.034) when compared to non-radiation-induced carotid plaques (3.0%+/- 5.7% and 46.3%+/- 17.9% respectively). Age, radiation and number of CVRF were significantly associated with the carotid atherosclerosis burden (P<0.001). Besides, age was significantly associated with the amount of lipid and calcification within carotid plaques (P<0.001). Conclusions: Radiation caused more severe carotid artery disease than CVRF with larger CIMT and more prevalent of carotid plaque. Radiation-induced carotid plaques tended to have more intra-plaque calcifications, whereas non-radiation-induced carotid plaques had more lipids. Ultrasound aided by computer-assisted image analysis has potential for more accurate assessment of carotid atherosclerosis.