The prevalence of shoulder pain is high in many countries, and a lifetime prevalence of shoulder pain is up to 70%. In America, shoulder pain costs the health care system 7 billion per year. With respect to the etiologies, up to 70% of shoulder pain is attributed to rotator cuff lesions. According to Neer’s classification system, the lesions of rotator cuff can be classified into inflammation, calcific tendinitis, and full or partial thickness tears. Patients with rotator cuff lesions have shoulder pain, positive impingement sign, limited forward elevation, weak abduction, and external rotation which may induce the difficulties of holding things. Rotator cuff tears is the most severe type which would cause severe shoulder pain and impingement sign, limited forward elevation and weak abduction and external rotation. Especially for current population with more and more aging people, the prevalence rate of rotator cuff tears would be expected to become higher. Clinical assessment relies on imaging modalities to evaluate the integrity of rotator cuff tendons. Previous literatures recommend shoulder ultrasound as a useful imaging tool to detect rotator cuff lesions1 and full-thickness rotator cuff tears. To strengthen the clinical use of ultrasound, the inter-operator variabilities should be further reduced. Computer-aided Detection (CAD) system has been proposed to improve the examination quality. The advantages of CAD systems are quantitative, efficiency, and consistent. Thus, a CAD system based on multi-layer active contour model of supraspinatus detection in should ultrasound image was proposed in this project to reduce operator dependence. After the construction of the CAD system, the lesion location and delineation can be standardized to provide quantitative analysis and quality control.
|Effective start/end date||8/1/17 → 7/31/18|
- rotator cuff lesions
- shoulder ultrasound
- computer-aided detection
Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.