Adaptive ROC analysis for Carotid Atherosclerosis Treatment Assessment

Atherosclerosis disease of the carotid is one of the major reasons for stroke. An accurate diagnosis and severity quantification of the disease is therefore needed to prevent symptoms and decide the appropriate treatment. The most important indicator for the risk of symptoms is the Degree of Stenosis (DS). Very recently, a new score was proposed, called Enhanced Activity Index (EAI), which is more complex but more accurate than DS. EAI, however, only takes into account medical information about the patient. The optimal EAI cut-off for endarterectomy is computed under the Receiver Operating Characteristics (ROC) framework. A Graphical User Interface (GUI) is described to help the medical staff in the decision making process. The optimal decision point is adjusted according to population characteristics, inherent to the ROC curve, patient information and economic aspects. The method is tested with real data acquired at Instituto Cardiovascular de Lisboa and Department of Vascular Surgery, Hospital de Santa Maria, Lisbon. The created tool outputs coincident results to the ones presented in different studies of the bibliography.