Diagnosis of Severe Aortic Stenosis Using Implemented Expert System

Aortic stenosis [AS] is set on the forth place of all cardiovascular disease in western world. Severe aortic stenosis [SAS] is one of the most common disease that in treatment demand cardiac surgery-replacement of native aortic valve with mechanical or biological prosthesis in our environment. SAS is often a consequence of a relapsed rheumatic heart disease. This paper present the possible modality of diagnosis SAS. Testing was done on 107 samples, 70 samples were SAS, and the rest of 37 were patients who were sent to some another examination, after ultrasound was done and diagnosis of SAS was excluded. In this paper, the application of basic kind of ANN architecture-the feedforward networks with back-propagation training-is explained, since this structure is able to perform nonlinear multiple regressions in a reliable manner, avoid overfitting. Dataset consists of 12 parameters as shown in Table X; sex, age, dizziness/loss of consciousness, dyspnoea, palpitation, systolic murmur of AS, EF, AV morphology, AVVmax, AVPGmean, AVPGmax, Area. Although limited number of data were present in this study, such automated systems can be used as assistant tool during diagnosis in real-time clinical settings. In the future it is planed to test and use this automated system for classification of aortic stenosis.