Expert systems in orthodontic clinical diagnosis.

Computer-aided diagnosis of orthodontic anomalies is of particular interest in the screening of potential patients. General practitioners and public dental health workers would find considerable benefit in a reliable automated system capable of reflecting the clinical assessments of orthodontists. This study investigates the application of an expert system (Xpert Rule) to the assessment of treatment need using thirteen traits of malocclusion to construct a knowledge base. Decision trees were generated using a pruning option available with the expert system. The results of the study are encouraging and suggest that an expert system is able to generate a useful clinical screening device. Further modification of the knowledge base is required to provide a broader range of case types, particularly those of less common occurrence. It is also suggested that the random aggregation of cases is not the most effective means of developing a knowledge base.