Expert systems in dermatology: the computer potential. The example of facial tumour diagnosis

The expert system approach to computer diagnosis uses a non-algorithmic method to represent and manipulate an expert's knowledge and reasoning. This information, which may be provided by a dermatologist, is represented by rules in a logic-based computer language in order to provide interactive and explanatory features. The major advantages of using expert system techniques for computer-aided diagnosis in dermatology are that knowledge is made explicit, the heuristic nature of an expert's knowledge can be more easily captured and the more easily readable programs make modification easier. In the example described, the differential diagnosis of nine facial skin tumours was considered. The program, using the language 'micro-PROLOG' in an expert system shell 'APES', consists of 'rules' and 'facts' which define a relationship between patient and symptom, symptom and disease or disease and therapy. The 'strength' of each relationship is defined and 'key symptoms' identified. The system finally offers a diagnosis, an estimate of certainty and simple management advice.