A program called ORAD was designed using Bayes' theorem to evaluate the radiographic and clinical features of patients with intrabony lesions in order to assist in their identification. Ninety-eight jaw lesions were described by their prevalence and their distribution by age, sex, race, presence of pain, number, size and location of lesions, association with teeth, expansion, locularity, borders, contents and impact on adjacent teeth. The user follows a menu of 16 questions in order to characterize a specific lesion. The program output is a list of the lesions in order of their estimated probability. In addition, an estimate of the extent of match between the lesion in question and the typical appearance of each lesion in the knowledge base is calculated. Preliminary trials indicate that ORAD is useful in assisting clinicians in formulating a differential diagnosis.