Results of computer-aided diagnosis of peripheral bronchial carcinoma

Radiographic images can be converted into numerical sequences that can be evalueate by computer to determine the significance of certain radiographic findings in peripheral pulmonary nodules. Of three methods (discriminant analysis. Bayes' theoroem, and the Calm system), discriminant analysis has proven the best with an accuracy rate of 85%. By taking into account serial radiographs and statements on the possible growth of the neoplasm, the accuracy of the method can be increased to 90%. A simplified, inexpensive and effective method for application of this diagnostic tool is described.