ACUTE PULMONARY-EMBOLISM - ARTIFICIAL NEURAL-NETWORK APPROACH FOR DIAGNOSIS

PURPOSE: To investigate use of an artificial neural network (ANN) as a computer-aided diagnostic (CAD) tool for predicting pulmonary embolis m (PE) from ventilation-perfusion lung scans and chest radiographs. MA TERIALS AND METHODS: The data base consisted of cases extracted from t he collaborative study of the Prospective Investigation of Pulmonary E mbolism Diagnosis (PIOPED). Initially, scan findings from 1,064 patien ts (383 with PE, 681 without PE) were used to train and test the netwo rk by using the ''jackknife'' method. Then, a receiver-operating-chara cteristic analysis was applied to compare the performance of the netwo rk with that of the physicians involved in the PIOPED study. RESULTS: The ANN significantly outperformed the physicians involved in the PIOP ED study (two-tailed P value = .01). CONCLUSION: The findings suggest that an ANN can form the basis of a CAD system to assist physicians wi th the diagnosis of PE.