Clinical Technology Jpn J Radiol Technol 2000 ;
Differential Diagnosis of Interstitial Lung Disease on Chest Radiograph:
Application of Artificial Neural Network as a Decision Support System

KAZUTO ASHIZAWA, KUNIAKI HAYASHI, TAKAYUKI ISHIDA,1) SHIGEHIKO KATSURAGAWA,2) and KUNIO DOI3)
Department of Radiology, Nagasaki University School of Medicine
1) Department of Radiology, Hiroshima International University
2) Department of Medical Engineering, Iwate Medical School
3) Kurt Rossmann Laboratories for Radiologic Image Research, Department of Radiology,
University of Chicago
Received Nov. 5, 1999; Revision accepted Dec. 27, 1999; Code No. 242
Summary
We applied an artificial neural network to the differential diagnosis of interstitial lung disease on chest radiographs. Our neural network consisted of 26 input units, including 16 radiological findings and 10 clinical parameters, and 11 output units corresponding to the 11 types of interstitial lung disease. Our results indicate that the neural network has a high level of performance for the differential diagnosis of interstitial lung disease and can significantly improve the diagnostic accuracy of observers. In conclusion, the decision support system using the neural network can assist observers in the differential diagnosis of interstitial lung disease on chest radiographs, when the output of the neural network is used as a "second opinion."
Key words: Chest radiography, Neural network, Differential diagnosis, Interstitial lung disease