Effect of an artificial neural network on radiologists' performance in thedifferential diagnosis of interstitial lung disease using chest radiographs

OBJECTIVE. We developed a new method to distinguish between various interst itial lung diseases that uses an artificial neural network. This network is based on features extracted from chest radiographs and clinical parameters . The aim of our study was to evaluate the effect of the output from the ar tificial neural network on radiologists' diagnostic accuracy. MATERIALS AND METHODS. The artificial neural network was designed to differ entiate among 11 interstitial lung diseases using 10 clinical parameters an d 16 radiologic findings. Thirty-three clinical cases (three cases for each lung disease) were selected. In the observer test, chest radiographs were viewed by eight radiologists (four attending physicians and four residents) with and without network output, which indicated the likelihood of each of the 11 possible diagnoses in each case. The radiologists' performance in d istinguishing among the 11 interstitial lung diseases was evaluated by rece iver operating characteristic (ROC) analysis with a continuous rating scale . RESULTS. When chest radiographs were viewed in conjunction with network out put, a statistically significant improvement in diagnostic accuracy was ach ieved (p <.0001). The average area under the ROC curve was .826 without net work output and .911 with network output. CONCLUSION. An artificial neural network can provide a useful "second opini on" to assist radiologists in the differential diagnosis of interstitial lu ng disease using chest radiographs.