Deep Convolutional Neural Network–based Software Improves Radiologist Detection of Malignant Lung Nodules on Chest Radiographs
Summary
In a multicenter study, 12 radiologists assisted by deep convolutional neural network software detected malignant pulmonary nodules on chest radiographs with a higher sensitivity and fewer false-positive findings per image compared with radiologists alone, irrespective of radiologist experience, nodule characteristics, or the vendor of the radiography acquisition system.
Key Results
■ Use of deep convolutional neural network (DCNN)–based software improved the average sensitivity of radiologists for the detection of malignant pulmonary nodules on chest radiographs (from 65.1% before DCNN to 70.3% after DCNN, P < .001).
■ Radiologists’ use of DCNN software decreased the number of false-positive marks per chest radiograph by 0.02 (from 0.20 per radiograph before DCNN to 0.18 after DCNN, P < .001).
■ For the 12 radiologists in this study, 104 of 2400 radiographs were positively changed (from false-negative to true-positive or from false-positive to true-negative) using the DCNN, while 56 of 2400 radiographs were changed negatively.
Full article HERE