Prediction of Obstructive Lung Disease from Chest Radiographs via Deep Learning Trained on Pulmonary Function Data

Joyce D Schroeder,1 Ricardo Bigolin Lanfredi,2 Tao Li,3 Jessica Chan,1 Clement Vachet,4 Robert Paine III,5 Vivek Srikumar,3 Tolga Tasdizen2 1Department of Radiology and Imaging Sciences, School of Medicine, University of Utah, Salt Lake City, UT, USA; 2Department of Electrical and Computer Engineering, Scientific Computing and Imaging Institute (SCI), University of Utah, Salt Lake City, UT, USA; 3School of Computing, University of Utah, Salt Lake City, UT, USA; 4Biomedical Imaging and Data Analytics Core, SCI, University of Utah, Salt Lake City, UT, USA; 5Division of Pulmonary and Critical Care Medicine, School of Medicine, University of Utah, Salt Lake City, UT, USACorrespondence: Joyce D SchroederDepartment of Radiology and Imaging Sciences, School of Medicine, University of Utah, 30 North 1900 East, Rm #1A71, Salt Lake City, UT 84132, USATel +1 801 581 7553Fax +1 801 581 2414Email joyce.schroeder@hsc.utah.eduBackground: Chronic obstructive pulmonary disease (COPD), the third leading cause of death worldwide, is often underdiagnosed.Purpose: To develop machine learning methods to predict COPD using chest radiographs and a convolutional neural network (CNN) trained with near-concurrent pulmonary function test (PFT) data. Comparison is made to natural language processing (NLP) of the associated radiologist text reports.Materials and Methods: This IRB-approved single-institution retrospective study uses 6749 two-view chest radiograph exams (2012– 2017, 4436 unique subjects, 54% female, 46% male), same-day associated radiologist text reports, and PFT exams acquired within 180 days. The Image Model (Resnet18 pre-trained with ImageNet CNN) is trained using frontal and lateral radiographs and PFTs with 10% of the subjects for validation and 19% for testing. The NLP Model is trained using radiologist text reports and PFTs. The primary metric of model comparison is the area under the receiver operating characteristic curve (AUC).Results: The Image Model achieves an AUC of 0.814 for prediction of obstructive lung disease (FEV1/FVC < 0.7) from chest radiographs and performs better than the NLP Model (AUC 0.704, p< 0.001) from radiologist text reports where FEV1 = forced expiratory volume in 1 second and FVC = forced vital capacity. The Image Model performs better for prediction of severe or very severe COPD (FEV1 < 0.5) with an AUC of 0.837 versus the NLP model AUC of 0.770 (p< 0.001).Conclusion: A CNN Image Model trained on physiologic lung function data (PFTs) can be applied to chest radiographs for quantitative prediction of obstructive lung disease with good accuracy.Keywords: machine learning, chronic obstructive pulmonary disease, quantitative image analysis, natural language processing