Potential of digital chest radiography-based deep learning in screening and diagnosing pneumoconiosis

Background To improve the accuracy of pneumoconiosis diagnosis, a computer-assisted method was developed. Methods Three CNNs (Resnet50, Resnet101, and DenseNet) were used for pneumoconiosis classification based on 1,250 chest X-ray images. Three double-blinded experienced and highly qualified physicians read the collected digital radiography images and classified them from category 0 to category III. The results of the three physicians in agreement were considered the relative gold standards. Subsequently, three CNNs were used to train and test these images and their performance was evaluated using multi-class classification metrics. We used kappa values and accuracy to evaluate the consistency and reliability of the optimal model with clinical typing. Results ResNet101 was the optimal model among the three CNNs. The AUC of ResNet101 was 1.0, 0.9, 0.89, and 0.94 for detecting pneumoconiosis categories 0, I, II, and III, respectively. The micro-average and macro-average mean AUC values were 0.93 and 0.94, respectively. The accuracy and Kappa values of ResNet101 were 0.72 and 0.7111 for quadruple classification and 0.98 and 0.955 for dichotomous classification, respectively, compared with the relative standard classification of the clinic. Conclusion The ResNet101 model performed relatively better in classifying pneumoconiosis than radiologists. The dichotomous classification displayed outstanding performance, thereby indicating the feasibility of deep learning techniques in pneumoconiosis screening.