Computer-Aided Diagnosis for Pneumoconiosis Staging Based on Multi-scale Feature Mapping

Yuan Zhang

doi:10.1007/s44196-021-00046-5

Abstract

In this research, we explored a method of multi-scale feature mapping to pre-screen radiographs quickly and accurately in the aided diagnosis of pneumoconiosis staging. We utilized an open dataset and a self-collected dataset as research datasets. We proposed a multi-scale feature mapping model based on deep learning feature extraction technology for detecting pulmonary fibrosis and a discrimination method for pneumoconiosis staging. The diagnostic accuracy was evaluated using under the curve (AUC) of the receiver operating characteristic (ROC) curve. The AUC value of our model was 0.84, which showed the best performance compared with previous work on datasets. The diagnosis results indicated that our method was highly consistent with that of clinical experts on real patient. Furthermore, the AUC value obtained through categories I–IV on the testing dataset demonstrated that categories I (AUC = 0.86) and IV (AUC = 0.82) obtained the best performance and achieved the level of clinician categorization. Our research could be applied to the pre-screening and diagnosis of pneumoconiosis in the clinic.

Highlights

Pneumoconiosis is an occupational disease that mainly develops through the long-term inhalation of large amounts of industrial dust into the lungs, causing diffusive fibrosis in lung tissues during occupational activities
The results showed that the method was significantly better than the artificial neural network (ANN) method or the rule-based method alone, indicating that the method could help radiologists classify pneumoconiosis
To assess the clinical accuracy of the proposed method, we randomly applied the discriminant method of pulmonary fibrosis to 10 diagnosed digital radiography (DR) patients (Fig. 7) from a self-collected dataset while hiding the patient information by monitoring the experimental parameters and setting the overlap threshold of the prediction and the GT box to 0.67

Summary

Introduction

Pneumoconiosis is an occupational disease that mainly develops through the long-term inhalation of large amounts of industrial dust into the lungs, causing diffusive fibrosis in lung tissues during occupational activities. The latest diagnostic criteria of pneumoconiosis are based on high kilo-voltage (HKV) chest X-rays and digital radiography (DR) [1] combined with a comprehensive analysis of the patient’s dust exposure history, clinical manifestations and assisted examinations. With the development of artificial intelligence technology and medical image research, medical image-aided diagnosis/detection technology has been gradually applied in the clinic and pneumoconiosis can be prevented through prescreening. We use multi-scale feature mapping technology based on deep learning to pre-screen radiographs quickly and accurately in the aided diagnosis of pneumoconiosis. The results on testing dataset and validation dataset showed that the mode is better than two other researchers for recognition of pulmonary fibrosis

Methods

Results

Conclusion