Deep Learning-Based Computed Tomography Features in Evaluating Early Screening and Risk Factors for Chronic...

Area of Science:

• Medical Imaging
• Artificial Intelligence in Medicine
• Pulmonology

Background:

• Chronic obstructive pulmonary disease (COPD) poses a significant global health challenge.
• Accurate and early diagnosis of COPD is crucial for effective management.
• Current diagnostic methods may have limitations in sensitivity and specificity.

Purpose of the Study:

• To evaluate the diagnostic performance of a deep learning double residual convolution neural network (DRCNN) model for COPD detection using CT images.
• To identify independent risk factors associated with COPD.
• To assess the potential of DRCNN-based CT for early COPD screening.

Main Methods:

• A questionnaire survey identified 980 residents aged ≥ 40 years.
• 84 diagnosed COPD patients and 25 healthy individuals underwent CT scans.
• A DRCNN model was developed for image noise reduction and COPD diagnosis from CT scans.
• Multivariate logistic regression analyzed risk factors.

Main Results:

• The prevalence of COPD in the surveyed population was 8.57%.
• Age, family history of COPD, and smoking were identified as significant independent risk factors (P < 0.05).
• DRCNN-based CT demonstrated superior diagnostic sensitivity, specificity, and accuracy compared to standard CT (P < 0.05).

Conclusions:

• Advanced age, family history of COPD, and smoking are confirmed independent risk factors.
• The DRCNN model significantly enhances the diagnostic accuracy of CT for COPD.
• DRCNN-based CT shows promise for early screening and improved diagnosis of COPD.