Using a Deep Learning Model to Explore the Impact of Clinical Data on COVID-19 Diagnosis Using Chest X-ray

Area of Science:

• Medical Imaging
• Artificial Intelligence in Medicine
• Infectious Disease Diagnostics

Background:

• The COVID-19 pandemic necessitates rapid and accurate diagnostic tools.
• Machine learning, particularly deep learning, shows promise in automating disease diagnosis.
• Integrating diverse data sources can enhance diagnostic model performance.

Purpose of the Study:

• To propose and evaluate a deep learning model for automated COVID-19 diagnosis.
• To investigate the impact of combining chest X-ray (CXR) images with clinical data for improved diagnosis.
• To assess the model's performance against expert diagnosis.

Main Methods:

• A deep learning model was developed using patient data (270 records) from King Fahad University Hospital.
• Experiments were conducted using clinical data alone, CXR images alone, and a fusion of both.
• A fusion technique combined clinical features with image-extracted features.

Main Results:

• The model achieved high diagnostic performance when integrating clinical data and CXR images.
• Specific metrics included accuracy (0.970), recall (0.986), precision (0.978), and F-score (0.982).
• The system's performance was validated against expert diagnoses, demonstrating its potential utility.

Conclusions:

• Integrating clinical data with CXR images significantly enhances automated COVID-19 diagnostic accuracy.
• The proposed deep learning system serves as a valuable assistive tool for clinicians.
• This approach offers a promising method for timely and accurate COVID-19 detection.