Improvement of lung abnormality detection in computed radiography using multi-objective frequency processing:...

Area of Science:

• Radiology
• Medical Imaging
• Image Processing

Background:

• Computed radiography (CR) exhibits low sensitivity for detecting pulmonary nodules.
• This limitation hinders its use in screening despite advantages like low cost and dose.

Purpose of the Study:

• To apply multi-objective frequency processing (MFP) to CR images.
• To evaluate MFP's effectiveness in diagnosing subtle lung abnormalities.

Main Methods:

• Fifty CR images with simulated subtle lung abnormalities were acquired from 50 volunteers.
• Images were processed using MFP, and 200 half-chest images (100 processed, 100 unprocessed) were analyzed by five radiologists.
• Receiver Operating Characteristic (ROC) analyses were performed.

Main Results:

• MFP processing significantly improved the detection rate of simulated subtle lung abnormalities.
• The area under the ROC curve (Az) was higher for MFP-processed images (0.8508) compared to unprocessed images (0.7925).

Conclusions:

• MFP processing enhances the diagnostic utility of CR for subtle lung abnormalities.
• MFP shows promise in improving sensitivity and specificity for detecting pulmonary nodules, ground-glass opacity (GGO), and reticular shadows.