Automatic detection of large pulmonary solid nodules in thoracic CT images

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Radiology

Background:

• Current computer-aided detection (CAD) systems show limitations in detecting larger pulmonary nodules (>10 mm) on computed tomography (CT) scans.
• Larger pulmonary nodules are statistically more likely to be cancerous, necessitating improved detection methods.

Purpose of the Study:

• To develop and evaluate a novel CAD system specifically designed for the accurate detection of solid pulmonary nodules exceeding 10 mm in size.
• To address the performance gap in existing CAD systems for larger, more clinically significant lung nodules.

Main Methods:

• A 3D lung segmentation algorithm was optimized using morphological processing to include large nodules, even those attached to the pleural wall.
• Preprocessing steps masked extra-pleural structures to standardize nodule appearance. Nodule candidates were identified through thresholding and morphological operations.
• A support vector machine (SVM) classifier analyzed 24 features (intensity, shape, context) for candidate classification, with evaluation via ten-fold cross-validation on the LIDC database.

Main Results:

• The proposed CAD system achieved a sensitivity of 98.3% for large pulmonary nodules.
• A secondary sensitivity of 94.1% was recorded for large nodules.
• The system demonstrated an average of 4.0 and 1.0 false positives per scan at these sensitivity levels.

Conclusions:

• The developed CAD system effectively detects the majority of large pulmonary nodules in thoracic CT scans.
• The system shows promise for identifying highly suspicious lesions with a low false positive rate, aiding clinical diagnosis.