Artificial Intelligence-Aided Diagnosis Software to Identify Highly Suspicious Pulmonary Nodules

Area of Science:

• Medical Imaging
• Artificial Intelligence in Radiology
• Pulmonary Diagnostics

Background:

• Lung nodules are a significant concern in pulmonary diagnostics.
• Accurate diagnosis of lung nodules is crucial for effective treatment planning.
• Current imaging techniques have limitations in fully characterizing lung nodules.

Purpose of the Study:

• To evaluate the diagnostic value of AI-assisted software for lung nodules.
• To compare the efficacy of combined low-dose computed tomography (LDCT) and high-resolution computed tomography (HRCT) with conventional methods.
• To assess AI's role in improving nodule visibility and malignancy rate determination.

Main Methods:

• 113 patients with pulmonary nodules underwent LDCT screening.
• HRCT local-target scanning (combined scheme) and conventional CT were performed for larger nodules.
• AI-assisted software analyzed nodule size and malignancy probability, comparing two imaging schemes.

Main Results:

• Significant differences in nodule volume and malignancy probability for subsolid nodules were observed between groups with improved and identical visibility.
• The combined scanning protocol showed significant between-group differences in subsolid nodule malignancy rates.
• AI analysis revealed variations in nodule characteristics based on imaging scheme effectiveness.

Conclusions:

• AI-assisted software, coupled with a combined scanning scheme, shows promise in lung nodule diagnosis.
• The combined LDCT and HRCT approach may improve the characterization of subsolid nodules.
• AI-driven analysis can aid in identifying high-risk populations for lung cancer screening.