Interobserver variability in Lung CT Screening Reporting and Data System categorisation in subsolid nodule-enriched...

Area of Science:

• Radiology
• Medical Imaging
• Thoracic Imaging

Background:

• Low-dose computed tomography (LDCT) screening is crucial for lung cancer detection.
• Subsolid nodules on LDCT require careful categorization using systems like Lung Imaging Reporting and Data System (Lung-RADS).
• Interobserver agreement in categorizing these nodules is essential for consistent patient management.

Purpose of the Study:

• To evaluate the interobserver agreement among radiologists when categorizing subsolid nodules using Lung-RADS on LDCT scans.
• To identify factors influencing the accuracy and consistency of Lung-RADS categorization.

Main Methods:

• Retrospective review of 162 LDCT scans enriched with subsolid nodules.
• Categorization of scans by five radiologists using Lung-RADS criteria.
• Calculation of concordance and discordance rates (minor and major).
• Analysis of interobserver agreement using Cohen's kappa statistics.

Main Results:

• Overall accurate Lung-RADS categorization was 60.3% for all cases and 45.0% for positive screens.
• Minor and major discordance rates were 12.3% and 27.3% overall, respectively.
• Interobserver agreement was moderate (mean kappa = 0.45).
• Experienced radiologists demonstrated higher concordance rates compared to residents.
• A higher proportion of part-solid nodules was associated with lower accurate categorization.

Conclusions:

• Interobserver agreement in Lung-RADS categorization of subsolid nodules is variable and moderate.
• Radiologist experience significantly influences the accuracy and consistency of Lung-RADS categorization.
• The presence of part-solid nodules may contribute to categorization discrepancies, potentially affecting lung cancer screening management recommendations.