Interobserver reliability in describing radiographic lung changes after stereotactic body radiation therapy

Area of Science:

• Radiology
• Radiation Oncology
• Pulmonary Medicine

Background:

• Radiographic lung changes post-stereotactic body radiation therapy (SBRT) exhibit significant patient variability.
• Standardized classification systems for acute (≤6 months) and late (>6 months) benign lung changes have been proposed.
• The reliability of applying these classification systems for lung changes after SBRT requires validation.

Purpose of the Study:

• To evaluate the interobserver reliability of classifying acute and late radiographic lung changes following thoracic SBRT.
• To assess the consistency of applying proposed standardized descriptions for post-SBRT lung abnormalities.

Main Methods:

• Analysis of 280 follow-up CT scans from 100 patients treated with thoracic SBRT at 3, 6, and 12 months post-treatment.
• Five physicians independently and blindly classified acute lung changes (3- and 6-month scans) using defined categories (e.g., consolidation, ground glass opacity, no change).
• Late lung changes (12-month scans) were classified into patterns (e.g., modified conventional, mass-like, scar-like, no change), with interobserver agreement quantified using Fleiss' kappa scores.

Main Results:

• Interobserver agreement (kappa scores) for classifying lung changes was fair: 0.30 at 3 months, 0.20 at 6 months, and 0.25 at 12 months.
• Distribution of acute changes at 3/6 months: diffuse consolidation (11%/21%), patchy consolidation/GGO (15%/28%), diffuse GGO (10%/11%), patchy GGO (15%/15%), no change (49%/25%).
• Distribution of late changes at 12 months: modified conventional (46%), mass-like (16%), scar-like (26%), no change (12%). Uniform scoring occurred in only 26, 8, and 14 cases at 3, 6, and 12 months, respectively.

Conclusions:

• Interobserver reliability for classifying radiographic lung changes post-SBRT indicates only fair agreement.
• Qualitative descriptions are insufficient for consistent categorization, as most patient scans do not fit clearly into a single classification.
• The 6-month time point presents the most classification difficulty due to the potential co-occurrence of acute and late lung changes.