AI-Based HRCT Quantification in Connective Tissue Disease-Associated Interstitial Lung Disease

Area of Science:

• Radiology
• Artificial Intelligence
• Pulmonology

Background:

• Interstitial lung disease (ILD) is a common, progressive complication in connective tissue diseases (CTDs).
• Accurate quantification of lung abnormalities on high-resolution computed tomography (HRCT) is crucial for treatment monitoring, especially with antifibrotic therapies.
• Artificial intelligence (AI) may enhance consistency in visual assessment and aid radiologists in longitudinal follow-up.

Purpose of the Study:

• To evaluate the performance of an AI-based quantitative imaging tool in assessing ILD patterns on HRCT.
• To compare AI interpretation with expert and non-expert radiologists in patients with CTD-related ILD.
• To assess AI's ability to detect changes during antifibrotic treatment follow-up.

Main Methods:

• Retrospective analysis of 48 patients with CTD-related ILD undergoing antifibrotic treatment.
• Four HRCT scans per patient evaluated by an expert and a non-expert radiologist using semi-quantitative scoring for four patterns (hyperlucency, GGO, reticulation, honeycombing).
• AI analysis using Imbio Lung Texture Analysis platform for volumetric percentages of each pattern; comparison of AI with human readers via concordance and Mean Absolute Error (MAE).

Main Results:

• AI showed high concordance (81%) with the expert radiologist, with low MAE (1.8%-2.6%).
• AI concordance with the non-expert radiologist was significantly lower (60-70%) with higher MAE (3.9%-5.2%).
• AI demonstrated superior alignment with expert interpretation (p < 0.01) and effectively detected subtle changes during follow-up, especially when visual assessment was inconsistent.

Conclusions:

• AI-driven quantitative imaging matches expert radiologists in assessing ILD patterns on HRCT.
• AI significantly outperforms less experienced readers, offering improved reproducibility and sensitivity to change.
• AI can standardize ILD follow-up, aid multidisciplinary decisions, and assist in managing progressive fibrosing CTD-ILDs on antifibrotic therapy.