Related Experiment Video
Updated: Jun 14, 2025
Imaging Features of Systemic Sclerosis-Associated Interstitial Lung Disease
Published on: June 16, 2020
Use of artificial intelligence algorithms to analyse systemic sclerosis-interstitial lung disease imaging features
Jing Zhao1, Ying Long2,3, Shengtao Li4
1Department of Rheumatology, People's Hospital of Xiangxi Tujia and Miao Autonomous Prefecture (The First Affiliated Hospital of Jishou University), Intersection of Shiji Avenue and Jianxin Road, Jishou, 416000, Hunan, People's Republic of China.
Artificial intelligence (AI) effectively analyzes high-resolution computed tomography (HRCT) for diagnosing systemic sclerosis-associated interstitial lung disease (SSc-ILD). AI reveals imaging patterns correlating with clinical features and prognosis, aiding in SSc-ILD assessment.
Area of Science:
- Radiology
- Artificial Intelligence
- Pulmonology
Background:
- Systemic sclerosis-associated interstitial lung disease (SSc-ILD) diagnosis and prognosis are challenging.
- High-resolution computed tomography (HRCT) is crucial for SSc-ILD assessment, but AI's role is underexplored.
- Understanding SSc-ILD imaging characteristics is vital for patient management.
Purpose of the Study:
- To analyze lung HRCT images of SSc-ILD patients using AI.
- To correlate AI-derived imaging features with clinical manifestations and prognosis.
- To explore the distinct imaging features of diffuse SSc-ILD (dSSc-ILD) and limited SSc-ILD (lSSc-ILD).
Main Methods:
- Collected 72 lung HRCT images and clinical data from 58 SSc-ILD patients.
- Utilized AI to identify and evaluate ILD lesion type, location, and volume on HRCT.
- Performed statistical analysis on imaging characteristics of dSSc-ILD and lSSc-ILD, and correlated findings with clinical indicators and prognosis.
Main Results:
- AI efficiently analyzed SSc-ILD imaging characteristics, showing potential for generalization.
- dSSc-ILD and lSSc-ILD showed different prevalence based on disease duration (<1 year vs. ≥5 years).
- Common ILD types included NSIP, UIP, and unclassifiable IIP; lesion diversity increased with disease progression. Ground-glass opacity was absent in SSc-ILD with pulmonary hypertension.
Conclusions:
- AI demonstrates significant potential in analyzing complex HRCT images for SSc-ILD.
- AI-driven analysis can reveal imaging patterns linked to SSc-ILD subtypes and clinical outcomes.
- This approach may enhance the diagnostic accuracy and prognostic assessment of SSc-ILD.
