Automatic Quantification of Interstitial Lung Disease From Chest Computed Tomography in Systemic Sclerosis

Area of Science:

• Radiology and Imaging
• Pulmonary Medicine
• Rheumatology

Background:

• Interstitial lung disease (ILD) is a frequent complication of systemic sclerosis (SSc).
• Early diagnosis of SSc-ILD is crucial for timely treatment interventions.
• Current assessment methods can be subjective.

Purpose of the Study:

• To develop and validate an automated method for quantifying ILD in SSc patients from chest CT images.
• To differentiate between limited and extensive SSc-ILD based on quantitative CT analysis.

Main Methods:

• Ninety-four SSc patients and 73 controls underwent CT scans, FVC, and DLCO tests.
• An automated method calculated the ratio of interstitial opacity weight to total lung weight as an ILD indicator.
• Receiver operator characteristic (ROC) curve analysis determined normality cutoffs and SSc-ILD classification (Limited vs. Extensive).

Main Results:

• 83% of SSc patients were classified with SSc-ILD, with an optimal threshold of 23.4% (AUC 0.94).
• 38% of patients had SSc Extensive-ILD (Z score ≥ 3 or FVC < 70%), and 45% had SSc Limited-ILD.
• SSc Extensive-ILD showed significantly lower DLCO and total lung volume compared to SSc Limited-ILD.

Conclusions:

• The proposed automated quantification method offers an objective alternative for identifying and assessing ILD extent in SSc patients.
• This method can help mitigate the subjectivity associated with traditional visual scoring techniques.
• Accurate quantification aids in classifying SSc-ILD severity, potentially improving patient management.