Computer-aided quantification of interstitial lung disease from high resolution computed tomography images in...

Area of Science:

• Radiology
• Pulmonology
• Medical Imaging Analysis

Background:

• Systemic sclerosis (SSc) frequently leads to interstitial lung disease (ILD).
• Accurate quantification of ILD extent is crucial for patient management and clinical trials.
• Conventional visual reader-based scoring (CoVR) can be subjective.

Purpose of the Study:

• To assess the performance of a computerized-aided method (CaM) for ILD quantification in SSc patients.
• To correlate CaM findings with conventional visual scoring (CoVR) and pulmonary function tests (PFTs).
• To evaluate CaM as a reliable tool for SSc-ILD assessment.

Main Methods:

• Seventy-nine SSc patients underwent chest high-resolution computed tomography (HRCT).
• HRCT images were scored using CoVR by two radiologists and analyzed by CaM software.
• Linear regression, Pearson's correlation, and ROC analysis were employed.

Main Results:

• A strong correlation was found between CaM and CoVR (P < 0.0001).
• CaM demonstrated significant negative correlations with forced vital capacity (FVC) and diffusing capacity of the lung for carbon monoxide (DLco) (P < 0.0001).
• An optimal CaM extent threshold of 20% achieved 75.6% sensitivity and 97.4% specificity.

Conclusions:

• Computerized-aided method (CaM) offers reliable quantification of SSc-ILD extent.
• CaM shows strong agreement with visual scoring and PFTs.
• CaM is a promising tool for clinical practice and research in SSc-ILD.