An Exploratory Radiomics Approach to Quantifying Pulmonary Function in CT Images

Area of Science:

• Quantitative medical imaging
• Radiomics
• Pulmonary function assessment

Background:

• Medical imaging is increasingly quantitative, with radiomics emerging to extract data from images.
• Radiomics transforms unstructured image data into structured data for phenotype characterization.
• Pulmonary function assessment traditionally relies on spirometry and gas exchange tests.

Purpose of the Study:

• To explore the association between quantitative radiomic features from CT images and pulmonary function.
• To identify potential radiomic biomarkers for pulmonary function.
• To investigate if radiomic patterns can stratify patients based on lung function.

Main Methods:

• Extracted 39 radiomic features from lung CT images of 64 patients.
• Features captured lung morphology, intensity variations, and tissue texture.
• Compared radiomic data with conventional pulmonary function tests (spirometry, diffusing capacity).

Main Results:

• Larger lungs with homogeneous, low-attenuating tissue (radiomic measures) associated with poorer spirometry and lower diffusing capacity for carbon monoxide.
• Unsupervised clustering identified patient subsets with similar radiomic patterns and comparable forced expiratory volume in one second (FEV1) measurements.
• Radiomic feature vectors correlated with spirometry performance, indicating imaging-based stratification of pulmonary function.

Conclusions:

• Radiomics can identify imaging phenotypes associated with pulmonary function.
• Quantitative imaging features may serve as biomarkers for lung function.
• Radiomic analysis offers potential for non-invasive assessment and stratification of pulmonary disease.