Cyst-based measurements for assessing lymphangioleiomyomatosis in computed tomography

Area of Science:

• Pulmonary Medicine
• Radiology
• Medical Imaging Analysis

Background:

• Lymphangioleiomyomatosis (LAM) is a rare lung disease characterized by cystic lung changes.
• Accurate assessment of LAM severity is crucial for patient management and treatment.
• Current CT-based assessments may not fully capture disease severity.

Purpose of the Study:

• To evaluate a new set of quantitative measurements derived from individual pulmonary cysts on CT scans.
• To determine the efficacy of these novel measurements in assessing LAM severity.
• To compare the performance of new CT-based cyst measurements against existing density-based CT metrics.

Main Methods:

• Individual pulmonary cysts were extracted from CT scans of LAM patients using the watershed algorithm.
• A novel family of measurements quantifying cyst amount, distribution, and boundary appearance was computed.
• Feature selection identified key measurements, and linear regression compared their correlation with spirometry against traditional CT measures.

Main Results:

• The proposed CT-based cyst measurements demonstrated adjusted R-squared values ranging from 0.42 to 0.59 (p < 0.05) when correlated with spirometry.
• These results indicate a stronger correlation compared to previously used density-based CT measurements, which showed limited R-squared values (best R²=0.46, often <0.3).

Conclusions:

• The novel family of CT-based cyst measurements exhibits superior correlation with spirometric measures in LAM patients.
• These findings suggest that the proposed measurements are a promising, sensitive tool for quantitative assessment of LAM severity.