Choice of Segmentation Algorithm to Determine Ventilated Volume Fraction from Oxygen-Enhanced MRI in Cystic Fibrosis and Healthy Volunteers

Motivation: We aim to refine lung ventilation quantification using R2* based oxygen-enhanced MRI (OE-MRI) in cystic fibrosis (CF). Goal(s): This study seeks to evaluate various segmentation methods for calculating ventilated volume fraction (VVF) from OE-MRI and to assess their correlation with lung function metrics in CF and healthy volunteers. Approach: We applied multiple segmentation algorithms and compared their repeatability in healthy volunteers, and correlation with the lung clearance index (LCI2.5) in CF patients. Results: Several methods studied demonstrated good repeatability and good correlation with LCI2.5. Only one method found a significant difference in VVF between CF and HV. Impact: This study provides a direct comparison of segmentation methods for quantifying ventilated lung volume in cystic fibrosis from dynamic oxygen-enhanced MRI. Our results could improve the assessment of lung function in patients, by improving repeatability and sensitivity to disease.