Morpho-Functional 1H-MRI of the Lung in COPD: Short-Term Test-Retest Reliability

Purpose Non-invasive end-points for interventional trials and tailored treatment regimes in chronic obstructive pulmonary disease (COPD) for monitoring regionally different manifestations of lung disease instead of global assessment of lung function with spirometry would be valuable. Proton nuclear magnetic resonance imaging (1H-MRI) allows for a radiation-free assessment of regional structure and function. The aim of this study was to evaluate the short-term reproducibility of a comprehensive morpho-functional lungMRI protocol in COPD. Materials and Methods 20 prospectively enrolled COPD patients (GOLD I-IV) underwent 1H-MRI of the lung at 1.5T on two consecutive days, including sequences for morphology, 4D contrast-enhanced perfusion, and respiratory mechanics. Image quality and COPD-related morphological and functional changes were evaluated in consensus by three chest radiologists using a dedicated MRI-based visual scoring system. Test-retest reliability was calculated per each individual lung lobe for the extent of large airway (bronchiectasis, wall thickening, mucus plugging) and small airway abnormalities (tree in bud, peripheral bronchiectasis, mucus plugging), consolidations, nodules, parenchymal defects and perfusion defects. The presence of tracheal narrowing, dystelectasis, pleural effusion, pulmonary trunk ectasia, right ventricular enlargement and, finally, motion patterns of diaphragma and chest wall were addressed. Results Median global scores [10(Q1:8.00; Q3:16.00) vs. 11(Q1:6.00; Q3:15.00)] as well as category subscores were similar between both timepoints, and kappa statistics indicated "almost perfect" global agreement (\(\kappa\)= 0.86, 95%CI = 0.81-0.91). Most subscores showed at least "substantial" agreement of MRI1 and MRI2 (\(\kappa\)= 0.64-1.00), whereas the agreement for the diagnosis of dystelectasis/effusion (\(\kappa\)= 0.42, 95%CI = 0.00-0.93) was "moderate" and of tracheal abnormalities (\(\kappa\)= 0.21, 95%CI = 0.00-0.75) "fair". Most MRI acquisitions showed at least diagnostic quality at MRI1 (276 of 278) and MRI2 (259 of 264). Conclusion Morpho-functional 1H-MRI can be obtained with reproducible image quality and high short-term test-retest reliability for COPD-related morphological and functional changes of the lung. This underlines its potential value for the monitoring of regional lung characteristics in COPD trials