PurposeMonitoring of regional lung function in interventional COPD trials requires alternative endpoints beyond global parameters such as FEV1. T1 relaxation times of the lung might allow to draw conclusions on tissue composition, blood volume and oxygen fraction. The aim of this study was to evaluate the potential value of lung Magnetic resonance imaging (MRI) with native and oxygen-enhanced T1 mapping for the assessment of COPD patients in comparison with contrast enhanced perfusion MRI.Materials and Methods20 COPD patients (GOLD I-IV) underwent a coronal 2-dimensional inversion recovery snapshot flash sequence (8 slices/lung) at room air and during inhalation of pure oxygen, as well as dynamic contrast-enhanced first-pass perfusion imaging. Regional distribution of T1 at room air (T1), oxygen-induced T1 shortening (ΔT1) and peak enhancement were rated by 2 chest radiologists in consensus using a semi-quantitative 3-point scale in a zone-based approach.ResultsAbnormal T1 and ΔT1 were highly prevalent in the patient cohort. T1 and ΔT1 correlated positively with perfusion abnormalities (r = 0.81 and r = 0.80; p&0.001), and with each other (r = 0.80; p<0.001). In GOLD stages I and II ΔT1 was normal in 16/29 lung zones with mildly abnormal perfusion (15/16 with abnormal T1). The extent of T1 (r = 0.45; p<0.05), ΔT1 (r = 0.52; p<0.05) and perfusion abnormalities (r = 0.52; p<0.05) showed a moderate correlation with GOLD stage.ConclusionNative and oxygen-enhanced T1 mapping correlated with lung perfusion deficits and severity of COPD. Under the assumption that T1 at room air correlates with the regional pulmonary blood pool and that oxygen-enhanced T1 reflects lung ventilation, both techniques in combination are principally suitable to characterize ventilation-perfusion imbalance. This appears valuable for the assessment of regional lung characteristics in COPD trials without administration of i.v. contrast.

The evaluation of the developed methods indicates good accuracy and shows that automatically generated lung masks differ from expert segmentations about as much as segmentations from different experts.