Repeatability and Reproducibility of Quantitative Whole-lung Perfusion Magnetic Resonance Imaging

Abstract

Magnetic resonance imaging (MRI) allows for quantitative evaluation of pulmonary perfusion and has shown high clinical usefulness for the evaluation and differentiation of different lung pathologies. The reproducibility of quantitative analysis of whole-lung perfusion has not been investigated previously. Our aim was to assess the intraobserver and interobserver repeatability and reproducibility of perfusion MRI to prove the concept that perfusion is suitable for therapy monitoring. The study was approved by the International Review Board. Fourteen healthy volunteers were examined using a time-resolved FLASH 3-dimensional perfusion sequence (1.5-T MRI, TREAT, GRAPPA 2, coronal orientation, voxel size 3.9×3.9×6.3 mm(3)). Perfusion was assessed initially and after 24 hours during an inspiratory and an expiratory breath hold. For each examination, 0.05 mmol/kg BW of Gd-DTPA was injected. Perfusion parameters such as pulmonary blood flow (PBF), pulmonary blood volume, and mean transit time were calculated. The evaluation was performed independently by 2 blinded observers. Intraobserver and interobserver differences were determined. The intraobserver differences between the initial and follow-up examinations for pulmonary blood volume, mean transit time, and time to peak were not significantly different for observers 1 and 2. PBF showed a significant difference for both observers only on inspiration (P<0.006 for observer 1 and P<0.009 for observer 2). For interobserver evaluation, all parameters, except inspiratory PBF, were significantly different (P<0.0001). Intraobserver quantitative perfusion MRI showed reproducible results. However, the evaluation is highly dependent on the observer. Therefore, quantitative analysis of the serial examinations should be performed by the same observer.