Automated Volumetric Quantification of Emphysema Severity by Using Ultrashort Echo Time MRI: Validation in Participants with Chronic Obstructive Pulmonary Disease

Background The validity of three-dimensional (3D) ultrashort echo time (UTE) MRI for the assessment of emphysema in patients with chronic obstructive pulmonary disease (COPD) at high spatial resolution is, to the knowledge of the authors, unknown. Purpose To assess whether noncontrast agent–enhanced 3D UTE MRI at submillimeter spatial resolution can be used to determine the extent of emphysema by using both qualitative visual scoring and fully automated volumetric quantification. Materials and Methods Twenty-eight participants with COPD and 10 control participants (mean age, 70 years ± 7 [standard deviation] and 64 years ± 4, respectively) were prospectively enrolled between 2015 and 2017. Participants underwent pulmonary function testing, CT, and MRI. CT was used as the reference standard. Qualitative scoring of emphysema extent was performed by two readers. Fully automated quantification of percentage of low-attenuation volume by using a threshold of −950 HU (%LAV−950HU) at CT and percentage of low-signal-intensity volume by using an adaptive threshold of 0.20 (%LSV0.20) at MRI were the respective emphysema indexes. Comparison of means was performed by using Student t test, correlation was determined by using Pearson test, agreement was found by using weighted κ index, and reproducibility was determined by using intraclass correlation coefficient. Diagnostic performance was assessed by calculating the area under the receiver operating characteristics curve (AUC). Results With qualitative scoring, agreement between UTE MRI and CT was good (weighted κ, 0.79; 95% confidence interval: 0.71, 0.83). With automated volumetric quantification, %LSV0.20 was significantly correlated with %LAV−950HU in participants with COPD (r, −0.80; P < .001) and correlated with forced expiratory volume in 1 second percentage predicted (r, −0.55; P = .002). %LSV0.20 was significantly higher in participants with COPD than in control participants (P < .001). The diagnostic performance and reproducibility of %LSV0.20 were good (AUC, 1.00 [95% confidence interval: 0.88, 1.00], and intraclass correlation coefficient, > 0.99, respectively). Conclusion A fully automated method with three-dimensional ultrashort echo time MRI reproducibly quantified the volumetric extent of emphysema in participants with chronic obstructive pulmonary disease. © RSNA, 2019 Online supplemental material is available for this article.