Purpose To assess the repeatability of real-time cine pulmonary MRI measures of metronome-paced tachypnea (MPT)-induced dynamic hyperinflation and its relationship with chronic obstructive pulmonary disease (COPD) severity. Materials and Methods SubPopulations and InteRmediate Outcome Measures In COPD Study (SPIROMICS) (ClinicalTrials.gov identifier no. NCT01969344) is a multicenter prospective cohort study that recruited individuals with COPD with 20 or more pack-years smoking history and controls aged 40-80 years. SPIROMICS-HF is a cross-sectional study evaluating cardiopulmonary interactions in COPD from December 2019 to April 2024. Two-dimensional coronal real-time cine pulmonary MRI (3.2 frames/sec) was performed during MPT twice. Lung masks and lung MR images from an independent study were used for transfer learning to segment real-time cine pulmonary MR images. Dynamic hyperinflation was evaluated as the increase in end-expiratory lung volume (EELV) during tidal breathing to the end of MPT. Repeatability was assessed with intraclass correlation coefficients, and multivariable associations with COPD severity were examined. Results Of the 70 participants (mean age, 67 years ± 10 [SD]; 37 male, 33 female) included in the study, 59% had COPD. The transfer learning model achieved high accuracy in lung segmentation (Dice similarity coefficient, 0.94 ± 0.03). There was good scan-rescan agreement for EELV and dynamic hyperinflation (intraclass correlation coefficient, 0.99 and 0.87, respectively). Dynamic hyperinflation was associated with COPD severity (P trend = .01, with a mean difference between severe COPD and controls of 0.24 L). Conclusion A transfer learning model yielded reproducible quantification of MPT-induced dynamic hyperinflation at real-time cine pulmonary MRI, with greater dynamic hyperinflation in participants with more severe COPD. Keywords: Chronic Obstructive Pulmonary Disease, MR Imaging, Pulmonary, Lung, Technology Assessment Supplemental material is available for this article. ClinicalTrials.gov identifier: NCT01969344 © RSNA, 2025.
Keywords: Chronic Obstructive Pulmonary Disease; Lung; MR Imaging; Pulmonary; Technology Assessment.