Rationale and objectives: We examined the effect of varying window settings and contrast enhancement on detecting pulmonary abnormalities on magnetic resonance (MR) images in patients with usual interstitial pneumonitis (UIP).
Methods: HRCT scans and MR images from 10 patients with UIP were evaluated. T1-weighted MR images were obtained before and after administration of gadopentetate dimeglumine and were photographed at conventional windows and at windows chosen to increase the conspicuity of the lung parenchyma ("lung windows"). The four MR image configurations were mixed with the high-resolution computed tomography (HRCT) scans of these patients and randomized. Corresponding scan levels in each patient were evaluated conjointly by two thoracic radiologists for the presence of "honeycomb lung," ground-glass opacity, parenchymal bands, and reticular abnormalities. Lung signal intensity in areas containing ground-glass signal intensity was measured on MR images using regions of interest.
Results: The administration of gadopentetate dimeglumine significantly improved the detection of honeycomb lung on scans photographed at lung windows, but it did not significantly influence the detection of ground-glass abnormalities. The use of lung windows improved the detection of ground-glass abnormalities for both enhanced and unenhanced scans, but lung windows improved the detection of honeycomb lung only for enhanced scans. All MR image configurations were insensitive compared with HRCT scans for detecting parenchymal bands and reticular abnormalities. There was a good correlation between measured lung signal intensity and visual ground-glass profusion score.
Conclusion: The visibility of pulmonary abnormalities on MR images of patients with UIP is limited compared with that of HRCT scans. The improved visibility of some parenchymal abnormalities after intravenous administration of gadopentetate dimeglumine and with the use of lung windows is insufficient to warrant their routine use in thoracic MR imaging.