Diagnostic accuracy of ultrashort echo times-MRI in patients with and without idiopathic pulmonary fibrosis.

Abstract

High-resolution computed tomography (HRCT) plays a key role in the diagnosis of interstitial lung diseases (ILD). Lung MRI with ultrashort echo time (UTE-MRI) is a promising alternative, yet its accuracy in identifying usual interstitial pneumonia (UIP) patterns, a key feature of idiopathic pulmonary fibrosis (IPF) management, remains unknown. This study aimed to evaluate UTE-MRI's diagnostic accuracy for UIP hallmarks compared to HRCT. A monocentric prospective study was conducted between 2017 and 2020. All patients underwent HRCT (reference test) and UTE-MRI (index test). Diagnostic accuracy was assessed for the main UIP hallmarks, i.e., honeycombing, reticulation with or without bronchiolectasis, subpleural and basal distribution and the absence of abnormalities suggestive of an alternative diagnosis. A comprehensive evaluation of UIP and non-UIP patterns was also conducted. Sixty patients were included (median age = 64 years; 46 males), with (n = 25) and without IPF (n = 35). UTE-MRI demonstrated high (sensitivity/specificity) for honeycombing (86.9%/91.9%), reticulation with bronchiolectasis (88.0%/91.4%) or without bronchiolectasis (83.9%/89.7%), subpleural and basal distribution (88.0%/97.1%) and absence of alternative diagnoses (81.2%/77.8%), respectively. It accurately discriminated patients with UIP patterns versus alternative diagnoses with 100% (89.7; 100) sensitivity and 100% (86.7; 100) specificity. Intra- and inter-reader reproducibility was very good (κ ≥ 0.80). UTE-MRI can diagnose the main structural hallmarks of IPF with high sensitivity and specificity, and was found to be comparable to CT for discriminating patients with UIP patterns among other fibrosing or non-fibrosing ILDs. Future research should explore morpho-functional combinations for monitoring IPF progression. Question The diagnostic accuracy of UTE-MRI to diagnose the usual interstitial pneumonia (UIP) patterns is currently unknown. Findings UTE-MRI demonstrated high (sensitivity/specificity) for honeycombing (86.9%/91.9%), reticulation with (88.0%/91.4%) or without bronchiolectasis (83.9%/89.7%), and accurately discriminated patients with UIP patterns versus alternative diagnoses. Clinical relevant statement UTE-MRI can diagnose the main structural hallmarks of IPF with high sensitivity and specificity, and was found to be comparable to CT for discriminating patients with UIP patterns among other fibrosing or non-fibrosing ILDs.