Felicia Tang, Timothy Chen, Sayedomid Ebrahimzadeh, Brandon K K Fields, Yoo Jin Lee, Jonathan Liu, Adam Yen, Kiara Bowers, Richard B Schonour, Pan Su, Pedro Itriago Leon, Omar Darwish, Peder Larson, Yang Yang, Jae Ho Sohn
Purpose To assess image quality and detection performance of 0.55-T MRI for common lung pathologies in a multireader study. Materials and Methods Twenty-eight participants were prospectively enrolled at a tertiary academic medical center between January 1, 2023, and August 31, 2023, and underwent same-day chest CT and 0.55-T MRI with respiratory-triggered T2-weighted axial BLADE and T1-weighted ultrashort echo time sequences. Six radiologists evaluated the quality of the MRI using a Likert-type scale and assessed common lung abnormalities. Sensitivity was calculated as the proportion of CT-confirmed lesions detected on MRI and was compared between readers by training status using the Kruskal-Wallis test. Fleiss κ values were calculated for interreader agreement. Nodule measurements were compared between modalities using the Pearson correlation coefficient. Results Twenty-eight participants (mean age ± SD, 59 years ± 19; 17 female) were included. The overall quality of 80% (134 of 168) of the images was scored as good or higher. 0.55-T MRI achieved sensitivities of at least 85% for consolidations and large solid nodules (≥6 mm) and less than 40% for emphysema, ground-glass nodules, and small nodules (<6 mm). Readers achieved moderate agreement in lesion detection (κ, 0.51; 95% CI: 0.37, 0.61), with no evidence of a difference in sensitivity when stratified by training level (P = .56). Nodule size at CT and MRI was strongly correlated (r = 0.79-0.95; P < .001). Conclusion Consolidation and large solid nodules were well detected on 0.55-T MRI, whereas the detection performance for small nodules, ground-glass nodules, and emphysema was limited. Keywords: MR-Imaging, MR-Diffusion Weighted Imaging, Thorax, Lung, Mediastinum, Pleura, Pulmonary Arteries, Imaging Sequences, Observer Performance, Technology Assessment Supplemental material is available for this article. © RSNA, 2026.