Comparison of MRI and CT in Detecting Pulmonary Lesions: Consistency and Prospects for Application

Background: Computed tomography (CT) is the primary tool for lung examinations, but it poses radiation risks to sensitive populations, such as children and adolescents. Advances in Magnetic Resonance Imaging(MRI) technology have made radiation-free lung examinations feasible. However, the heterogeneity in the presentation of lung lesions with varying densities (solid and subsolid lesions) across different MRI sequences challenges clinicians and radiologists in selecting the most appropriate sequence. This study aims to systematically compare the consistency of MRI with CT for pulmonary lesions and analyze the appearance of solid and subsolid lesions across different MRI sequences to guide the optimization and selection of MRI imaging strategies. Methods: This study involved 45 patients (age 61.65±12.72years) with 46 suspected lung malignancies, scanned between November 2023 and October 2024. MRI scans, done within a week of CT, used utilizing respiratory-triggered zero echo time (ZTE) (RTr ZTE), breath-hold ZTE (BH ZTE), T1 LAVA-Flex, and RTr T2WI-FS sequences. CT findings were the benchmark for analyzing lesion morphological features (diameter, lobulation, spiculation, vacuole sign) and for comparing image quality of the two ZTE sequences. The Mann-Whitney U, Cohen's Kappa, and Wilcoxon signed-rank tests were used to assess statistical differences. Results: MRI showed high consistency with CT in detecting solid lesion diameters, with no significant sequence differences. The RTr ZTE sequence demonstrated the closest agreement with CT, averaging a 0.6 mm bias. For subsolid lesions, ZTE sequences showed no significant difference from CT (p>0.05), while T1 LAVA-Flex and RTr T2WI-FS significantly underestimated lesion diameters (p<0.05). All MRI sequences were consistent with CT in displaying lobulation and vacuole signs (p>0.05). However, only the RTr ZTE sequence showed comparable performance to CT in visualizing spiculation (p=0.011), while other sequences exhibited significant differences (p<0.05). In image quality, RTr ZTE outperformed BH ZTE significantly (p<0.05), with good inter-observer agreement (p>0.05). Conclusion: The RTr ZTE sequence correlates closely with CT for lung lesion detection, particularly subsolid lesions and spiculation, suggesting it potentially serving as a substitute for CT in the long-term follow-up of radiation-sensitive populations. Clinical trial number: Not applicable.