Low-dose computed tomography (LDCT) has become a well-established approach for lung cancer screening in individuals at high risk, particularly those with a history of heavy smoking and patients diagnosed with chronic obstructive pulmonary disease (COPD). However, repeated LDCT scans over time raise concerns about cumulative radiation exposure, which may inadvertently increase the risk of malignancy. As a radiation-free imaging modality, magnetic resonance imaging (MRI) is increasingly being explored for pulmonary applications.
A nationwide multicentre trial was conducted to assess the longitudinal performance of MRI compared to LDCT for evaluating incidental pulmonary nodules in patients with or at risk of COPD. Over a three-year interval, the study examined whether MRI could serve as a reliable and safe alternative for monitoring nodules in this high-risk population.
MRI Detection and Growth Monitoring of Pulmonary Nodules
The study involved 239 participants from 16 centres, all of whom underwent the same-day LDCT and MRI during two imaging rounds spaced approximately three years apart. In total, 240 nodules were detected on the initial LDCT scan. Of these, 224 nodules persisted across the second round, representing over 93% of the original findings. Eight patients also developed new nodules between the two rounds. MRI demonstrated a detection sensitivity of 82.8% and 81.5% for the two readers, respectively, when compared against LDCT as the standard of reference.
Despite the moderate sensitivity for identifying all nodules, MRI achieved excellent agreement with LDCT when assessing nodule growth, defined as a size increase of more than 2 mm. In this respect, MRI yielded a kappa statistic of up to 0.90, indicating high concordance in classifying nodules as progressed, stable or regressed. Axis measurements of nodules between the two MRI rounds were consistent across both readers, with minimal variance and no statistically significant discrepancies. Importantly, there was also strong inter-method agreement between MRI and LDCT for these longitudinal measurements. These findings suggest that while MRI may be less sensitive in detecting all nodules, particularly those smaller than 6 mm, it performs reliably in tracking changes over time, a key factor in identifying potential malignancies.
Morphological and Risk Classification Capabilities of MRI
The ability to evaluate changes in nodule morphology is essential for assessing malignancy risk. MRI demonstrated consistent sensitivity across both imaging rounds in identifying features such as spiculation, cavitation, lobulation and location relative to pleural and fissural structures. Agreement between readers for these morphological changes ranged from moderate to excellent. Notably, the contrast-enhanced T1-weighted sequences offered superior detection and characterisation of nodules, suggesting that this sequence alone may suffice in certain screening contexts, thereby reducing overall MRI examination time.
Must Read: Pulmonary Nodule Risk Estimation With Deep Learning and Uncertainty Analysis
Nodules were also categorised using the Lung-RADS classification system, which integrates size and morphology to stratify malignancy risk. When assessed per nodule, MRI showed substantial agreement with LDCT (κ = 0.62–0.70), and this agreement improved to excellent (κ = 0.86–0.88) when evaluated per patient. This per-patient approach relies on identifying the most suspicious nodule in each individual, aligning closely with how clinical decisions are made. MRI’s ability to detect diameter changes and morphological evolution supported accurate Lung-RADS classification, reinforcing its role in guiding follow-up strategies and potential interventions.
Prognostic Accuracy and Implications for Cancer Detection
The study also assessed MRI’s prognostic capabilities by comparing Lung-RADS classifications at baseline with observed nodule progression. MRI demonstrated slightly higher accuracy in predicting which nodules would grow or transform, as indicated by Lung-RADS categories ≥3. While LDCT achieved an accuracy of 15.8% in identifying progressing nodules at baseline, MRI’s accuracy reached 23.1% and 21.4% for the two readers. These findings imply that MRI, when used longitudinally, may offer greater predictive value than single timepoint assessments.
Out of the participants, 11 patients underwent surgical procedures for nodules deemed suspicious, with histological confirmation of lung cancer in each case. In a pooled analysis of both imaging rounds, MRI and LDCT both identified seven out of eight patients with confirmed malignancies as having Lung-RADS category ≥3. These results demonstrate that MRI’s diagnostic accuracy in a per-patient context is comparable to that of LDCT, supporting its application in routine monitoring of high-risk patients.
The study illustrated that standardised morpho-functional MRI is a promising alternative to LDCT for the longitudinal follow-up of incidental pulmonary nodules in patients with or at risk of COPD. While MRI exhibits moderate sensitivity in detecting all nodules, particularly small ones, its accuracy in measuring growth, evaluating morphology and applying Lung-RADS classification closely mirrors that of LDCT over time. These findings suggest that MRI could serve as a radiation-free tool for ongoing surveillance, particularly valuable in reducing long-term imaging risks in high-risk populations. Further research is warranted to expand these findings to broader clinical settings and evaluate the use of MRI across diverse patient cohorts.
Source: European Radiology
Image Credit: iStock
References:
Zhu L, Li Q, von Stackelberg O et al. (2025) Longitudinal MRI in comparison to low-dose CT for follow-up of incidental pulmonary nodules in patients with COPD—a nationwide multicenter trial. Eur Radiol.
Image:
