Diagnostic performance of a single breath-hold lung MRI scan with AI-powered compressed sensing for nodule detection in comparison to photon counting detector-CT
Randomized trial evaluates lung MRI's effectiveness for nodule detection, suggesting a potential radiation-free alternative to CT.
Key Points
The study aims to assess the diagnostic performance of a single-breath-hold lung MRI sequence for nodule detection compared to photon-counting detector CT.
Prospective design involving 148 healthy adults for lung 3-T MRI and PCD-CT comparison.
Used AI-aided compressed sensing with an acceleration factor of 9 for MRI scanning.
Evaluated Lung-RADS scores and agreement using Lin’s Concordance Correlation Coefficient.
MRI achieved an overall sensitivity of 83.1% for solid nodules, increasing to 98.1% for nodules ≥ 4 mm.
MRI and CT had complete agreement in Lung-RADS 4A and 4B categories.
Lin’s CCC for nodule size was 0.985, indicating excellent agreement, with slight underestimation in MRI.