Assessment of geometric distortions and dosimetric impact from implementing 2D-T2-PROPELLER and 3D-T1-LAVA-FLEX in an...

Magnetic Resonance Imaging (MRI) is the gold standard image guidance of cervical brachytherapy treatments. Many clinics are moving towards MRI-only planning procedures. In this work, the reconstructions using images acquired from the commercially-available 2D, T2- weighted PROPELLER sequence, and 3D, T1-weighted LAVA-Flex sequences are compared to reconstructions on CT images and assessed for geometric and dosimetric accuracy.

A cohort of 11 cervical brachytherapy patients was obtained during the commissioning of a new MRI-only gynecological brachytherapy program. Patients underwent CT, 2D-T2-PROPELLER, and T1-3D-LAVA-Flex image acquisition. Geometric distortions in the MRI reconstructed images were evaluated. The applicator and catheters for each image were independently reconstructed in the Oncentra treatment planning system (Elekta), and plans were manually copied to each imaging modality. A gamma analysis using acceptance criteria of 3%, 2mm was used to access the dosimetric consistency between the reconstructions on MRI versus CT.

An average geometric distortion of 2.4mm was measured for both the 2D-T2-PROPELLER and T1-3D-LAVA-Flex sequence reconstructions; maximum distortions were 7.5 and 6.4mm respectively. Larger geometric distortions are attributed to obtaining measurements on the periphery of the field-of-view and not in the region of applicator and catheter reconstruction. Gamma pass rates of 96.1±2.8% and 96.6±3.1% were found for the 2D-T2-PROPELLER and 3D-T1- LAVA-Flex MRI reconstructions when compared to CT reconstructions. Minimum gamma pass rates were 90.2% and 90.5% respectively. Significant differences in gamma pass rates were not found between T2-2D-PROPELLER and T1-3D-LAVA-FLEX; however, T1-3D-LAVA-FLEX results were skewed towards higher pass rates. T1-3D-LAVA-FLEX also performed significantly better than T2-2D-PROPELLER for a larger number of catheters.