DISCUSSION
To our knowledge, this is the first treatment planning study of craniospinal irradiation using the RefleXion X1 system. We have previously reported on treatment planning comparison between RefleXion X1 and Eclipse VMAT for 42 patients across 6 cancer sites [25]. In this study, we tested the feasibility of CSI using RefleXion X1. We have successfully generated clinically acceptable RefleXion CSI plans for 5 pediatric medulloblastoma patients with target length less than 50 cm. Dosimetric indices were comparable between RefleXion X1 and Eclipse VMAT modalities except for statistically significantly improved bowel sparing with RefleXion X1.
Due to 2 cm field size and long PTV CSI targets, the average beam-on time was approximately 4.5 times greater using RefleXion X1 compared to Eclipse VMAT. For VMAT CSI delivery using 2-isocenter plans and treated on Varian C-arm linear accelerator, treatment times for the first fraction from start of pre-treatment imaging to the end of treatment session had a wide range of 16.3 – 43.5 min (mean, 29.2 min) signifying existing challenges in imaging and aligning each isocenter separately. RefleXion X1 has a capability of overcoming this challenge with imaging a long extent of the patient (up to 90 cm), localizing, and delivering the whole treatment using 1 isocenter in axial mode moving the couch in the cranio-caudal direction with 2.1 mm increments. This may potentially reduce beam matching and shifting errors which may arise from multi-isocenter delivery. In addition, X1 was recently upgraded to enable 1000 MU/min dose rate from the initial dose rate of 850 MU/min. This upgrade promises to improve the beam-on time but needs to be validated.
While there are currently no studies comparing RefleXion CSI and VMAT CSI, literature discussing TomotherapyTM delivering the treatment using 2.5 cm jaws in helical fashion may be useful as a comparison due to its similar delivery approach to X1. A study in 2019 by Sun et al. [26] comparing VMAT, IMRT, and Tomotherapy plans found that the Tomotherapy plans offered superior PTV homogeneity, conformity, and brainstem, optic chiasm, and optic nerve sparing compared to VMAT plans. IMRT was superior to VMAT and Tomotherapy in terms of OAR sparing in the mid body region (esophagus and heart). Results of this study by Sun et al. differed from the results of the current RFX study, which found difference in Dmean to the bowel bag as the only statistically significant dosimetric parameter. However, just as the average beam-on time for RFX plans were estimated to be longer than average beam-on time for VMAT plans in our study, delivery time of HT was found to be longer than the delivery time of VMAT for Sun et al.
Another study by Herdian et al. [27] found that differences in oral cavity Dmean, kidneys Dmean, and mean D2% to the spinal PTV were statistically significant between IMRT and Tomotherapy plans. Differences in oral cavity Dmean, kidneys Dmean, mean D2% to the cranial PTV, and mean D2% to the spinal PTV were also statistically significant between 3D-CRT plans and HT plans. Additionally, Tomotherapy plans resulted in longer mean beam-on time than both IMRT and 3D-CRT [27].
One limitation of this study is the small sample size (n=5) due to the maximum target length threshold of 50 cm. The vendor is planning to upgrade the system to enable treating targets greater than 50 cm in their next clinical release. This will enable us to expand the patient selection. Another limitation is that this study only focuses on comparison between VMAT and RefleXion X1 plans. It would be interesting to include Tomotherapy plans in the testing cohort. Based on our communication with the vendor, RefleXion Medical Inc will lift the maximum target length limit in the future software releases of X1 treatment planning system. This will enable us to include larger target sizes in the investigation and collect and analyze the treatment delivery times. This work shows the feasibility of CSI planning using RefleXion X1 paving the way to potentially use RefleXion X1 for CSI treatments. There is a potential benefit of single-isocenter treatment with near-diagnostic image-quality kVCT guidance enabling long scanning length of 90 cm. This can simplify the IGRT workflow and streamline the treatment delivery for CSI patients, especially important for pediatric patients treated under anesthesia.