Purpose: An intramuscular tumor is not visible on CT but delineable on MR scans. We compared dosimetric differences between CT-guided radiotherapy (CTgRT) and MR-guided radiotherapy (MRgRT) to demonstrate the advantages of treating such tumor with MRgRT.
Methods: A patient with a renal cell carcinoma gluteal intramuscular oligometastasis was treated with a 1.5T MR-Linac using daily adaptive MRgRT. The prescription was 60 Gy to GTV and 45 Gy to PTV delivered in 15 fractions. Fractional delivered dose of MRgRT was accumulated to the simulation MR scan by aligning the GTV on daily treatment MR scans to the simulation MR scan using rigid registration. In parallel, we developed an in-house cycle-consistent generative adversarial network (cycleGAN) model to generate synthetic CT for each daily MR scan. The synthetic CT was used to simulate daily CBCT alignment using rigid registration based on nearby bony landmarks because tumor is invisible in CT. The fractional dose based on CTgRT was accumulated to the simulation MR scan for comparison with MRgRT.
Results: MRgRT D95% dose accumulation was 60.1 Gy and 47.5 Gy with average dose of 61.1 Gy and 55.0 Gy to GTV and PTV respectively. CTgRT D95% dose accumulation was 37.3 Gy and 24.4 Gy with average dose of 44.8 Gy and 38.8 Gy to GTV and PTV respectively. CTgRT dose accumulation resulted in >35% and >45% under-dose compared to the planned GTV and PTV D95% dose respectively. Qualitatively, the low-dose regions in CTgRT were comparable to MRgRT, however the high-dose regions were significantly different. The GTV was partially missed in 10/15 fractions and completely missed in 5/15 fractions to receive full planned dose in CTgRT simulated treatments.
Conclusion: Treating intramuscular tumors using CTgRT can significantly under-dose the tumor without excessively generous margin. MRgRT is ideal for cases where tumor is not visible on CT.