MR-Guided Liver Stereotactic Body Radiotherapy (SBRT): To Adapt, or Not to Adapt?

Abstract

Long-term disease-free survival is possible in liver oligometastatic cancer with ablative local treatment such as SBRT, however in some studies, long term local control remains an issue. Planning target volume (PTV) for tumors accounts for positional uncertainty, but not intra-fractional changes in tumor size, which are assumed to be negligible. Magnetic resonance imaging (MRI)-guided radiotherapy (MRgRT) to the liver using gadoxetate contrast allows for radiation oncologists to evaluate daily fractional dose to liver metastases and to adapt for anatomical variation. The purpose of this study is to determine the extent of gross tumor volume (GTV) changes between simulation and treatment start, as well as during each fraction of SBRT. Patients who underwent 5-fraction maximum inspiratory breath hold MR-guided liver SBRT for metastases were retrospectively screened for use of gadoxetate during simulation and at least the first treatment. An experienced radiation oncologist contoured the GTV on the MR planning scan and subsequent daily setup MRIs. Sørensen–Dice coefficients (DC) were calculated. Wilcoxon signed rank test was used to test significance of the median change in tumor volume using SPSS version 26.0. A total of 37 patients were treated with MR-guided SBRT for metastatic liver lesions, 20 of which had gadoxetate enhanced MRs for simulation and at least the first fraction of treatment. SBRT doses ranged from 40-80 Gy over 5 fractions. 17 out of 20 patients had all treatments using MRgRT (91 evaluable fractions). Median simulation GTV was 5.4 ml (R: 0.29-90.47 ml). From simulation to fraction 1, the mean DC was 0.78 (R: 0.55-.877). Mean DC for tumors treated solely using MRgRT for fractions 1 to 2, 2 to 3, 3 to 4, and 4 to 5 were 0.83, 0.82, 0.82, 0.79, 0.82, respectively. Following simulation, GTV volume was found to increase peaking at fraction 2 with a median 37% increase over simulation (95% CI: 20% to 59% increase, p<0.001) with a median absolute increase of 1.9 ml. Thereafter, volume decreased to a median of 17% (95% CI: -1%-25%) greater than simulation by the final fraction (p<0.01). At the time of the second fraction, 70% of tumors had increased by 20% over simulation volumes with 35% having a volume that was ≥ 50% larger than at the time of simulation. At the time of treatment completion, tumors had begun to regress and only 43% measured ≥ 20% larger than at the time of simulation. Volumes of GTVs visualized on initial gadoxetate enhanced MRI simulation demonstrated notable tumor volume changes throughout SBRT treatment. Potential causes for GTV changes include tumor growth, inflammation, treatment response, or radiation effects on gadoxetate distribution. These changes suggest there may be a role for daily adaptive radiation planning to improve plan coverage using MRgRT. In the absence of MR guidance, this study suggests the importance of minimizing the time to treatment and/or using a larger PTV margin to account for GTV changes.