The objective of this study was to compare radiation dose distributions created by two radiosurgery modalities for treating a large number (≥7) brain metastases: single-isocenter volumetric modulated arc therapy (VMAT) and GammaKnife Perfexion (GK). We examined 103 brain metastases (range 0.01- 19.95 cc) from twelve patients. The mean number of targets per patient was 9 (range 7-14). Highly-optimized GK plans were compared to highly-optimized single-isocenter, multiple non-coplanar arc VMAT plans using both 6FFF (1400 MU/min) and 10FFF (2400 MU/min) high intensity modes from a Truebeam linear accelerator equipped with a high-definition multileaf collimator. Three layers of ‘tuning rinds’ were used in VMAT planning to control dose-volume constraints corresponding to high (Rx of each target), medium (12 Gy), and low-dose levels (6 Gy) [1-4]. Dosimetric parameters analyzed included RTOG conformity index (CI), 12 Gy (a known predictor of radiation necrosis [5]), 6 Gy, 3 Gy isodose volumes (V12Gy, V6Gy, V3Gy), mean and maximum hippocampus dose, maximum skin dose, and beam-on time. Each lesion had > 99.5% volume covered by the Rx dose. Direct comparison was performed using a Kruskal-Wallis test with Dunn’s post hoc comparisons (significance criteria: p < 0.05). There was a significant CI difference between GK and both 6FFF and 10FFF (2.52 ± 1.64 vs 1.59 ± 0.81 and 1.68 ± 0.87, p < 0.001). For V12Gy, there was a significant difference between GK and both 6FFF and 10FFF (2.79 ± 6.13 cc vs 2.97 ± 5.22 cc and 3.13 ± 5.39 cc, p <0.001). For V6Gy, there was no significant difference between GK and 6FFF, however, GK was significantly smaller compared to 10FFF (81.12 ± 72.86 cc vs 167.55 ± 87.55 cc, p=0.011). For V3Gy, GK was significantly smaller compared to both 6FFF and 10FFF (323 ± 295 cc vs, 880 ± 369 cc and 938 ± vs 362 cc, p < 0.001). For mean hippocampus dose, there was no significant difference between GK and 6 FFF, however, GK was significantly smaller compared to 10FFF (1.88 ± 1.35 Gy vs 3.59 ± 1.36 Gy, p = 0.013). No significant differences existed for maximum hippocampus or skin doses. There was a significant beam-on time difference between GK and both 6FFF and 10 FFF (148 ± 49 min vs 11 ± 2 min and 6 ± 1 min, p < 0.001). This study focused on cases with at least 7 brain metastases. Highly-optimized VMAT produced improved conformity at the expense of a higher V12Gy and V3Gy volume when compared to highly-optimized GK. However, the increased V12Gy with VMAT may not be clinically significant while increased V3Gy is comparable to a single fraction of whole-brain radiotherapy. [1] Liu et al. Front. Oncol. 6:26 (2016) [2] Thomas et al. Neurosurgery. 75(4): 409-418 (2014) [3] McDonald et al. J Neurosurg (Suppl 2) 121: 51-59 (2014) [4] Clark et al. Pract. Radiat. Oncol. 2: 306-313 (2012) [5] Minniti et al. Radiat. Oncol. J. 6: 48 (2011).
Read full abstract