Development and validation of a treatment decision model for optimal delivery of liver stereotactic body radiation therapy (SBRT): Photons versus protons.

Abstract

264 Background: Treatment of liver tumors with SBRT provides excellent local control rates, but is often limited by liver dose constraints. Protons offer potential for more liver sparing, but clinical situations where protons may be superior to photons are not well described. We developed and validated a treatment decision model to determine whether patients with liver tumors of certain sizes and locations are more suited for photon vs proton SBRT. Methods: Six spherical mock tumors from 1-6 cm in diameter were contoured on axial CT images of one patient at each of 4 locations (24 total tumors): peripheral/dome, caudal, left medial, central. We generated photon and proton plans (48 plans) to deliver 50 Gy* in 5 fractions with ~135% dose max within the gross tumor volume. Mean liver dose was constrained to < 14 Gy and at least 700 cc received < 15 Gy. Using these plans, we developed a hypothesis-generating model to predict the optimal modality for maximal liver sparing based on tumor size and location. We validated this model in 10 actual liver tumor patients by generating and comparing photon and proton plans. Results: Proton plans spared significantly more liver volume and decreased mean liver dose, particularly for tumors ≥ 3 cm in the dome or central locations− peripheral/dome: 134 ± 21 cc (p = 0.03), 912 cGy vs 1303 cGy (p = 0.01); central: 108 ± 4 cc (p < 0.01), 1179 cGy vs 1454 cGy (p < 0.01). Our model correctly predicted the optimal SBRT modality for all 10 patients. There was a significant increase in the volume of liver spared (176 ± 21 cc, p < 0.01) and decrease in mean liver dose (841 cGy vs 1223 cGy, p < 0.01) with protons for patients with dome or central tumors ≥ 3 cm. As predicted, in patients with tumors < 3 cm at any location or caudal and left medial tumors of any size, there was no significant difference in liver sparing between photon and proton SBRT (33 ± 18 cc, p = 0.13). Conclusions: Tumor size and location are important determining factors in generating deliverable liver SBRT plans. When feasible, protons should be considered as the radiation modality of choice for dome or central tumors ≥ 3 cm to allow maximal liver sparing and potentially reduce radiation toxicity. *All doses are in Gy (RBE).