Dose Perturbation From Titanium Plates in Post-Operative Oral Cavity Volumetric Modulated Arc Therapy: The Utility of Model-Based Algorithm

Abstract

Computerized planning and 3D printing of titanium (Ti) plates attached during radical tumor resection allow for patient-specific, personalized solutions for complex osseous reconstructions in locally advanced oral cavity carcinoma. Here, we investigate dose perturbations related to such Ti plates and further interrogate whether the commonly used anisotropic analytical algorithm (AAA) dose calculation algorithm accurately predicts perturbations around Ti plates in comparison to the Boltzmann Solver Acuros XB model (AXB).Twenty-three cases of radically resected oral cavity carcinoma, with or without Ti plate reconstruction, followed by VMAT were reviewed with rates of spontaneous osteoradionecrosis (ORN) documented. Clinical treatment plans using AAA (2.5mm dose grid) were retrospectively re-planned with 1mm dose grids on AAA and AXB. Seven consecutive Ti plate reconstructed cases were compared to 16 control cases (postoperative cases without Ti plate) to understand the extent of dose perturbation near reconstructed bone and surrounding tissues using previously established dose metrics, such as maximum plan dose (Dmax) and maxillary/mandibular doses (Dmax, Dmean, V35, V44). Additionally, a 3mm shell was contoured around each plate to directly assess short-range dose perturbance at tissue-to-Ti interfaces.Of the 23 patients who underwent oral cavity resection followed by adjuvant VMAT, 4 underwent reconstruction with conventional low-profile Ti plates while 3 cases underwent reconstruction with high-profile custom plates. ORN was observed in 3 patients, all of whom had high profile custom plates. Global Dmax was significantly elevated in AXB 1mm (118.0 ± 3.3%) compared to AAA 2.5mm (109.1 ± 0.8%) and AAA 1mm (113.1 ± 2.0%), with most notable increases in the Ti plate cohort (P < 0.001; Table). Within the AXB 1mm plans, all a priori parameters were significantly higher in the Ti plate vs. no plate cohort (global Dmax: +5.7%, osseous Dmax: +4.9%, osseous Dmean: +10.9%, osseous V35: +13.4%, osseous V44: +15.9%). Furthermore, in the Ti plate population, AXB calculated significantly higher local hot spots within the 3mm shell contoured (115.6 ± 4.4% vs. 107.0 ± 1.7%; P < 0.002).Using the AAA model with a clinical standard 2.5mm dose grid does not appear to adequately account for dose perturbation at the tissue-to-Ti interface in comparison to the AXB algorithm at a finer dose grid, resulting in a significant underestimation of global and osseous maximum doses. Care should be taken when delivering radiation to patients with high profile Ti plates. The impact of beam entrance restriction and alternative plate designs to minimize the profile of high-Z materials are currently being explored.