A Retrospective Study of Predictors for Radiation Induced Severe Oral Mucositis in BPA-Mediated BNCT of Head and Neck Cancer Patients

Abstract

Radiation dose in boron neutron capture therapy (BNCT) consists of high-LET dose from 10B(n,a)7Li reaction, nitrogen capture dose, recoil proton dose and photon dose. Traditionally, the biological radiation dose in BNCT is obtained as a sum of the dose components after multiplying each with a constant relative biological effectiveness (RBE) factor. An alternative biological dose calculation method, the photon iso-effective dose formalism, considers also the dose rate, cumulative dose per fraction and synergistic interactions. Results of our previous study suggested that the photon iso-effective dose model predicts mucosal membrane toxicity after BNCT more reliably than the traditional RBE model or the physical absorbed doses. We aim here to verify this finding in a larger patient series. In addition, we study the relation of chemotherapy, tumor location, tumor histology, patient age, and gender to the perceived mucosal toxicity. Data of 93 patients with recurrent inoperable HN carcinoma, who were treated with BNCT in Finland from 2003 to 2011, was analyzed. Photon iso-effective dose formalism and traditional RBE factors were applied to calculate the biological dose delivered to mucosal membrane. Since exact nitrogen concentration of mucosal membrane is unknown, the dose was calculated using three different tissue concentrations (muscle 3.4 %, skin 4.2%, and melanoma 6.8%). The maximum mucosal membrane doses were compared with the clinically observed mucositis at 2 to 4 weeks after the 1st BNCT. Severe mucositis (CTCAE 4.0 grade 3) of oral mucositis was observed in 40 (43%) out of the 93 patients. The maximum RBE doses were 16-43% lower than the photon iso-effective doses (table 1), being below the reported tolerance dose limit of 15 Gy for a single radiotherapy fraction. All median photon iso-effective doses were higher than the tolerance dose. The other examined clinical factors did not predict the perceived mucositis in binary logistic regression analysis (p>0.05). The preliminary results confirm that the photon iso-effective dose model predicts mucosal membrane dose after BNCT more reliably than the traditional RBE model. However, maximum dose, tumor location, previously given chemotherapy, patient age and gender, and tumor histology do not seem to correlate with the mucosal toxicity. Dose volume distribution in the oral cavity, blood boron concentration, and irradiation time may be needed to predict the mucosal toxicity better.Abstract 2934; Table 1Maximum biological RBE and photon iso-effective dose to mucosal membrane from single BNCT calculated using nitrogen concentration of muscle, skin, and melanoma.Muscle Median (min, max)Skin Median (min, max)Melanoma Median (min, max)Gy-RBEGy-IsoEGy-RBEGy-IsoEGy-RBEGy-IsoE12.1 (5.6, 15.2)16.2 (9.8, 18.9)12.4 (5.8, 15.6)16.3 (10.1, 19.0)13.3 (6.1, 16.7)17.0 (10.3, 19.9) Open table in a new tab