Factors affecting mandibular complications in low dose rate brachytherapy for oral tongue carcinoma with special reference to spacer

Purpose: To evaluate the efficacy of a spacer in the prevention of mandibular complications in low dose rate (LDR) brachytherapy (BRT) for oral tongue carcinoma. Methods and Materials: A retrospective analysis was conducted using 103 patients with T1 or T2 tongue carcinoma treated by a single plane implantation of iridium (192Ir) pins between 1979-1994. Of these patients, 60 were treated by BRT alone, and the rest were combined with external irradiation (Ext) and/or chemotherapy (CHT). Forty-eight and 55 patients were given BRT with and without a spacer, respectively. Spacers were individually made of acrylic resin according to a prosthetic technique so as to obtain the thickness of 7-10 mm at the lingual part of the implanted side. Variables, including a spacer, which may be associated with the development of osteoradionecrosis (ORN) of the mandible, were analyzed by the Cox proportional hazards regression analysis. Results: Our spacer reduced about 50% of the absorbed dose at the lingual side surface of the lower gingiva (LSG) to that in the absence of a spacer. Absolute incidence of ORN was 2.1% (1 of 48) and 40.0% (22 of 55), with and without a spacer, respectively, and the difference was statistically significant by univariate analysis (p = 0.0004). It was revealed by the Cox analysis that the spacer (p = 0.0247), combined CHT (p = 0.0295), and combined Ext (p = 0.0279) were significant independent factors associated with the development of ORN. The spacer was shown to be a significant factor by univariate analysis (p = 0.0037), but not by multivariate analysis when analysis was restricted to the patients who did not receive CHT. The absorbed dose, dose rate, and biological effective dose (BED) reflecting early or late response were estimated at the LSG, and prognosticators associated with the incidence of ORN were also determined by the Cox analysis. Particularly, BED for late response by BRT, the total absorbed dose, and any BED by Ext plus BRT were highly significant factors in the whole population. Essentially similar results were obtained in the patients without receiving CHT. Conclusions: It was clarified that our spacer effectively prevents mandibular complications in LDR BRT by 192Ir for oral tongue carcinoma. Furthermore, introduction of a spacer provided novel information concerning the development of ORN, where BED particularly for late response given by BRT, the total absorbed dose, and any BED by Ext plus BRT could be good prognostic factors only when estimated at the LSG.