Dosimetric Analysis of Lingual Artery Bleeding Events after Stereotactic Ablative Radiotherapy Reirradiation of Head and Neck Cancers

Abstract

<h3>Purpose/Objective(s)</h3> Stereotactic ablative radiotherapy (SABR) for treatment of recurrent head and neck cancer within a previously irradiated field can provide durable local control, but may be associated with significant treatment-related adverse events. Lingual artery bleeding (LAB) is an underreported and potentially fatal consequence of reirradiation with SABR. The dose tolerance of the lingual artery is unknown. The objective of this work is to identify dosimetric predictors of LAB using normal tissue complication probability modeling. <h3>Materials/Methods</h3> We identified head and neck cancer patients treated with definitive-intent SABR reirradiation (≥36 Gy in 4 fractions) to the oropharynx or oral cavity at our institution from 2013-2021. Institutional guidelines for the lingual artery include Dmax of <20 Gy when >1 cm from target, Dmax of <30 Gy and V27Gy <0.5 cc when <1 cm from target, or minimizing hotspots when target overlaps the lingual artery. Medical records were reviewed for LAB events and the laterality of bleed identified. The bilateral lingual arteries and their branches, including the main lingual, sublingual, and deep lingual arteries were contoured. Logistic dose-response modeling was used to predict the risk for LAB. <h3>Results</h3> We analyzed 43 patients, of whom 6 had a LAB event. LAB was grade 3 in 4 cases and grade 4 in 2 cases. Four of 6 patients underwent an embolization procedure. Branch sites of LAB included the main lingual artery (n=5) and deep lingual artery (n=1). The median time to LAB was 8.1 months (range, 4.1-13.2). Prescription doses included 45 Gy / 5 fx (n=23, 53%), 42.5 Gy / 5 fx (n=11, 26%), 40 Gy / 5 fx (n=6, 14%), and 36 Gy / 4 fx (n=3, 7%). The median age was 65 years, 10 (23%) were female, the median initial dose was 70 Gy (IQR, 64.5 – 70), the median reirradiation interval was 2.3 years, and 34 (79%) received concurrent chemotherapy. Among patients who experienced LAB, the mean D0.01cc, D0.1cc, and D0.5cc were 46.6 Gy, 45.7 Gy, and 43.4 Gy, respectively, compared to 39.7 Gy, 37.9 Gy, and 33.2 Gy among those who did not, respectively (p=0.01, p=0.005, p=0.002, respectively). In logistic modeling, Dmax, D0.01cc, D0.1cc, D0.5cc, and D1cc were all significantly associated with risk for LAB (<b>Table 1</b>). A risk of LAB <3%, 5%, and 10% was associated with a D0.5cc of <30Gy, 35Gy, and 40Gy, respectively. <h3>Conclusion</h3> Lingual artery bleeding is an underreported and potentially fatal complication of SABR reirradiation for oropharyngeal and oral cavity cancers. To maintain a risk of LAB <3%, 5%, or 10%, recommended dose constraints include D0.5cc of <30 Gy, 35 Gy, and 40 Gy, respectively. This represents the first available dose-volume tolerance data for the lingual artery in SABR reirradiation and should be incorporated into treatment planning.