Head and Neck Stereotactic Body Radiation Therapy Re-Irradiation for Patients With Carotid or Skin Involvement Ineligible for RTOG 3507

Abstract

Patients with recurrent head and neck squamous cell carcinoma (HNSCC) who present with carotid encasement (CE) > 180˚, skin involvement and/or require bolus are not eligible for SBRT per RTOG 3507. The objective of this study was to characterize our institutional experience of SBRT for patients ineligible for RTOG 3507.Patients with recurrent HNSCC treated with SBRT from 2010-2020 with a history of prior radiation were extracted from an IRB approved database. Patients were evaluated based on CE > 180˚ or ≤180˚ and if there was skin involvement and/or use of bolus. The primary outcome was incidence of treatment related carotid bleeding event (CBE) and skin toxicity after SBRT in the absence of disease.Thirty-one patients were treated with SBRT to 37 sites with a median follow up of 8 months (mo) (range 2-47). The median time from prior radiation to SBRT was 13 mo (range 2-257). SBRT fractionation included 35 Gy (13%), 40 Gy (57%), and 45 Gy (30%) in 5 fractions. SBRT was delivered every other day (76%) or 2 fractions (24%) per week. Concurrent immunotherapy was given in 11% and cetuximab in 8%. A total of 30% (N = 11) of sites treated exhibited CE > 180˚ and 70% (N = 26) ≤180˚. There were no CBE related to SBRT, however 18% (N = 2) of patients with CE > 180˚ experienced fatal CBE related to progressive disease at 2.3 mo and 7.6 mo from SBRT. The risk of CBE was potentially related to CE > 180˚ (P = 0.08). Skin involvement and/or use of bolus was present in 43% (N = 16) of sites treated and absent in 57% (N = 21). The rate of treatment related skin toxicity was 19% (N = 3) and only occurred in those with pre-SBRT skin involvement and/or use of bolus. The cumulative incidence of treatment related skin toxicity at 3 mo and 6 mo was 13% (95% CI 1.9-33.6) and 27% (95% CI 4.0-58.8), respectively. Of the 3 treatment related skin toxicities there was one each of grade 2, 3, and 5. Grade 5 toxicity was related to untreated SBRT wound infection as the patient transitioned to hospice. The risk of treatment related skin toxicity was possibly related to pre-SBRT skin involvement and/or use of bolus (P = 0.007). Of the 13 sites with pre-SBRT skin involvement, 31% (N = 4) completely resolved, 54% (N = 7) had residual ulceration attributed to disease, and 15% (N = 2) had ulceration related to SBRT toxicity. The cumulative incidence of local failure at site of SBRT at 3 mo and 6 mo was 13% (95% CI 4-27.3) and 38% (95% CI 20.5-55.9), respectively. The median time of local and regional recurrence free survival were 7 and 5.5 mo, respectively. Median OS was 8 mo and the estimated 1-year OS was 25% (95% CI 9.9-40.9).The risk of toxicity following SBRT for patients with > 180˚ CE, skin involvement and/or use of bolus appears to be largely related to persistent or recurrent disease. There were no CBE related to SBRT in the absence of disease and 31% of patients with pre-SBRT skin involvement exhibited complete healing after SBRT. Consideration of SBRT may still be beneficial for such patients with appropriate counselling.