Patterns of Loco-Regional Failure and Outcomes after Intensity Modulated Radiation Therapy for Unresectable Anaplastic Thyroid Cancer

We evaluated the patterns of loco-regional failure (LRF) and patient outcomes after intensity modulated radiation therapy (IMRT) in patients with unresectable anaplastic thyroid cancer (ATC) to inform target volume delineation and treatment strategies. We reviewed patients with unresectable ATC treated from 2010-2019 who received therapeutic dose IMRT (≥ 45Gy) and had at least 1 post-treatment neck CT. LRF after IMRT was defined as increase in size of treated disease or development of new disease in the neck. Recurrent gross tumor volumes (rGTV) were contoured on diagnostic CTs and co-registered with initial planning CTs using deformable image registration (Velocity AI). LRFs were classified based on established spatial and dosimetric criteria: Type A- central high dose; B- peripheral high dose; C- central elective dose; D- peripheral elective dose; and E- extraneous dose. Survival was estimated using the Kaplan-Meier method and groups were compared using Wilcoxon or log-rank analysis. 30 patients formed the cohort; 17 were stage IVA/B and 13 IVC. Median follow up after IMRT was 6.2m (1d-63m). Median prescribed dose was 66Gy (45-68Gy). 24 patients (80%) were treated with multi-dose levels and 6 patients (20%) were treated with a single dose level. 27 patients (90%) received concurrent chemotherapy. 17 patients (57%) developed LRF. 35 rGTVs were identified. Median LRF free survival (LRFFS) was 8.4m. Fractionation (daily, twice daily, or 6x/week) did not predict LRFFS (p = 0.7). No difference in LRFFS was seen in patients treated to 1 dose level vs. multi-dose levels (p = 0.7). A majority of failures occurred in the high dose volume- Type A (n = 25, 71%) and Type B (n = 2, 5.7%). 3 recurrences in 3 patients (2 Type C, 1 Type D) occurred in elective dose volumes, while the remaining 5 recurrences in 4 patients were in extraneous dose regions (Type E). Mean high dose target volume in those with vs. those without LRF was 217 cc vs. 192 cc, respectively (p = 0.6). Median OS was 6.7m from end of IMRT (11.8m for stage IVA/B, 3.1m for IVC, p = 0.02). For those with stage IVA/B, median OS was 11m for those with LRF and 15m for those without (p = 0.7). Overall, 24 patients (80%) developed distant failure (13 with new sites, 11 with progressive DM). Only 4 patients (13%) required tracheostomy or tracheal stent during or after IMRT. 15 (50%) required a feeding tube. Despite an aggressive chemoRT schedule, central high dose failures predominated. These findings are consistent with a classic radio-resistant disease profile. Given the high competing risk for distant failure, no survival difference was detected between those with vs. those without local regional control. However, the low rate of airway interventions required suggests a palliative benefit. Strategies to maximize both local and distant control including neoadjuvant targeted therapy in those with actionable mutations, followed by surgical resection and post-operative chemoRT are under current investigation.