2309 - Neoadjuvant Immunotherapy Followed by Response-Adapted Definitive Radiotherapy for Unresectable, Locally Advanced Cutaneous Squamous Cell Carcinoma of the Head and Neck: Dosimetric Analysis of a Phase II Clinical Trial

Purpose/Objective(s): After neoadjuvant chemotherapy, guidelines suggest that definitive radiotherapy (RT) targets should be based on the extent of disease present before drug therapy. In an ongoing prospective phase II clinical trial of neoadjuvant immunotherapy (IO) with 2-4 doses of cemiplimab before definitive RT for unresectable, locally advanced (T3/T4 or N+) cutaneous squamous cell carcinoma (ULACSCC), RT was planned for delivery to the extent of disease present after neoadjuvant IO based on a different mechanism of drug action. We hypothesized that IO response-adapted RT would decrease planning target volumes (PTVs) and in turn reduce radiation doses to organs at risk (OAR) for injury, owing to a high response rate after neoadjuvant IO.

Materials/Methods: Patients with ULACSCC of the head and neck participating in a prospective phase II clinical trial of neoadjuvant IO prior to definitive intensity modulated RT (IMRT) at our institution were analyzed. Imaging performed prior to treatment was used to generate an IMRT plan according to protocol guidelines, with 3 possible dose levels (low, intermediate and high risk, planned to receive an equivalent total dose at 2 Gy/fraction of 50, 60 and 70 Gy, respectively). Patients with progressive disease after 6 weeks of IO underwent early IMRT planning and delivery, while those with response or stable disease underwent IMRT planning and delivery after 12 weeks of IO. Group average doses to PTVs and OARs from pre-IO and post-IO IMRT plans were compared using 2-tailed paired t-tests.

Results: Eleven patients were identified and met criteria for analysis. Composite response (complete and partial) rate to IO according to RECIST v1.1 at 6 and 12 weeks was 36% (n=11) and 100% (n=10); 1 patient experienced progression at week 6 and received early IMRT and was not re-evaluated for response at week 12. IMRT plans using pre-IO vs post-IO imaging demonstrated significantly smaller PTVs post-IO (compared to pre-IO) for the low, intermediate and high dose PTVs by 6%, 19% and 28%, respectively (p=0.04, 0.03, 0.06). The pre-IO and post-IO IMRT plans demonstrated no significant difference in dose to PTV at any dose level as measured by maximum point dose, dose to 95% and 99% of the PTV, or volume receiving prescription dose, suggesting plan quality in target coverage was comparable. However, post-IO IMRT plans demonstrated statistically significant reductions (compared to pre-IO IMRT plans) in maximum dose to brainstem (9%, p=0.02), spinal cord (14%, p=0.004), contralateral lens (36%, p=0.04), cochlea (36%, p=0.04), brachial plexus (15%, p=0.03), and mean dose to lips (8%, p=0.04), larynx (11%, p=0.04), pharynx (8%, p=0.008), and contralateral parotid (29%, p=0.01).

Conclusion: These results imply that IO response-adapted RT for ULACSCC of the head and neck may reduce doses to OARs compared to definitive RT alone. Further investigation will be required to determine if this approach is also more effective than alternative treatment strategies.