251 - Immune Profiling of "Radvax" Abscopal Responses in Metastatic Melanoma Patients Progressing on Anti-PD1 Therapy Treated with Hypofractionated Radiotherapy (HFRT) to One Lesion Combined with Pembrolizu

J. N. Lukens¹, A. C. Huang², A. J. Minn III¹, and A. Maity³; ¹University of Pennsylvania, Philadelphia, PA, ²U. Pennsylvania, Philadelphia, PA, ³Department of Radiation Oncology, Huntsman Cancer Institute, University of Utah, Salt Lake City, UT

Purpose/Objective(s): This prospective “Radvax” trial combined HFRT to one metastatic lesion with pembrolizumab (pembro) in melanoma patients progressing through anti-PD1 therapy. Immune profiling was undertaken to characterize dynamic immune features occurring before and after RT, and correlate these changes to clinical response. Materials/Methods: In this prospective trial, melanoma patients were required to have disease progression (by RECIST v1.1) after = 2 doses of anti-PD1, an index lesion =1 cm amenable to HFRT, and =1 other lesion that was not radiated to follow for “abscopal” response. Pembro was given 1 week prior to RT, and then continued after RT. The HFRT dose was 8 Gy x 3 fractions or 17 Gy x 1 fraction. Response of unirradiated metastatic lesions (abscopal response) was measured by RECIST. High dimensional flow cytometry and single-cell RNA/TCR-sequencing were performed on peripheral CD8 T cells from Radvax patients and from a pembro-only control cohort.

Results: In metastatic melanoma patients progressing through anti-PD1, 7 of 16 had an objective response to HFRT + pembro (abscopal response), including 4 complete and 3 partial responses. Responses are durable, with 3/4 complete responders alive with no evidence of progression >5 years after HFRT, and 1/3 partial responders alive, now with slow progression >6 years after HFRT. Of the 5 patients who received HFRT to a nodal metastasis, 4 had CR/PR. In contrast to a cohort of pembro-only patients who exhibited a robust increase in proliferating PD1⁺ CD8 T cells that peaked one week after the first cycle of pembro, RadVax patients had a marginal increase in proliferating PD1⁺ CD8 T cells at this timepoint, consistent with relative refractoriness to PD1 blockade. However, an increase in proliferating CD8 T cells was observed in RadVax patients two weeks after HFRT. In clinical responders, the composition of proliferating CD8 T cells after HFRT was characterized by an increase in early effector-memory (EM_early) and central memory (CM) CD8 T cells. Moreover, a post-HFRT increase in developmental relatedness between EM_early/CM T cells with recently activated CD8 T cells was observed, suggesting the EM_early/CM population may have been newly primed after HFRT. In non-responders, the post-HFRT composition of proliferating CD8 T cells did not increase in either recently activated or EM_early/CM CD8 T cells.

Conclusion: This analysis provides insights into how addition of HFRT can elicit durable clinical response to anti-PD-1 in a subset of melanoma patients who fail to respond to anti-PD1 alone. After addition of HFRT to pembro, patients who subsequently respond show an increase in treatment-responsive effector and memory CD8 T cells that may result from new T cell priming. The expansion of these non-exhausted effector and memory CD8 T cells after HFRT may explain the durability of observed responses. In patients who received HFRT to lymph node metastases, response rates were particularly high, suggesting the site of irradiation may influence efficacy.