The ADscopal Effect? Control of Partially Irradiated Versus Completely Irradiated Tumors on a Prospective Trial of Pembrolizumab and SBRT Per NRG-BR001

Abstract

Preclinical data suggests SBRT may augment the effects of PD1 blockade. We conducted a phase I study for patients with metastatic solid tumors to evaluate the combination of pembrolizumab with multi-organ site ablative radiation therapy (MOSART) per the NRG-BR001 trial (NCT02608385). The maximum gross tumor volume (GTV) to be irradiated on protocol was 65cc and partial tumor irradiation was permitted with the hypothesis that SBRT may boost the immunotherapy locally and systemically. Herein we report the local tumor control for partially and completely irradiated tumors. Between January 2016 and December 2016, 48 patients with at least one imaging follow-up were enrolled at a single institution. SBRT dose varied per NRG-BR001 (15 Gy x 3 for peripheral lung, liver, and abdominal pelvic; 10 Gy x 5 for central lung and mediastinal/cervical, 10 Gy x 3 for osseous and spinal). Two lesions were targeted for each patient and not all sites of disease were treated with SBRT. GTV was delineated for both lesions. If the GTV was >65 cc, a SUCITV was created within the GTV as a target structure with volume contraction down to 65 cc. A PTV expansion was used for all lesions except spinal. Pembrolizumab 200 mg IV Q3W was initiated within 7 days after the final SBRT fraction. RECIST 1.1 was used to assess radiated tumor progression. Control (defined as CR, PR, or SD) was calculated by the Kaplan Meier method and compared with the log-rank test. A total of 96 lesions were radiated including 25 peripheral lung, 18 central lung, 8 mediastinal/cervical, 15 liver, 21 abdominal-pelvic, 5 osseous, and 4 spinal lesions. Median follow-up was 14 wks (range 4–52 wks). Fourteen lesions were partially irradiated using a SUCITV (partial group) and 82 lesions were completely irradiated using the original GTV (complete group). Mean GTV size was significantly larger in the partial group (177 cc vs 11 cc, P < 0.001). In the partial group, the mean volume of GTV excluded from the target SUCITV was 113 cc (range = 23 cc-326 cc). Coverage of the original GTV by the prescription dose was significantly lower in the partial group (median 13% isodose line, IQR = 5%-51%) vs complete group (median 100% isodose line, IQR = 86%-104%) (P < 0.001). Overall, 5 lesions had a complete response, 19 partial response, 65 stable disease, and 7 progressive disease with no difference between groups (P = 0.42). At 6 months, the treated lesion control was 93% in the partial group and 96% in the complete group (P = 0.32). Large partially irradiated tumors exhibited control similar to smaller tumors that were fully encompassed in the target volume. To our knowledge, this is the first human data supporting preclinical studies showing enhanced local effects of SBRT and PD-1 blockade. These preliminary data suggest the effects of combining SBRT with pembrolizumab may be abscopal, or “close to the target.”