Safety and Efficacy of a Randomized Phase I Trial to Evaluate Concurrent or Sequential Ipilimumab, Nivolumab, and Stereotactic Body Radiotherapy in Patients with Stage IV Non-small Cell Lung Cancer (COSINR Study)

Stereotactic body radiotherapy (SBRT) provides high rates of treated metastasis control, stimulates innate and adaptive immune pathways, and is safe in patients treated with anti-PD1 monotherapy. We hypothesized that SBRT may improve outcomes for patients receiving immunotherapy through direct cytoreduction and increased immunogenicity. We designed a phase 1 trial to evaluate the safety of combination immune checkpoint blockade with nivolumab and ipilimumab (N/Ip) plus sequential (Seq) or concurrent (Con) multisite SBRT as first-line therapy for patients with stage IV NSCLC. Chemotherapy-naïve patients with metastatic NSCLC received SBRT to 1 to 4 metastases. Not all metastases were targeted, and metastases > 65mL were partially irradiated. Brain metastases were allowed, and those > 3mm were treated prior to enrollment. SBRT dose varied by anatomic site and ranged from 45 to 50 Gy in 3 to 5 fractions with predefined dose de-escalation if excess dose-limiting toxicities (DLTs) were observed. Patients in the Seq arm received N/Ip 1 to 7 days after completion of SBRT. Patients in the Con arm received N/Ip prior to completion of SBRT. N/Ip continued until progression, toxicity, or up to 2 years. To assess cytoreduction 3 months (mo) after treatment, deep learning algorithms were used to contour all metastases in patients, which were verified by radiologists. A total of 35 patients (n = 19 Seq, n = 16 Con) were enrolled and evaluable for toxicity analysis (SBRT and at least 1 cycle N/Ip). Brain metastases were present in 27% of patients. PD-L1 expression was: 0% (n = 16), 1-49% (n = 10), or >50% (n = 9). The median number of metastases treated with SBRT was 3.2. Six patients experienced DLT (4 pneumonitis), resulting in a 20% reduction in organ-at-risk constraint in the Seq arm. Median PFS by RECIST (total/Seq/Con) was 5.9mo (95% CI: 4.9-13.1)/ 6.2mo (95% CI: 3.5-12.6)/ 5.9mo (95% CI: 3.1-18.0). RECIST best response was 11% CR, 57% PR, 6% SD, and 26% PD. When comparing mean volumetric cytoreduction of non-irradiated metastases at 3mo, the Con arm was superior by 3-fold (p<0.05). Treatment past progression was allowed, and time to second-line therapy (chemotherapy) was NR/17.5mo (Seq/Con). Median OS has not been reached with median follow-up of 15mo. PD-L1 status did not impact PFS (p = 0.64) nor OS (p = 0.77). Multisite SBRT and N/Ip were well tolerated, with median OS and time to chemotherapy not reached (>17mo). Fewer DLT events were observed in the Con arm. Deep learning algorithms demonstrated significant differences in cytoreduction (abscopal response) of non-irradiated metastases favoring the Con arm. Based on the lower toxicity and improved cytoreduction, the Con arm will be used in the phase 2 expansion. Clinical trial information: NCT03223155.