Neoadjuvant nivolumab, ipilimumab, and celecoxib in MMR-proficient and MMR-deficient colon cancers: Final clinical analysis of the NICHE study.

3511 Background: The combination of PD-1 and CTLA4 blockade has changed the treatment landscape for several cancer types. Although this treatment is highly effective in metastatic mismatch-repair deficient (dMMR) colorectal cancers, metastatic MMR-proficient (pMMR) tumors do not respond. The NICHE study was the first neoadjuvant immunotherapy study in colon cancer (CC) to show impressive responses in 100% of dMMR (n= 20) and 27% of pMMR (n= 15) CC. In contrast, pathologic response to neoadjuvant chemotherapy using standard of care folfox is approximately 5% in dMMR tumors. Here we present the final efficacy data for the original NICHE study cohorts. Methods: Patients with non-metastatic, resectable dMMR or pMMR CC were treated with a single dose of ipilimumab 1mg/kg and two doses of nivolumab 3mg/kg and underwent surgery within 6 weeks. In addition, patients with pMMR tumors were randomized to receive celecoxib. The primary endpoints were safety and feasibility, and secondary endpoints included pathologic response rate and disease-free survival in 30 patients with dMMR and 30 with pMMR tumors. Pathologic response was defined as 50% or less viable tumor rest (VTR), and major pathologic response (MPR) as <10% VTR. Results: Thirty patients with pMMR and 32 with dMMR tumors were evaluable for the efficacy analyses. In the pMMR cohort, pathologic response was observed in 9/30 (30%, 95% CI 14-46%) patients, consisting of 7 MPR (including 3 pathologic complete responses pCR) and 2 partial responses. Four out of 9 pMMR responders had received celecoxib. Five patients received adjuvant chemotherapy. At a median follow-up of 25 months (IQR 12-35 months), 3 patients (all non-responders) in the pMMR group had disease recurrence. In the 32 patients with dMMR tumors, pathologic response was observed in 100% of patients, with 31/32 MPR (97%, 95% CI 91-100%) and one partial response. Pathologic complete response was observed in 22/32 (69%, 95% CI 53-85%) patients. None of the patients in the dMMR cohort had disease recurrence. Surgery was delayed in one patient with a pMMR tumor due to myositis. Grade 3 immune-related adverse events were observed in 12% of patients, consistent with our previous report on the primary endpoint. There were no grade 4 immune-related AEs nor unexpected surgical complications. Conclusions: These data confirm our previously published results of the NICHE study, with responses to neoadjuvant nivolumab plus ipilimumab in 30% of pMMR and 100% of dMMR CC in the completed original cohorts. Validation of the dMMR responses in a large group of dMMR patients is ongoing and has the potential to change current practice. Clinical trial information: NCT03026140.