Natural killer cells as a predictive biomarker for response to anti-PD-1 therapy in patients with advanced solid tumors

e21055 Background: Assessment of PD-L1 expression in tumor biopsy has been widely used for predicting anti-PD-1 therapy. Several factors, including tumor heterogeneity, false negative staining and unavailability of biopsy samples limit the application. Natural killer (NK) cells exert cytotoxicity against cancer cells. Therefore, we explored using NK cells percentage in peripheral blood (PB) as a biomarker to identify pts who are more likely to benefit from anti-PD-1 therapy. Methods: Pts treated with anti-PD-1 antibody (JS001) 1 mg/kg, 3mg/kg or 10mg/kg every 2 weeks in a Phase I clinical trial (ClinicalTrials.gov Identifier: NCT02836795) had PB collected at baseline and at time of radiographic evaluation performed every 8 wks. Disease status (shrinking stable disease, growing stable disease, partial response, complete response, or progressive disease) was characterized using RECIST v1.1 and immune-related response criteria. Frequencies of NK cells (CD3-CD16+CD56+) were measured by flow cytometry. Results: Thirty-six patients were enrolled in this trial, including 22 melanoma, 9 urothelial carcinoma and 5 renal cell carcinoma. As of 1/27/2017, median treatment duration: 3.92 months. Radiographic tumor evaluation has been performed in 32 pts: 1/32 CR, 6/32 PR, 6/ 32 shrinking SD, 4/32 growing SD and 15/32 PD. Pts with high NK cells at baseline observed 60% (6/10) Clinical Benefit Rate (CR/PR/ shrinking SD) versus 40% (13/32) in overall pts. Patients with clinical benefit had higher frequency of NK cells at baseline compared to pts with radiographic growing SD and PD (mean 24.3% vs. 17.9%, P = 0.076). Furthermore, pts with NK cell level consistently above UN or increasing above UN during treatment are more likely to get clinical benefit than those under UN (7/10 vs 3/22, P = 0.013). Conclusions: Baseline high NK cells frequency may predict a favorable response to anti-PD-1 therapy, which provides a non-invasive way to monitor response to PD-1 blockade. These results are being validated in a phase II trial. Clinical trial information: NCT02836795.