Background Elevated baseline neutrophil lymphocyte ratios (NLR) are now well established as a poor predictor of survival in renal cell carcinoma (RCC) and other cancers. Platelet Lymphocyte Ratios (PLR) have also recently shown similar effects. Despite these findings, the practical use of these ratios is still somewhat limited. We have previously shown that higher NLRs may be associated with increased concentrations of myeloid derived suppressor cells (MDSC). We hypothesized that increases in NLR or PLR (NLR/PLR failure) at 2 months while on immunotherapy could be a predictor of eventual treatment failure and overall survival.

Methods We analyzed patients who received nivolumab therapy for RCC at our institution from 3/2016 to 6/2019. Patients with complete data on NLR and PLR at time = 0 and +2 months and those who had accurate response and overall survival information available were selected (n = 37). Information on comorbidities, previous therapy, demographics were collected for adjusted analysis. NLR failure was defined as an increase of 3 or more compared to baseline NLR. Cox proportional hazard models were used to analyze the risk of progression and death with NLR/PLR failure at 2 months (± 2 weeks). Kaplan Meier graphs were constructed to trace survival functions for PFS and OS by NLR

Results NLR failure was associated with a statistically significant increase in the risk of progression on nivolumab therapy (HR 4.26, 95% CI [1.47–12.3], p = 0.007), in an adjusted cox regression model that included baseline NLR. In this adjusted model, the value of baseline NLR to predict treatment failure was non-significant (HR 1.01, p – 0.69). Similarly, NLR failure increased the risk of death (HR 3.83 95% CI [1.23–11.9], p = 0.02), with a similar non-significant contribution of baseline NLR (HR 1.06, p = 0.06). NLR failure predicted PFS less than 6 months with 90% positive predictive value (9/10) and a 48% (12/25) negative predictive value, and survival less than 1 year with a 56% negative predictive value and 100% positive predictive value (10/10). PLR changes failed to show any association with PFS (HR 0.99, p = 0.93) or OS (HR 1.00, p = 0.93).

Conclusions An increase in NLR of 3 or more at 2 months of therapy with nivolumab appears to predict for impending treatment failure and increasing risk of death with a high positive predictive value. NLR failure if validated in larger studies could be useful in treatment management

Ethics Approval The study was approved by UAB Comprehensive Cancer Centers Ethics Board