Associations of dietary fructose with survival of patients (pts) with metastatic cancer of the urothelium (UC) and renal cell carcinoma (RCC) on immune checkpoint blockade (ICB).

Abstract

4571 Background: The gut microbiome is a modifiable determinant of ICB efficacy. Dietary fiber is associated with longer progression-free survival (PFS) on ICB in melanoma and UC, putatively via effects on the gut microbiome. Other dietary factors may also impact ICB efficacy. Fructose promotes adverse health effects and induces ICB resistance in pre-clinical models via upregulation of the antiapoptotic enzyme heme oxygenase-1. Thus, we set out to explore effects of fructose intake on ICB efficacy. Methods: In a retrospective analysis of a prospectively collected cohort, we leveraged dietary data collected with the validated Harvard Willett Food Frequency Questionnaire from pts with advanced UC and RCC initiating ICB at Memorial Sloan Kettering to assess for associations between baseline dietary fructose and PFS and overall survival (OS). Tumor mutational burden (TMB) was estimated via MSK-IMPACT. Associations of fructose with clinical outcomes were evaluated using univariate (UV) and multivariable (MV) Cox proportional hazards regression. Fructose was normalized via log transformation and treated as a continuous variable. MV models adjusted for fiber intake and prior ICB exposure. UC models adjusted for TMB and Bellmunt risk score and RCC models for histology and IMDC score. Results: From 2/2021-6/2022, 88 eligible pts enrolled. Median follow-up was 10.2 and 17.0 months among pts with UC (n = 40) and RCC (n = 48), respectively; 32 RCC pts (67%) were IMDC intermediate/poor risk and 21 UC pts (52%) had visceral metastases. High fructose and fiber intake correlated (rho 0.66). Among UC pts, higher fructose was associated with shorter PFS after adjusting for covariables (Table). In the MV model, high fiber was associated with longer PFS (HR 0.2, 95% CI 0.07-0.62, p = .004). Among RCC pts, fructose was not associated with PFS. Fructose was associated with OS in the RCC UV model, but this association was not statistically significant in the MV model (Table). In an exploratory model for PFS pooling all pts, a test for interaction between fructose intake and disease type (UC vs RCC) showed p = .01. Conclusions: Dietary fructose is associated with shorter PFS in UC pts treated with ICB, but not RCC pts. Discrepant results between UC and RCC require further study and may be attributable to differences in therapy or cancer biology. Our findings provide insights into potentially beneficial dietary interventions. Efforts to characterize effects of fructose on the gut microbiome are ongoing. [Table: see text]