Abstract 247: Molecular markers of the vitamin D pathway and esophageal adenocarcinoma survival

Abstract

Abstract We investigated the potential prognostic association between vitamin D levels (25-Hydroxyvitamin D [25(OH)D]) and vitamin D pathway single nucleotide polymorphisms (SNPs) with overall survival (OS) in esophageal adenocarcinoma (EA) patients. EA patients were recruited from Massachusetts General Hospital near the time of their diagnosis between 1999 and 2016. At the time of recruitment, patients completed a baseline questionnaire and provided blood and serum samples. Serum 25(OH)D levels were measured using radioimmunoassay and adjusted for month of blood draw. Genotypes for 48 tag SNPs from seven candidate genes in the vitamin D pathway (GC, CYP2R1, CYP27A1, CYP27B1, CYP24A1, VDR and RXRA) were determined from whole blood DNA. For survival outcomes, we used log-rank test and a multivariable extended Cox model to estimate adjusted hazard ratio (HR) of death. HR models for each SNP were restricted to patients who identified as White and adjusted for age, sex, smoking status, diagnosis year, and treatment, stratifying baseline hazard by clinical stage and modeling surgical resection as a time-dependent covariate. HR models by quartiles 25(OH)D, additionally adjusted for body mass index (BMI), and timing of blood draw. Our analyses included EA patients with complete information on relevant confounders and predictors of OS. The 25(OH)D analysis included 463 EA patients and 337 deaths, with median follow-up of 23.1 months. The mean 25(OH)D level at diagnosis was 20.7 ng/mL. We found no evidence that OS curves differed across quartiles of vitamin D (log rank p=0.83). In the adjusted extended cox model, we found no difference in OS among the highest quartile and quartiles 2-4 (respectively, HR (95% confidence interval), 0.95 (0.70, 1.31), 1.03 (0.76,1.39), 1.01 (0.73, 1.38), global p=0.97). We did not find evidence of interaction between 25(OH)D and clinical stage (p=0.88) or BMI (p=0.43) on OS, and this relationship did not differ by timing of blood draw. Preliminary SNP analysis included 424 EA patients and 305 deaths, with median follow-up of 24.1 months. Two SNPs in CYP24A1 (SNP, HR (95% confidence interval) per minor allele: rs1570669, 1.23 (1.04-1.45); rs927650, 1.22 (1.04-1.44)), two SNPs in GC (rs4588, 1.25(1.04-1.49); rs7041, 0.84(0.71-0.99)), and one SNP in RXRA (rs7039190, 1.57(1.06, 2.31)) were marginally associated with OS in the multivariable survival model. SNPs were not statistically significant after adjusting p-value for multiple testing. Based on the available evidence from our analyses, markers in the vitamin D pathway at the time of diagnosis are not clinically relevant markers of EA survival. Citation Format: Elizabeth Loehrer, Rebecca Betensky, Ruyang Zhang, Li Su, Edward Giovannucci, David Christiani. Molecular markers of the vitamin D pathway and esophageal adenocarcinoma survival [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 247.