Molecular Markers of Vitamin D and Esophageal Adenocarcinoma Prognosis

Abstract

Esophageal cancer is the ninth leading cause of cancer and the fifth leading cause of cancer death globally, with two predominant histologies: squamous cell carcinoma and adenocarcinoma. In the last few decades, esophageal adenocarcinoma (EAC) has become the predominant histology in the U.S. Despite some improvements in treatment, EAC still has a five-year survival of less than 20%. Moreover, EAC has few prognostic markers and few if any modifiable prognostic factors. Vitamin D is one potential factor that has garnered attention for it’s role in tumor suppressive effects and it’s potential to be modified. To understand a the role of vitamin D in stalling cancer progression, we also need to study underlying physiologic and metabolic factors of the vitamin D pathway. Therefore, we utilized an ongoing survivor study of esophageal cancer patients from Massachusetts General Hospital to study the association between circulating serum levels of 25(OH)D, genetic variants in candidate genes in the vitamin D pathway, and body mass index with overall survival in EAC patients. We recruited patients around their time of diagnosis, and collected serum and whole blood from patients for biomarker and genetic analyses. Patients completed a baseline questionnaire about demographic and covariate information. Medical records were used to obtain clinical variables such as treatment regimen, pathology, and outcome information. We used extended cox models to estimate adjusted hazard ratios in all studies, adjusting for relevant confounders, modeling surgery as a time dependent covariate, and stratifying the baseline hazards by clinical stage. We found circulating levels of 25(OH)D were not associated with overall survival in EAC, and the relationship was not modified by stage or by body mass index. Genetic variants in the CYP24A1 and CYP27B1 genes were marginally associated with overall survival, but were not significant after multiple testing. Overweight and obese patients at diagnosis had reduced hazards of death, whereas patients who lost substantial weight leading up to diagnosis had significantly increased hazard of death, independent of BMI at diagnosis. Moreover, the association between change in BMI and overall survival was modified by average adult weight, with substantial weight loss being worse for patients who were in the healthy to underweight range as adults (≤27.5 kg/m2). The findings in this dissertation help guide our understanding of the physiologic and molecular differences that might drive different clinical courses in EAC.