Finding metabolites relevant to esophageal adenocarcinoma

Science Spotlight

Finding metabolites relevant to esophageal adenocarcinoma

From the Vaughan Lab, Public Health Sciences Division

April 17, 2017

Esophageal adenocarcinoma is on the rise, with an eight-fold increase in incidence over recent decades. As treatment has improved only modestly over time – median survival with this cancer is currently only about one year – prevention and early diagnosis remain the most potent means to reduce mortality. While several risk factors and precursor states for esophageal adenocarcinoma have been identified, current screening and surveillance modalities are invasive, expensive, and have room for improvement. Metabolomic approaches may provide a new and less-invasive approach for identifying novel biomarkers associated with cancer development. Dr. Matthew Buas, a former Fred Hutch post-doc working with Dr. Thomas Vaughan and colleagues in the Public Health Sciences Division, evaluated 57 metabolites for differences between patients at different stages of progression to cancer. In a recent issue of Metabolomics, the authors found several metabolites that demonstrated discriminatory abilities.

Roughly one in five American adults have recurrent symptoms of gastroesophageal reflux disease (GERD), where stomach acid escapes up into the esophagus. Chronic GERD can lead to Barrett’s esophagus (BE), an inflammatory condition where the normal lining of the esophagus is replaced by a metaplastic epithelium. While most esophageal adenocarcinoma cases develop within this BE tissue, 95% of BE patients will never progress to cancer. Current screening methods for esophageal adenocarcinoma in BE patients are invasive, and involve inserting a viewing tube down the throat and taking a biopsy. As such, there is a need for new methods that could more easily identify patients who have, or are at risk for progressing to, esophageal adenocarcinoma. Metabolomics approaches provide a sensitive and comprehensive snapshot of the state of biological systems, and are a promising avenue for identifying molecular signatures associated with cancer.

To evaluate whether serum metabolite levels might differ between patients with GERD, BE, and early stage esophageal adenocarcinoma, the authors utilized data and archived blood samples from 322 participants across two studies based in the Seattle region: the Study of Reflux Disease and the Seattle BE Project. The authors measured 57 metabolites representing multiple metabolic pathways including glycolysis, the citric acid cycle, and amino acid, nucleotide, and fatty acid metabolism. These metabolite levels were then assessed for differences between patient groups. The authors found nine metabolites that differed for BE versus GERD patients, and four that were different between BE and early esophageal adenocarcinoma patients. Several of these differential metabolites have been linked to inflammatory processes, which are thought to contribute significantly to BE and carcinogenesis.

ROC curve for metabolite classifier

Estimated accuracy of a multivariate classifier based on four metabolites found to differ between patients with Barrett’s esophagus and esophageal adenocarcinoma, based on 100 iterations of Monte Carlo cross validation.

Image modified from the publication

The authors went on to evaluate whether combining these serum metabolites into multi-marker signatures might be useful for differentiating between these groups. Classification models incorporating the identified metabolites yielded moderate discrimination between BE and esophageal adenocarcinoma, with an area under the curve of 0.75 (see figure). The ability of identified metabolites to distinguish GERD from BE was more modest, with an area under the curve of 0.64. These results suggest that the serum metabolome might provide important clinical information for the identification of esophageal disease and progression to cancer. Combining these metabolites with other markers or risk factors has the potential for further improvements.

While follow-up studies are needed, this study provides evidence that differences in serum metabolites are detectable in intermediate disease states prior to the development of esophageal adenocarcinoma. Importantly, this study focused on metabolite differences between different stages of disease progression, rather than simply cancer versus cancer free, providing information that is more relevant to clinical settings. If further validated, the biomarkers from this study may help improve future screening and surveillance approaches for this often-fatal cancer.


Also contributing to this project from the Fred Hutch were Ms. Lynn Onstad, and Drs. Brian Reid and Daniel Raftery.

Funding for this study was provided by the National Cancer Institute.


Buas MF, Gu H, Djukovic D, Zhu J, Onstad L, Reid BJ, Raftery D, Vaughan TL. 2017 Candidate serum metabolite biomarkers for differentiating gastroesophageal reflux disease, Barrett's esophagus, and high-grade dysplasia/esophageal adenocarcinoma. Metabolomics;13(3). pii: 23. doi: 10.1007/s11306-016-1154-y.