Predicting the Risk of Esophageal Adenocarcinoma in People With Barrett Esophagus

In patients with Barrett esophagus, what is the annual risk of developing invasive esophageal adenocarcinoma?

In Western countries, the prevalence of Barrett esophagus in adults is estimated to be 1% to 2% and, of those, about one-third have long-segment (greater than 2 cm) disease.¹ Risk factors for Barrett esophagus include gastroesophageal reflux disease (GERD), male sex, age older than 50 years, tobacco use, and obesity. Alcohol is not associated with the development of Barrett esophagus, and some studies have shown a protective effect.^2,3 In patients with Barrett esophagus, approximately 0.1% to 0.4% of cases progress to esophageal adenocarcinoma each year. Risk factors for progression include the presence of dysplasia (especially higher grade), a longer segment of Barrett esophagus, and a longer duration of Barrett esophagus.¹

Guidelines from the American College of Gastroenterology (ACG) and the European Society for Gastrointestinal Endoscopy recommend surveillance for esophageal adenocarcinoma in people with Barrett esophagus. Although more frequent upper endoscopy screening intervals are recommended for those with dysplasia, a 5-year interval is recommended for patients with less than a 3-cm segment of Barrett esophagus without dysplasia, and a 3-year interval for those with a longer segment and no dysplasia.^4,5 A risk score that identifies patients at low, intermediate, and high risk for progressing to esophageal adenocarcinoma could help physicians allocate patients to lower vs higher intensity surveillance protocols.

A risk model was developed using data from an Australian case-control study of risk factors for esophageal adenocarcinoma and included 364 patients with the disease and 1,580 controls. The study showed that a model using level of education, body mass index, tobacco use, frequency of GERD symptoms, sex, and use of nonsteroidal anti-inflammatory drugs could classify patients into risk groups ranging from less than 20 to more than 600 cases of esophageal adenocarcinoma per 100,000 men over 5 years.⁶ The area under the receiver operating characteristic curve (AUROC) was 0.76, but the model has not been prospectively validated.

A retrospective cohort study that included 2,591 patients with Barrett esophagus based on an electronic health record query found that 133 cases progressed to dysplasia or esophageal adenocarcinoma, with a median follow-up of 4.4 years. Independent risk factors, including male sex and age 70 years or older (1 point each), and no use of proton pump inhibitors, a segment greater than 3 cm, and esophageal candidiasis (2 points each), were used to create the Barrett’s Esophagus Assessment of Risk score.⁷ The AUROC was 0.81, with a score of 3 or higher identifying high-risk patients and an odds ratio of 9.0 for dysplasia or esophageal adenocarcinoma. This risk score also has not been prospectively validated in a new population.

The largest study to date identified 2,697 patients with Barrett esophagus who were followed up for a median of 5.9 years. During that period, 154 (5.7%) patients developed high-grade dysplasia or esophageal adenocarcinoma. Variables significantly associated with disease progression included male sex, ever smoking cigarettes, increasing length of Barrett esophagus, and low-grade dysplasia at baseline.⁸ Predicting the risk of high-grade dysplasia and esophageal adenocarcinoma is clinically useful because treatment for both conditions is recommended by ACG guidelines.⁵ This risk score was validated in a retrospective cohort study that included a Northern Irish population and was found to have an AUROC of 0.70 (95% CI, 0.63–0.78) in distinguishing patients who will progress to high-grade dysplasia or esophageal adenocarcinoma from those who will not. The risk score and its interpretation are summarized in Table 1.⁸