Validation of an Epigenetic Prognostic Assay to Accurately Risk-Stratify Patients With Barrett Esophagus

Abstract

Introduction: Esophageal adenocarcinoma (EAC) is the second-most lethal cancer in the United States, with Barrett esophagus (BE) being the strongest risk factor. Assessing the future risk of neoplastic progression in patients with BE is difficult; however, high-grade dysplasia (HGD) and early EAC are treatable by endoscopic eradication therapy (EET), with survival rates of 90%. Thus, it would be beneficial to develop a molecular assay to identify high-risk patients, who merit more frequent endoscopic surveillance or EET, as well as low-risk patients, who can avoid EET and undergo less frequent surveillance.

Methods: Deidentified endoscopic biopsies were acquired from 240 patients with BE at 6 centers and confirmed as future progressors or nonprogressors. Tissues were analyzed by a set of methylation-specific biomarker assays. Test performance was assessed in an independent validation set using 4 stratification levels: low risks, low-moderate risks, high-moderate risks, and high risks.

Results: Relative to patients in the low-risk group, high-risk patients were 15.2 times more likely to progress within 5 years to HGD or EAC. For patients in the high-risk category, the average risk of progressing to HGD or EAC within 5 years was 21.5%, 4-fold the BE population prevalence within 5 years, whereas low-risk patients had a progression risk of only 1.85%.

Discussion: This clinical assay, Esopredict, stratifies future neoplastic progression risk to identify higher-risk patients with BE who can benefit from EET or more frequent surveillance and lower-risk patients who can benefit from reduced surveillance.

Keywords: clinical decisions; epigenetics; esophageal cancer; personomics; predictive biomarkers.

Graphical abstract

Figure 1.

Study enrollment cohorts and inclusion criteria. (a) Diagram describing the retrospective patient collection cohorts of training, validation, and independent validation. (b) The inclusion criteria are based on biopsy interval years and progression status. EAC, esophageal adenocarcinoma; HGD, high-grade dysplasia.

Figure 2.

Model development and training. All of these data represent the training set (n = 99). (a) Bivariate plot of biomarkers HPP1 and FBN1 with the x-axis showing normalized methylation values of HPP1 and the y-axis showing normalized methylation values of FBN1. Normalized to the internal control β-Actin. (b) Coefficients and low and high confidence intervals (CI) of fitting multivariate logistic regression models to the data with standardized predictors including SE and z-score (z). (c) Predictiveness curve with prevalence-adjusted probability compared with the model score with cut points chosen represented by dashed vertical lines indicating risk levels. The bold gray line represents the average probability of progression based on prevalence within 5 years. The rug plot across the x-axis shows lines representing patient scores. The red lines indicate a patient that progressed to high-grade dysplasia or esophageal adenocarcinoma, and the black lines indicate a nonprogressor (n = 99). (d) Table reflecting population-adjusted average risk, and estimated number (n) and percent (%) of population-adjusted patients in each risk category and risk level. To adjust for prevalence, the number of nonprogressors in this cohort was multiplied by a factor of 12 (n = 99). NMV, normalized methylation value.

Figure 3.

Validation of model. All of these data represent the validation set (n = 110). (a) Predictiveness curve with prevalence-adjusted probability compared with the model score with locked cut points, represented as dashed lines. The bold gray line represents the average probability of progression based on prevalence in 5 years. The rug plot across the x-axis shows lines that represent patient scores, the red lines indicate a patient that progressed to high-grade dysplasia or esophageal adenocarcinoma, and the black lines indicate a nonprogressor (n = 110). The bold black vertical line in the high-risk level represents a 95% confidence interval, included 5 patients with scores ranging from 43 to 56, with 3 progressors and 2 nonprogressors. (B) A table reflecting population-adjusted average risk, and estimated number (n) and percent (%) of population-adjusted patients in each risk category and risk level. To adjust for prevalence, the number of nonprogressors in this cohort was multiplied by a factor of 12 (n = 110). (c) Estimate the likelihood of progression, i.e., OR, within each risk category. Top figure and table: lower-risk category (i.e., risk levels low and low-moderate) compared with higher-risk category (i.e., risk levels high-moderate and high). Bottom table: Comparing all risk levels to the low-risk level, including LCL and UCL. Owing to prevalence, the number of nonprogressors in this cohort was represented by a multiplication of 12 (n = 110). LCL, lowest confidence level; OR, odds ratio; UCL, upper confidence level.

Figure 4.

Performance of Esopredict. All of these data represent the validation set (n = 110). (a) Table with the number of patients (n) in each risk level and overall separated by nonprogressor and progressor. 1.1 Includes all patients in the validation cohort (n = 110) and 1.2 includes only patients with an index biopsy of NDBE in the validation cohort (n = 93). (b) Includes only patients who had an index biopsy (i.e., assayed) of NDBE, and table with the number of patients (n), population-adjusted percent of patients, and the average prevalence-adjusted risk percent across each risk category and risk level. NDBE, nondysplastic Barrett esophagus.

Figure 5.

Performance in an independent validation cohort. (a) Boxplots of the Esopredict scores (y-axis) of nonprogressor (black, n = 70) and progressor patients in the first validation (gray, n = 40) for the 2 left plots, the right plot (represents independent validation patients that progressed >5 years [n = 31]). The bold line represents the median score, and the box represents the first and third quartiles. (b) Analysis of variance of the 3 patient populations: non-progressor (n = 70), progressor (n = 40) from the validation, and progressions with outcome biopsies >5 years (n = 31).