Skip to content. | Skip to navigation

National Cancer Institute U.S. National Institutes of Health www.cancer.gov

Navigation

Personal tools

You are here: Home / Biomarkers / Meltzer 8-marker panel for Esophageal Adenocarcinoma

Meltzer 8-marker panel for Esophageal Adenocarcinoma

Basics

Attributes

QA State: Accepted
Short Name:

Eight methylation biomarkers - p16, RUNX3, HPP1 (HGNC name TMEFF2), NELL1, TAC1, SST, AKAP12 and CDH13 - were tested in a restrospective multicenter double-blinded validation study for their accuracy in predicting neoplastic progression in Barrett's Esophagus. Hypermethylation of p16, RUNX3 and HPP1 has been show to occur in early Barrett's Esophagus-related neoplastic progression and predicts progression risk. Several of the panel (NELL1, TAC1, SST, AKAP12 and CDH13) were also shown to be methylated early and often in Barrett's Esophagus-related neoplastic progression.

Panel Details

Organs

The following organs have data associated with this biomarker…

Esophagus

Attributes

Phase: Two
QA State: Under Review

Esophageal adenocarcinoma risk in Barrett's esophagus is increased 30- to 125- fold versus the general population. Among all Barrett's esophagus patients neoplastic progression occurs only once per 200 patient-years. Molecular markers (individual or in panel) would be useful to risk-stratify patients for more efficient surveillance endoscopy and to improve the early detection of progression. In 2005, Meltzer et al reported that hypermethylation of p16, RUNX3, and HPP1 occurs early in Barrett's esophagus-associated neoplastic progression and predicts risk. Later, the group developed a tiered risk stratification model to predict progression in Barrett's esophagus using epigenetic and clinical features, and also studied methylation levels and frequencies of individual genes using real-time quantitative methylation-specific PCR in 259 endoscopic esophageal biopsy specimens of different histologies. Among ten genes evaluated, five (NELL1, TAC1, SST, AKAP12, and CDH13) were methylated early and often in Barrett's esophagus-associated progression. In these studies, methylation status and levels correlated inversely with mRNA expression levels.

Supporting Study Data

The following studies/protocols provide evidence supporting Meltzer 8-marker panel for Esophageal Adenocarcinoma indications for the Esophagus…

Barrett's Esophagus Methylation Profiles

We propose a nested case-control study of biomarkers in the setting of BE. By bringing together research institutions with large populations of patients with BE, we will perform a multi-center study of FISH and hypermethylation markers as possible prognostic factors in BE. The centers will select from their cohorts who have progressed to HGD or to adenocarcinoma of the esophagus ("progressors"), and who also donated samples prior to the development of cancer, when their histology was felt to be benign. These subjects will be compared to individuals who have been under endoscopic surveillance, but who have not progressed to HGD or EAC ("non-progressors"). Using this approach, we hope to identify promising markers for risk stratification in BE. We expect to be able to make successful application for a prospective study of markers identified in this case-control study.

View more about this study
Sensitivity/Specificity Information
Notes Sensitivity Specificity Prevalence NPV PPV
Panel-plus-age, combined (2-yr and 4-yr progression models). 195 biopsies of Barrett's Esophagus (145 nonprogressors and 50 progressors). 76.0 80.0 N/A N/A N/A
Panel alone, model of progression within 2 years. 195 biopsies of Barrett's Esophagus (145 nonprogressors and 50 progressors). 61.0 90.0 N/A N/A N/A
Panel alone, model of progression within 2 years. 195 biopsies of Barrett's Esophagus (145 nonprogressors and 50 progressors). 72.0 80.0 N/A N/A N/A
Panel-plus age, model of progression within 2 years. 195 biopsies of Barrett's Esophagus (145 nonprogressors and 50 progressors). 54.0 90.0 N/A N/A N/A
Panel alone, combined (2-yr and 4-yr progression models). 195 biopsies of Barrett's Esophagus (145 nonprogressors and 50 progressors). 44.0 90.0 N/A N/A N/A
Panel-plus-age, combined (2-yr and 4-yr progression models). 195 biopsies of Barrett's Esophagus (145 nonprogressors and 50 progressors). 46.0 90.0 N/A N/A N/A
Panel alone, combined (2-yr and 4-yr progression models). 195 biopsies of Barrett's Esophagus (145 nonprogressors and 50 progressors). 63.0 80.0 N/A N/A N/A
Panel-plus age, model of progression within 2 years. 195 biopsies of Barrett's Esophagus (145 nonprogressors and 50 progressors). 79.0 80.0 N/A N/A N/A
Panel alone, model of progression within 4 years. 195 biopsies of Barrett's Esophagus (145 nonprogressors and 50 progressors). 47.0 90.0 N/A N/A N/A
Assessment of the classification accuracy of a single marker using ROC curve analyses show AUC for HPP1, p16, and RUNX3 are all significantly greater than 0.50. Summarized in Related Resources section, file "NP-P suppl table4.ppt". [8] N/A N/A N/A N/A N/A
Risk stratification of BE patients by predictiveness curves of 8-marker panel, age alone, and 8-marker-plus-age panel in combined, 2-year and 4-year prediction models. A, Predictiveness curve of 8-marker panel in combined model. After rigorous overfitting correction, at risk = 0.1 and = 0.5, 45% and 4% of subjects had estimated risks below 0.1 (LR group) and above 0.5 (HR group), respectively; while remaining 51% had estimated risks between 0.1 and 0.5 (IR group). B, Predictiveness curves of age alone and of 8-marker-plus-age panel in combined model. After rigorous overfitting correction, BE patients were stratified into LR (15%) and IR (85%) groups by age alone. BE patients were stratified into LR (51%), IR (44%) and HR (5%) groups by 8-marker-plus-age panel. C, Predictiveness curve of 8-marker panel in 2-year model. After rigorous overfitting correction, BE patients were stratified into LR (45%), IR (51%) and HR (4%) groups. D, Predictiveness curves of age alone and of 8-marker-plus-age panel in 2-year model. After rigorous overfitting correction, BE patients were stratified into LR (11%) and IR (89%) groups by age alone. BE patients were stratified into LR (52%), IR (44%) and HR (4%) groups by the 8-marker-plus-age panel. E, Predictiveness curve of the 8-marker panel alone in the 4-year model. After rigorous overfitting correction, BE patients were stratified into LR (44%), IR (51%) and HR (5%) groups. F, Predictiveness curves of age alone and of the 8-marker-plus-age panel in the 4-year model. After rigorous overfitting correction, BE patients were stratified into LR (15%) and IR (85%) groups by age alone. BE patients were stratified into LR (51%), IR (44%) and HR (5%) groups by 8-marker-plus-age panel. LR: low-risk; IR, intermediate-risk; HR: high-risk; OC: overfitting correction. Graphs in Related Resources section, file "NP-P Supplementary fig 2.ppt". [8] N/A N/A N/A N/A N/A
Predictiveness curves (risk plots) were used to assess the clinical utility of the combined classification rules in stratifying patients according to risk levels in the target population. Summarized in Related Resources section, file "Jin_Table4_PredictCurve.doc". [7] N/A N/A N/A N/A N/A
Logistic regression analyses of the eight-marker panel: summarized in Related Resources section, file "Jin_Table1_LogisticRegression.doc". [3] N/A N/A N/A N/A N/A
Receiver-operator characteristic (ROC) curves for the uncorrected 8-marker and 8-marker-plus-age panels, overfitting-corrected ROC curves for the 8-marker and 8-marker-plus-age panels, and ROC curves for age alone in the 2-, 4-year, and combined prediction models. A, Uncorrected ROC curve (AUC=0.843) and overfitting-corrected ROC curve (AUC=0.745) for 8-marker panel in 2-year prediction model; shrinkage due to overfitting correction minimal, at 0.098. B, Uncorrected ROC curve (AUC=0.829) and overfitting-corrected ROC curve (AUC=0.720) for 8-marker panel in 4-year prediction model; shrinkage minimal, at 0.109. C, Uncorrected ROC curve (AUC=0.840) and overfitting-corrected ROC curve (AUC=0.732) for 8-marker panel in combined prediction model; shrinkage minimal, at 0.108. D, Uncorrected ROC curve for 8-marker-plus-age panel (AUC=0.858), overfitting-corrected ROC curve (AUC=0.756), and ROC curve for age alone (0.604) in 2-year prediction model; increment over age alone substantial, at 0.152. E, Uncorrected ROC curve for 8-marker-plus-age panel (AUC=0.850), overfitting-corrected ROC curve (AUC=0.744), and ROC curve for age alone (0.630) in 4-year prediction model; increment over age alone substantial, at 0.114. F, Uncorrected ROC curve for 8-marker-plus-age panel (AUC=0.855), overfitting-corrected ROC curve (AUC=0.753), and ROC curve for age alone (0.635) in combined prediction model; increment over age alone substantial, at 0.118. Graphs in Related Resources section, file "NP-P Supplementary fig 1.ppt" . [4] N/A N/A N/A N/A N/A
Clinicopathic features of biopsies are summarized in the Related Resources section, in file "NP-P suppl table1.ppt". [1] N/A N/A N/A N/A N/A
Normalized methylation values of HPP1, p16 and RUNX3 were significantly higher in progressors versus nonprogressors by Wilcoxon test. The remaining five markers did not differ significantly in progressors versus nonprogressors. Summarized in the Related Resources section, in file "NP-P suppl table 3.ppt". [2] N/A N/A N/A N/A N/A
Panel-plus age, model of progression within 4 years. 195 biopsies of Barrett's Esophagus (145 nonprogressors and 50 progressors). 72.0 80.0 N/A N/A N/A
Definition of Barrett's esophagus progressor patients and specimens: Barrett's esophagus progressor patients are Barrett's esophagus subjects with either no dysplasia or low-grade dysplasia who later developed high-grade dysplasia or esophageal adenocarcinoma. Progressor specimens (any Barrett's esophagus tissues obtained prior to the progression endpoint) were divided into three subgroups - those obtained at 0-2 years, 2-4 years or more than 4 years prior to the progression endpoint. Non-progressor specimens were defined as Barrett's esophagus and low-grade dysplasia tissues obtained at 0-2 years, 2-4 years, or more than 4 years before the non-progression follow-up date. Progressors were considered both as a single combined group, and in two tiers: progression within 2 years (tier 1) or 4 years (tier 2). N/A N/A N/A N/A N/A
Panel alone, model of progression within 4 years. 195 biopsies of Barrett's Esophagus (145 nonprogressors and 50 progressors). 61.0 80.0 N/A N/A N/A
Panel-plus age, model of progression within 4 years. 195 biopsies of Barrett's Esophagus (145 nonprogressors and 50 progressors). 45.0 90.0 N/A N/A N/A
Incremental AUC values contributed by an eight-marker-plus-age biomarker panel to that of age alone, summarized in Related Resources section, file "Jin_Table2_IncrementalValues.doc". [5] N/A N/A N/A N/A N/A
Maintaining high specificity to minimize false-positive results, the model still predicted a number of new early diagnoses (diagnoses that would have not occurred earlier without the panel). Summarized in Related Resources section, file "Jin_Table3_SpecSens.doc". [6] N/A N/A N/A N/A N/A
Additional Study-Specific Protocols
Study-Specific Publications
Study-Specific Resources

Organ-Specific Protocols

No organ-specific protocols defined.

Organ-Specific Publications

No organ-specific publications defined.

Organ-Specific Resources

No organ-specific resources defined.

Organs

No associated resources found.

News

The final report of the 2013 Cancer Biomarkers Bioinformatics Workshop is now available.

Announcement 03/06/2014

Thank you to everyone who helped make the 27th EDRN Steering Committee Meeting a success. We look forward to seeing everyone at the 9th EDRN Scientific Workshop from September 8-11, 2014 in Washington D.C.

Announcement