372 An Automated, Quantitative Multiplex Immunofluorescence Assay Accurately Risk Stratifies Barrett’s Esophagus Patients With a Community-Based Diagnosis of Low-Grade Dysplasia at a Rate Comparable to Expert Pathologists

Abstract

INTRODUCTION: The majority of community-based LGD diagnoses are down-staged to non-dysplastic (ND) BE when reviewed by an expert pathologist, whereas patients with confirmed LGD have a 10% annual risk of progression to high-grade dysplasia (HGD) or cancer (Ca). Although recommended by guidelines, expert review is poorly defined, prone to observer variation, and not widely available. Recent studies suggest that an automated, quantitative multiplex immunofluorescence assay may identify NDBE patients with an increased risk of malignant progression. We investigated if this assay also allows objective risk stratification for cases with a community-based diagnosis of LGD at a rate comparable to expert pathologists. METHODS: A blinded nested case-control cohort was derived from the screening cohort of a randomized controlled trial of SUrveillance vs. RadioFrequency ablation (SURF) for BE patients with LGD. 60 patients with a community-based diagnosis of LGD of whom 30 progressed to HGD/Ca and 30 patients who did not progress were matched for age, sex and maximal BE length. All random biopsy levels of the baseline endoscopy were independently reviewed by 3 expert pathologists with an international reputation. All levels were additionally tested by a multiplex biomarker assay. The assay classifies patients into low-risk (LR), intermediate-risk (IR) and high-risk (HR) for progression to HGD/Ca within 5 years. A histological revision diagnosis of NDBE was considered as low- (LR), while indefinite for dysplasia (IND) and LGD were considered as high-risk (HR) for progression. RESULTS: 60 BE patients (52 male), mean age 63 ± 9 years, were studied. Median time between baseline endoscopy and progression was 2.5 years (1.3-5.4). Accuracy, sensitivity and specificity of the biomarker assay vs. the 3 expert pathologists were 73% vs 72-77%, 66% vs 77-80% and 80% vs 63-77%, respectively. 20/26 (78%) patients classified as IR/HR by the biomarker panel and 69-77% classified as HR (IND/LGD) by the 3 expert pathologists progressed to HGD/Ca. CONCLUSION: A quantitative multiplex immunofluorescence assay risk stratified BE patients with a community-based diagnosis of LGD with an accuracy comparable to three renowned expert pathologists. The assay allows for automated, objective risk stratification which may be a practical and effective solution to the lack of standardization of expert pathology review of LGD as advocated by all guidelines.