Abstract
Background: Patients with non-dysplastic Barrett's esophagus (ND-BE) and low-grade dysplasia (LGD) are typically monitored by periodic endoscopic surveillance, while those with high-grade dysplasia (HGD) and esophageal adenocarcinoma (EAC) are usually treated by more aggressive interventions like endoscopic mucosal resection, ablation or surgery. Therefore, the accurate grading of dysplasia in Barrett's esophagus (BE) is essential for proper patient care. However, there is significant interobserver and intraobserver variability in the histologic grading of BE dysplasia. The objective of this study was to create an immunohistochemical (IHC) panel that facilitates the grading of BE dysplasia and can be used as an adjunct to histology in challenging cases. Methods: 100 BE biopsies were re-graded for dysplasia independently by 3 subspecialized gastrointestinal pathologists. IHC staining for CDX2, p120ctn, c-Myc and Jagged1 proteins was then performed and assessed by two separate methods of semi-quantitative scoring. Scores were integrated using a principal component analysis (PCA) and receiver operating characteristic (ROC) curve. Results: Principal component analysis demonstrated the ability of this panel of proteins to segregate ND-BE/LGD and HGD/EAC, as the expression of the four proteins is significantly altered between the two subsets. Analysis of the receiver operating characteristic curve showed that this panel has the potential to aid in the grading of dysplasia in these two subcategories with both high sensitivity and specificity. While not able to discriminate between ND-BE and LGD, this panel of four proteins may be used as an adjunct to help discriminate subsets of ND-BE/LGD from HGD/EAC. Conclusions: We propose that the maximum utility of this IHC panel of CDX2, p120ctn, c-Myc, and Jagged1 proteins would be to distinguish between LGD and HGD in histologically challenging cases, given the aggressive interventions still used for HGD in many institutions, and hence may aid in the optimal patient management. The results of this initial study are promising, though further validation is needed before this panel can be used clinically, including future randomized prospective studies with larger patient cohorts from diverse locations.