Accuracy of Endoscopic Biopsies for Determining Tumor Grade in Pre-resection Evaluation of Esophageal Adenocarcinoma

Abstract

Background and Aims

Endoscopic resection (ER) can be a curative treatment for early esophageal adenocarcinoma (EAC). Poor tumor differentiation is a histologic feature of noncurative ER and, when seen on preoperative biopsies, may lead to avoidance of endoscopic treatment. However, the accuracy of tumor grade on preoperative endoscopic biopsies remains unclear. The aim of this study was to determine the accuracy of tumor differentiation on endoscopic forceps biopsy compared with endoscopically or surgically resected tissue.

Methods

This retrospective cohort study from 2 tertiary referral centers compared tumor differentiation from preoperative biopsies with endoscopically or surgically resected EAC samples. Accuracy (%) of preoperative biopsy and agreement analyses (Gwet's AC2) were performed.

Results

In total, 346 EAC lesions (225 esophagectomies, 121 ERs) were included. The overall accuracy and reliability of EAC grade on preoperative biopsies compared with postoperative tissue from ER or esophagectomy was 74% (114/154), with a substantial agreement coefficient (Gwet's AC2 0.78 [0.72-0.85]; P < 0.001). The accuracy of ER was only 56%, with moderate agreement (Gwet's AC2 0.60 [0.46-0.73]; P < 0.001). Poorly differentiated tumors were downgraded to moderately differentiated in 19.6% (20/102) of tumors, and conversely, moderately differentiated tumors were upgraded to poorly differentiated in 22.7% (37/163) of tumors. For patients with T1a tumors, poorly differentiated tumors on preoperative biopsies were downgraded to moderately differentiated tumors in 40% (4/10) of cases after resection.

Conclusions

The accuracy of tumor grade of EAC on preprocedural biopsies is suboptimal and should not be depended upon solely to make decisions regarding endoscopic treatment.