Background and aims: Gastric cancer is classified into differentiated and undifferentiated types according to the degree of glandular differentiation. Undifferentiated-type early gastric cancer (EGC) carries a higher risk of lymph-node metastasis than differentiated type, and therefore the indication criteria for endoscopic resection differ. This study aimed to clarify the ability of clinical predictors to distinguish between differentiated-type and undifferentiated-type EGCs.
Methods: This was a post hoc study of a multicenter prospective trial carried out in 5 Japanese hospitals, including 343 patients with cT1 EGC of ≥10 mm. According to the protocol, age, sex, and endoscopic findings of cancer (diameter, location, macroscopic type, and invasion depth) were evaluated, and the final diagnosis was confirmed from resected specimens. We evaluated the associations between these clinical factors and the histologic type of cancer and calculated the ability of the factors to diagnose differentiated-type EGC. The diagnostic ability of forceps biopsy was also calculated as a reference.
Results: Multivariate analysis identified older age (≥72 years), male sex, larger tumor size (>30 mm), elevated type, and shallower invasion depth (cT1a) as independent significant predictors for differentiated-type EGC, with elevated type showing the highest positive likelihood ratio. The sensitivity, specificity, accuracy, and positive and negative likelihood ratios of elevated type for differentiated-type EGC were 24%, 99%, 38%, 15.7, and 0.77, respectively, compared with 96%, 86%, 95%, 7.0, and 0.04 for forceps biopsy.
Conclusions: Endoscopic elevated type is a significant predictor for differentiated-type EGC and may exclude undifferentiated-type EGC without the need for forceps biopsy.
Copyright © 2018 American Society for Gastrointestinal Endoscopy. Published by Elsevier Inc. All rights reserved.