Prevalence and Clinical Significance of Gastric‐Type Mucosal Lesions in Non‐Ampullary Duodenal Epithelial Tumors Based on Mucin Phenotype
Kazuya Ohno, Eiji Nakatani, Takafumi Kurokami, Asami Kawai, Masanori Matsuda, Yuichi Masui, Tatsunori Satoh, Taiyo Hirata, Shodai Takeda, Yoshihiro Uehara, Kodai Baba, Yuka Ozaki, Tomoki Hikichi, Haruna Takahashi, Koki Sakaguchi, Kazuhito Honda, Tomomi Maeda, Moeri Hotta, Naofumi Nakamura, Makoto Suzuki, Ken HarumaABSTRACT
Objectives
Non‐ampullary duodenal epithelial tumors differ in clinicopathological features according to mucin phenotype. Gastric‐type tumors have distinct biological features; however, their associated mucosal background remains unclear. In this study, gastric foveolar metaplasia, gastric heterotopia, and Brunner's gland hyperplasia were collectively referred to as gastric‐type mucosal lesions (G‐MLs). We aimed to assess their prevalence and associations with mucin phenotype and clinicopathological factors.
Methods
We retrospectively analyzed the data of 130 patients with 130 resected non‐ampullary duodenal epithelial tumors at a single center. Tumors were classified according to mucin phenotype. The prevalence of concomitant G‐MLs and associated factors was evaluated. Exploratory Firth‐penalized logistic regression analysis was used to assess the association between gastric‐type tumors and G‐ML presence.
Results
Concomitant G‐MLs were identified in 44 patients (33.8%). Their prevalence was higher in gastric‐type tumors (13/16; 81.3%) than in gastric–intestinal‐type tumors (10/37; 27.0%) and intestinal‐type tumors (21/77; 27.3%). G‐MLs were associated with mild endoscopic gastric mucosal atrophy and smoking status and were inversely associated with current Helicobacter pylori infection. In the exploratory fully adjusted model, gastric‐type tumors remained positively associated with the presence of G‐MLs (adjusted odds ratio, 7.00; 95% confidence interval, 1.98–30.95; p = 0.002).
Conclusions
G‐MLs were present in approximately one‐third of patients and were most frequent in gastric‐type tumors. These findings indicate an association between G‐MLs and gastric‐type tumors; however, the cross‐sectional design does not allow inference of a biological relationship.
Trial Registration
N/A.