Pathological Predictive Factors of Submucosal Invasion in the Pretreatment Biopsies of Early Gastric Carcinoma.
Sun M Lee, Choel K Park, Kyoung M Kim. The University of Texas Health Science Center, San Antonio; Samsung Medical Center, Seoul, Korea
Background: In an endoscopic resection of early gastric carcinoma (EGC), the presence of submucosal (SM) invasion is one of the most significant risk factors associated with lymph node metastasis and prognosis. Although the diagnostic accuracy rate for SM invasion by endoscopy and ultrasonography has approached to 80%, there are still some cases with SM invasion in the endoscopic resection specimen. Identifying pathologic risk factors to predict SM invasion in the pretreatment biopsy will help selecting cases that are most applicable for endoscopic mucosal/submucosal resection.
Design: Pretreatment biopsies from 60 patients with SM invasive GCs who underwent endoscopic mucosal resection (EMR) between 2002 and 2007 were reviewed. For control, 58 patients with intramucosal GCs during the same period were used. WHO histologic grade, single cell invasion, islands of muscularis mucosa, cribriform pattern, papillary features, desmoplastic reaction, and intraglandular eosinophilic necrotic debris (IEND) were investigated to evaluate pathologic factors associated with SM invasion. Pearson's chi-square test and Fisher's exact test were used for statistical analysis.
Results: The depth of SM invasion was varied from 50 µm to 3000 µm (mean 949 µm). In the biopsy of SM invasive GCs, well to moderately differentiated histology, single cell invasion, islands of muscularis mucosa, cribriform pattern, papillary features, desmoplastic reaction, and IEND were observed in 96.7%, 36.7%, 16.7%, 16.7%, 23.3%, 40%, and 46.7% of cases, respectively, while 100%, 5.2%, 0%, 1.7%, 5.2%, 19%, and 22.4% were observed in the biopsies of intramucosal GCs. WHO histologic grade was not associated with SM invasion. However, single cell invasion, islands of muscularis mucosa, cribriform pattern, papillary features, desmoplastic reaction, and IEND were significantly related to SM invasion in endoscopic mucosal resections (P <0.05). The coexistence of 4 or more of these “SM invasion-associated features” in pretreatment biopsies yielded a 94% sensitivity and 98% specificity in the prediction of SM invasion in the endoscopic mucosal resection specimens.
Conclusions: Our study showed that SM invasion in EGC was not associated with the degree of differentiation. Rather, combined SM invasion-associated features were significant pathologic findings predicting SM invasion and could be used to distinguish SM invasion from intramucosal gastric carcinoma with a high degree of accuracy.
Tuesday, March 1, 2011 9:30 AM
Poster Session III # 156, Tuesday Morning