[Purpose]The aim of this study was to clarify the discrepancies in the histologic diagnosis of early gastric cancer between biopsy diagnosis and postoperative pathologic diagnosis, and also to determine the optimal number of endoscopic biopsy specimens required to obtain an accurate diagnosis of early gastric cancer. [Methods]The efficacy of biopsy-based preoperative diagnosis was evaluated in 1330 cases (1083 cases of early gastric adenocarcinoma and 247 cases of gastric adenoma) treated by endoscopic submucosal dissection (ESD) and surgery between January 2004 and July 2010, by comparing the preoperative and postoperative diagnoses. According to the Japanese Gastric Cancer Association classification, epithelial changes present in the biopsy specimens were categorized into Groups III-IV, as well as into the differentiated, undifferentiated or mixed type. Early gastric carcinoma histology was classified into 4 subtypes : pure differentiated (PD), a mixture of differentiated and undifferentiated (differentiated-type predominant : MD ; undifferentiated-type predominant : MU) and pure undifferentiated (PU). [Result]The 359 cases diagnosed as Group III by preoperative biopsy were as follows : adenoma, 66.6% ; PD, 32.9% ; MD, 0.6% ; MU, 0% : PU, 0%. The corresponding percentages for the other classifications were 3.4%, 92.6%, 4.0%, 0% and 0%, respectively (Group IV ; 176 cases), 0.4%, 88.6%, 8.9%, 1.4% and 0.7%, respectively (undifferentiated ; 174 cases), 0%, 4.6%, 3.4%, 8.6% and 83.3%, respectively (differentiated ; 570 cases), and 0%, 25.5%, 33.3%, 19.6% and 21.6%, respectively (mixed ; 51 cases). This study demonstrated that two biopsy specimens were sufficient to provide the correct pathologic diagnoses in early gastrointestinal cancer, and additional biopsies did not increase the accuracy. [Conclusion]Discrepancies between preoperative biopsy-based diagnoses and postoperative pathological diagnoses are not uncommon. Therefore, clinical management based on careful endoscopic observation is of vital importance. Nevertheless, increasing the number of biopsies does not improve the accuracy of histological diagnosis (with two being the optimal number of biopsy specimens). In addition, ESD may at times be required for total biopsy.
View full abstract