Accuracy of Macroscopic Intraoperative Diagnosis of Serosal Invasion and Risk of Over- and Underestimation in Gastric Carcinoma

Background Accurate intraoperative diagnosis of serosal invasion is a prerequisite for proper application of invasive procedures, such as intraperitoneal chemohyperthermia, for serosa positive gastric carcinomas. Methods We reviewed the prospectively constructed data of 1,265 gastric cancer patients who underwent surgery between 2007 and 2009. Accuracies of macroscopic diagnoses of serosal invasion were determined by comparing with pathological findings. The risk factors of over- and underestimation of serosal invasion were analyzed in the univariate and multivariate model. Results The accuracy of macroscopic intraoperative diagnosis of serosal invasion was 88%. Serosal invasion was underestimated in 34 of 187 serosa positive patients and overestimated in 117 of 1,078 serosa negative patients; a sensitivity and specificity of 82 and 89%, respectively. When pT1 tumors were excluded, the accuracy, sensitivity, and specificity of macroscopic diagnosis of serosal invasion were 71.5, 81.8, and 65.3%, respectively. Univariate and multivariate analysis revealed that a tumor size of > 4 cm and preoperative CT finding of serosa positive were independent risk factors for macroscopic overestimation as serosal invasion in pT2 gastric cancer. Meanwhile, Borrmann type 1, preoperative CT finding of serosa negative, lesser/posterior surface location, and tumor size of < 4 cm were independent risk factors for underestimation of serosal invasion in pT3 gastric carcinoma. Conclusions The macroscopic diagnosis of serosal invasion is largely consistent with pathological findings. However, great care should be taken with regard to the risk of over- and underestimation of serosal invasion when making a decision for invasive treatments based on macroscopic findings of serosal invasion.