Endoscopic discrimination of gastric ulcers

Because endoscopists are concerned about misidentifying a gastric neoplasm as a benign gastric ulceration (GU), routine endoscopic biopsy and documentation of GU healing has been recommended. With the decreasing incidence of gastric cancer in the United States, the increased use of nonsteroidal anti-inflammatory drugs, and the concern over medical care costs, this practice standard has been questioned. To study the utility of endoscopic GU follow-up, we reviewed all cases of GU in the Duke GI-Trac database over a 7-year period. We found 1,189 patients diagnosed with GU who underwent 1,698 upper endoscopies. Of these, 130 patients underwent serial esophagogastroduodenoscopy until GU healing was documented. We identified 19 cases of gastric neoplasm. Endoscopic impression correlated with histology as follows: positive predictive value, 36%; negative predictive value, 99.3%; specificity, 90%; and sensitivity, 84%. Two of the three cases of GU in which the endoscopist's impression was benign but histology revealed malignancy occurred in the setting of an acute gastrointestinal bleed. We conclude that more than 99% of the time an endoscopist's initial impression that a GU is benign is correct. Using the Medicare reimbursement scheme, approximately $150,000 would be spent to detect one early gastric cancer. Our results further question the utility of serial endoscopic evaluation of GUs until healing.