[The reliability of fiber endoscopic estimation of the size of gastroduodenal ulcer].
The areas of 22 regions in a gastro-duodenal model were estimated endoscopically by 11 examiners. The procedure was then repeated with a biopsy forceps as the reference for measurement. Without the biopsy forceps, 86% of the areas were underestimated. Measurement with the biopsy forceps reduced the deviations of the estimates and the scatter of these but 75% were still underestimated. The estimates showed considerable inter-observer variation and limited reproducibility. Endoscopically estimated alteration in area of an ulcer should not be accepted as clinical documentation for healing.