Abstract
Replacing x-ray examinations, endoscopy has become the mainstream method applied to the detection of gastric cancers. However, failures can still occur. In order to investigate the accuracy of gastroscopy, 88,535 gastroscopic examinations conducted between 1984 and 1995 have been investigated for comparison with the most accurate case in the local governmental population-based cancer registry data in Japan. In the case of 343 examinations, no patients were initially diagnosed as having gastric cancer, and later these patients were diagnosed as having gastric cancer within three years after the initial examinations (false-negatives). For the same period, 1,201 examinations were identified as true-positives. Thus, the false-negative rate over the three-year period was 22.2 percent. This rate was significantly higher for the latter half of the period. Also, a higher rate was found among male subjects and among those who were examined by intern endoscopists with less than 10 years' experience. The percentage of advanced gastric cancers identified among those who had been diagnosed as false-negative was higher for tumors located in the cardia-fornix or upper gastric body than for those in other regions. It has been concluded that the false-negative rate associated with the detection of gastric cancer by means of endoscopy is higher than expected. We emphasize that endoscopists should be provided with adequate training in order to enable them to examine all parts of the stomach so that they can detect even small or flat lesions, and can conduct sufficiently competent biopsies as well.
