Background/aims: There are few reports regarding the accuracy of endoscopy in detecting cancer. We investigated the difference in the false-negative rates for cancer detection between gastroscopy and colonoscopy using the records of a population-based cancer registry, which is the most accurate in Japan.
Methodology: Between 1990 and 1995, 51,411 gastroscopic and 7756 colonoscopic examinations were carried out in our hospital. These subjects were matched with the cancer registry and the patients whose gastric or colorectal cancers were missed by examinations were identified.
Results: The false-negative rate for the detection with gastroscopy was 25.8 percent at three years but that with colonoscopy was 11.1 percent. This difference was statically significant p = 0.01.
Conclusions: We concluded that for routine examinations, surveillance after negative endoscopy should be discussed based on the difference in false-negative rates between gastroscopy and colonoscopy.