A cohort study of missed and new cancers after esophagogastroduodenoscopy

Am J Gastroenterol. 2010 Jun;105(6):1292-7. doi: 10.1038/ajg.2009.736. Epub 2010 Jan 12.


Objectives: Little is known about missed rates of upper gastrointestinal cancer (UGC) in Western populations, with most data originating from Japanese centers quoting high missed rates of 23.5-25.8%. The objective of this study was to better define missed rates of esophagogastroduodenoscopy (EGD) and the natural history of UGC in a Western population that underwent an initial EGD without cancer, but were subsequently diagnosed with a UGC. Our hypothesis was that a normal EGD rarely misses the detection of UGC.

Methods: This is a retrospective cohort study. A prospectively maintained electronic database was used to identify all patients who underwent EGD between 1990 and 2004 at the study institution. Patients in this cohort who were diagnosed with UGC before 2006 were identified through the Western Australian Cancer Registry. We defined missed cancers as those diagnosed within 1 year of EGD, possible missed cancers as those diagnosed 1-3 years after EGD, and new cancers as those diagnosed more than 3 years after EGD. This study had no interventions and was conducted at a tertiary referral center. The main outcome measurement included UGC.

Results: Of the 28,064 EGDs performed, UGC was diagnosed subsequent to the procedure in 116 cases (0.41%). There were 29 missed cancers, 26 possible missed cancers, and 75 new cancers. Of the missed cancers, 11 were esophageal, 15 were gastric, and 3 were duodenal. In 69% (n=20) of the missed cancers, an abnormality was described at the site of malignancy. In 59% (n=17) of the missed cancers, the indication for EGD was an alarm symptom of dysphagia or suspected blood loss. In an univariate analysis, the presence of an alarm symptom was related to missed cancers, whereas operator experience, trainee participation, and usage of newer equipment were not. One of the main limitations of this study is that it was a retrospective review.

Conclusions: UGC is rare after normal EGD, confirming the high accuracy of EGD. Institutional approval was granted for the conduct of this study.

MeSH terms

  • Cohort Studies
  • Diagnostic Errors*
  • Duodenal Neoplasms / diagnosis*
  • Endoscopy, Digestive System*
  • Esophageal Neoplasms / diagnosis*
  • Female
  • Humans
  • Male
  • Retrospective Studies
  • Stomach Neoplasms / diagnosis*