Objectives: Esophagogastroduodenoscopy (EGD) is considered a very effective method to identify gastric cancer (GC). However, the existence of missed lesions has been frequently discussed. This systematic review and meta-analysis aimed at assessing the magnitude of missing GC diagnosis with EGD and its predictive factors.
Methods: MEDLINE was searched to identify all studies assessing and reporting the proportion of missed GC diagnosis with EGD. Pooled proportion and negative predictive values were computed using the random-effects model and heterogeneity was assessed using the Cochrane Q-test and I.
Results: The studies included (n=22) were grouped by study design. The pooled negative predictive value was 99.7% (95% confidence interval 99.6-99.9%). Missed GCs proportion was 9.4% (95% confidence interval 5.7-13.1%), being 10.0% in studies including patients with negative EGD followed over time, 8.3% in studies including patients with GC, and 23.3% in studies evaluating the proportion of missed synchronous lesions. Mainly, missed cancers were located in the gastric body both in Eastern and in Western studies (39 and 47%, respectively). The majority of missed GCs were adenocarcinomas. Younger age (<55 years), female sex, marked gastric atrophy, gastric adenoma or ulcer, and inadequate number of biopsy fragments were reported as predictive factors for diagnostic failure.
Conclusion: EGD is a very effective method to rule out GC. However, missing GC with EGD is not uncommon, with one out of 10 cancers being potentially missed. Interestingly, lesions were more often missed in the body and therefore a more rigorous protocol for endoscopy and biopsy should be implemented worldwide.