Missed diagnosis of early gastric cancer or high-grade intraepithelial neoplasia

2013 
AIM: To investigate the causes of missed diagnosis of early gastric cancer (EGC) or high-grade intraepithelial neoplasia (HGIN) in Chongqing, China. METHODS: The present study summarizes 103 cases of EGC/HGIN detected by esophagogastroduodenoscopy (EGD) and pathological analysis from January 2010 to December 2011. Dimethyl silicone oil was administrated orally 15 min before the EGD procedures. The stomach was cleaned by repeated washing with saline when the gastroscope entered the stomach cavity. Suspected EGC lesions were subject to conventional biopsy sampling and pathological examinations. The correlation between lesion locations, endoscopic morphology of cancerous sites, training level of the examiners, pathological biopsies, and missed diagnosis was analyzed. RESULTS: Twenty-three cases were missed among the 103 cases (22.23%) of EGC/HGIN. The rate of missed EGC in the gastroesophageal junction (8/19, 42.1%) was significantly higher than at other sites (15/84, 17.86%) (χ2 = 5.253, P = 0.022). In contrast, the rate of missed EGC in the lower stomach body (2/14, 14.29%) was lower than at other sites (21/89, 23.6%), but there were no significant differences (χ2 = 0.289, P = 0.591). The rate of missed EGC in the gastric antrum (5/33, 15.15%) was lower than at other sites (18/70, 25.71%), but there were no significant differences (χ2 = 1.443, P = 0.230). Endoscopists from less prestigious hospitals were more prone to not diagnosing EGC than those from more prestigious hospitals (χ2 = 4.261, P = 0.039). When the number of biopsies was 4 biopsies (3/23, 13.04%) (P < 0.001). In addition, there was no significant difference in the rate of missed diagnosis in patients with 1-3 biopsy specimens (χ2 = 0.141, P = 0.932). CONCLUSION: Endoscopists should have a clear understanding of the anatomical characteristics of the esophagus/stomach, and endoscopic identification of early lesions increases with the number of biopsies.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    26
    References
    32
    Citations
    NaN
    KQI
    []