Tu1488 PREVALENCE AND RISK FACTORS FOR LYMPH NODE METASTASIS AFTER NONCURATIVE ENDOSCOPIC RESECTION FOR EARLY ASTRIC CANCER: A SYSTEMATIC REVIEW AND META-ANALYSIS

Abstract

Listen

Translate article

Additional surgery for patients who do not meet the curability criteria after endoscopic resection (ER) for early gastric cancer (EGC) may be excessive due to the relatively low rate of lymph node metastasis (LNM) in such patients. However, the prevalence and risk factors for LNM after noncurative ER have not been consistent across studies. Hence, we aimed to identify the prevalence and risk factors of LNM in patients with noncurative ER for EGC. We performed a systematic review of electronic databases through August 10, 2018 to identify cohort studies with patients who underwent additional surgery after noncurative ER for EGC. This is reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses 2009 guidelines. Missing data were collected from corresponding authors of the included studies. The primary outcome was a pooled prevalence rate of LNM at additional surgery in patients diagnosed as noncurative ER for EGC. The secondary outcome was to reveal pathological risk factors for LNM, including tumor size, deep submucosal invasion, differentiation, lymphatic invasion, vascular invasion, lymphovascular invasion, ulceration (scar), and positive vertical margin. The prevalence of LNM in such patients was extracted for all studies. Odds ratios were combined using random effects meta-analyses to assess the risk of LNM, when possible. We identified 24 studies comprising 3,877 patients with 311 having LNM (pooled prevalence, 8.1%, 95% confidence interval, 7.3%–9.0%). Summary results concerning the prevalence of LNM after noncurative ER for EGC in the subgroup analyses are present in Table 1. The risk of LNM was significantly increased in lymphatic invasion (odds ratio [95% confidence interval] =4.22 [2.88–6.19]), lymphovascular invasion (4.17 [2.90–5.99]), vascular invasion (2.38 [1.65–3.44]), positive vertical margin (2.16 [1.59–2.93]), deep submucosal invasion (2.14 [1.48–3.09]), and tumor size >30 mm (1.77 [1.31–2.40]) (Table 2). In contrast, there was no significant association between undifferentiated-type or ulceration (scar) and LNM. There was no heterogeneity across the studies in each analysis (I2 =0% in all analysis). Funnel plot asymmetry was not present among the studies reporting the prevalence and risk of LNM other than vascular invasion (Egger’s test, P =0.003). When studies were restricted to the studies that evaluated the adjusted odds ratio, the risk of vascular invasion for LNM did not reach statistical significance. In this meta-analysis including the largest cases to date, several pathological factors, most notably lymphatic invasion and lymphovascular invasion, were associated with LNM in patients with noncurative ER for EGC. Lymphatic and vascular invasion should be assessed separately instead of lymphovascular invasion. (PROSPERO CRD42018109996)View Large Image Figure ViewerDownload Hi-res image Download (PPT)