340 Proportion of Missed Cancer and High-grade Dysplasia During Barrett's Esophagus Diagnosis: a Systematic Review and Meta-Analysis

Abstract

340 Proportion of Missed Cancer and High-grade Dysplasia During Barrett’s Esophagus Diagnosis: a Systematic Review and MetaAnalysis Kavel Visrodia*, Siddharth Singh, Rajesh Krishnamoorthi, Prasad G. Iyer, David A. Katzka Mayo Clinic Rochester, Rochester, MN Background: Esophageal adenocarcinoma (EAC) and/or high-grade dysplasia (HGD) detected within one year of Barrett’s esophagus (BE) diagnosis is considered prevalent disease and therefore likely to have been “missed” during the initial endoscopy. Recent large BE surveillance studies suggest a sizeable portion of all EAC/HGD is found within this timeframe. This stands in contrast to the emphasis on BE surveillance occurring years after diagnosis. Aim: We conducted a systematic review and meta-analysis comparing the proportion of missed vs incident EAC/HGD detected after a negative index exam (IE). Methods: We performed a systematic literature search of the PubMed, EMBASE, and Web of Science databases (August 2014) for studies reporting the timing of EAC/HGD development in patients with BE during a follow-up of R 2 years. Missed EAC/HGD was defined as those developing within 12 years of BE diagnosis, after a negative IE. Incident EAC/HGD was defined as those occurring after the timeframe used in each study to define missed EAC/HGD. Prevalent EAC/HGD was defined as those diagnosed at the time of IE, regardless of presence or absence of symptoms. Pooled estimates for proportions of missed vs incident EAC/HGD, and prevalent+missed vs incident EAC/HGD were calculated for all studies; sensitivity analysis was performed for studies with BE cohorts comprised of only patients with baseline non-dysplastic BE (NDBE) or low-grade dysplasia (LGD). Results: 53 studies met the inclusion criteria; 51 had a missed EAC/HGD cutoff% 12 months. Meta-analysis of 38 studies and 1,321 patients with EAC showed that after a negative IE, 1/3 were classified as having missed EAC (pooled prevalence, 33.5%; 95% CI, 27.5-40.5) and the remainder as having incident EAC. Results from analysis of 20 studies including 1,481 patients with EAC/HGD were identical (33.5%; 95% CI, 23.9-44.7). On restricting analysis to a subset of patients with NDBE or LGD, the proportions of missed EAC (26.4%; 95% CI, 16.1-40.0) or EAC/HGD (34.8%; 95% CI, 24.7-46.5) were slightly lower or similar, respectively. As expected, with increasing duration of follow-up, the proportion of all EAC classified as incident cases increased: 2-5 years (nZ20 studies): 39.0% vs 61.0%; 5-8 years (nZ12 studies): 26.7% vs 73.3%; O 8 years (nZ4 studies): 12.8% vs 87.2%. In studies reporting data on prevalent, missed, and incident EAC (nZ20 studies and 615 patients with EAC), the majority of EAC were classified as prevalent or missed (83.5%; 95% CI, 77.4-88.2) vs incident (16.5%, 95% CI, 11.8-22.6). Conclusions: One-fourth EAC and 1/3 EAC/ HGD detected during surveillance are potentially present and missed at the time of initial BE diagnosis. Therefore, efforts to improve detection of EAC/HGD at or soon after BE diagnosis should be increased and perhaps even shifted from long-term surveillance. Table 1A. Proportion of missed and incident EAC and EAC/HGD in all BE cohorts, independent of baseline grade of dysplasia www.giejournal Missed (% Missed+Incident EAC)