Incremental benefit of preoperative EUS for the detection of pancreatic neuroendocrine tumors: a meta-analysis

Background Current guidelines recommend CT scan or magnetic resonance imaging as the initial imaging modalities for the work-up of suspected pancreatic neuroendocrine tumors (PNETs). Objective To determine the incremental benefit of preoperative EUS (IB EUS ) for the detection of suspected PNETs after other investigative modalities have been attempted. Design This systematic review searched MEDLINE, EMBASE, bibliographies of included articles, and conference proceedings for studies reporting original data regarding the preoperative detection of PNETs. Pooled IB EUS was calculated by using random effects models. Heterogeneity was explored by using stratified meta-analysis and meta-regression. Evidence of small-study effects was assessed by using funnel plots and the Begg test. Patients Patients with suspected PNETs. Interventions EUS evaluation. Main Outcome Measurements The pooled IB EUS for the detection of PNETs after CT scan, with or without additional investigative modalities. Results Among 4505 citations identified, we included 17 cohort studies (612 patients). EUS identified PNETs in 97% of cases. Improved PNET identification with EUS was observed in all of the studies. After adjusting for small-study effects, meta-analysis showed that EUS alone could identify PNETs in approximately 1 in 4 patients (adjusted IB EUS 26%; 95% confidence interval, 17%-37%). The pooled IB EUS varied based on the study design, study size, type of CT scan used, and the number of modalities used prior to EUS. Limitations The majority of included studies were retrospective. Small-study effects were observed. Conclusion Preoperative EUS is associated with an increase in PNET detection after other modalities are attempted.