Efficacy and Safety of Endoscopic Ultrasound-Guided Radiofrequency Ablation for Pancreatic Neuroendocrine Tumors: A Systematic Review and Metanalysis

Abstract

Introduction: The development of dedicated endoscopes and the technical evolution of endoscopic ultrasound (EUS) have allowed a direct approach to pancreatic neoplastic lesions both for diagnosis and treatment. Among the more promising targets are pancreatic neuroendocrine tumors (Pan-NETs). Aim: to describe the evolution of endoscopic ultrasound-guided radiofrequency ablation (EUS-RFA) with particular attention to the treatment of PanNETs, focusing on safety and clinical efficacy of the technique. Methods: MEDLINE, Scopus, and Cochrane Library databases were searched for studies reporting about EUS-RFA for the treatment of PanNETs. Studies with outcomes of interest were selected and results were reported to describe clinical success, complications, fol-low-ups, and electrodes used. Clinical success was defined as the disappearance of clinical symp-toms for functional (F-) PanNETs and as complete ablation per nonfunctional (NF)-PanNETs. The pooled data were analyzed by a random-effects model. Results: Nineteen studies were selected, including 183 patients (82 males, 44.8%) with 196 lesions (101 F-PanNETs and 95 NF-PanNETs). Pooled estimates for the overall AE rates for the clinical efficacy were 17.8% (95% CI 9.1-26.4%) and 95.1% (95% CI 91.2-98.9%) for F-PanNETs and 24.6% (95% CI 7.4-41.8%) and 93.4% (95% CI 88.4-98.4%) for NF-PanNETs. Conclusions: EUS-RFA appears to be a mini-invasive technique with a good safety and efficacy profile for the treatment of F- and NF-PanNETs. EUS-RFA could be of-fered as possible alternative to surgery for the treatment of low-grade NF- or F-PanNETs, especially for those patients that are not eligible or are at high-risk for surgery.

Keywords: EUS-guided ablation; PanNETs; RFA; endoscopic ultrasound; insulinoma; pancreas; pancreatic neuroendocrine tumors; radiofrequency ablation.

Figure 1

Basic principles of RFA (Courtesy of L. Portella).

Figure 2

Study flow-chart.

Figure 3

Forest plots reporting estimates for overall adverse event rates in F- (a) and NF- PanNETs (b). The pooled AE rates were 17.8% (95% CI 9.1–26.4%) and 24.6% (95% CI 7.4–41.8%) in F- and NF- PanNETs, respectively.

Figure 4

Forest plots reporting estimates for clinical efficacy of F- (a) and NF- PanNETs (b). The pooled efficacy rates were 95.1% (95% CI 91.2–98.9%) and 93.4% (95% CI 88.4–98.4%) in F- (a) and NF- PanNETs (b), respectively.