Endoscopic ultrasound-guided ethanol ablation versus surgical resection of insulinomas

Keywords

insulinoma - neuroendocrine tumor - interventional endoscopic ultrasound - pancreatic surgery - alcohol

Introduction

Pancreatic surgery is the standard of care procedure for patients suffering from insulinoma, a rare neuroendocrine tumor, which in the majority of cases is benign but can lead to life-threatening and repetitive hyperinsulinemia-induced hypoglycemia [1]. Despite improvements in morbidity and mortality rates over the last decades [2], pancreatic surgery is still associated with high mortality risk of up to 3.5% and morbidity rates of up to 35% with complications ≥ 3 or even higher according to the grading system by Clavien-Dindo [3] [4] [5] [6] [7]. Since insulinomas are usually encapsulated, non-metastatic, and of small tumor size, recent improvements of surgical procedures with less radical resection methods such as enucleation have improved outcome [8]. A recent study in patients with insulinomas due to multiple endocrine neoplasia type 1 demonstrated that depending on the type of surgery, postoperative pancreatic fistulas occur in up to 20% of patients [9]. Considering the high rates of surgery-associated complications and the inability to perform surgery in a subgroup of multimorbid patients, the demand for alternative less-invasive procedures is high.

About 16 years ago, the first patient with insulinoma was successfully treated with endoscopic ultrasound (EUS)-guided ethanol injection [10]. Ever since, EUS-guided ethanol ablation of insulinomas has been reported in sporadic case reports showing promising data with regard to safety and efficacy. Armelli et al. provided a case report including a valuable review of the literature on this topic [11]. Thus, accumulating evidence indicates that EUS-guided ethanol injection is an effective, less invasive, and well tolerated procedure for the treatment of insulinomas. However, according to current practice guidelines [12], surgery is recommended for all non-metastatic insulinomas with curative intention. This underscores the need for larger study cohorts and comparison with state-of-the-art pancreas surgery.

We here provide a retrospective analysis of nine patients treated with EUS-guided ethanol injection compared with 24 insulinoma patients undergoing pancreatic surgery.

Patients and methods

Retrospective data were derived from a single-center cohort consisting of all consecutive patients with insulinoma admitted and treated between April 2010 and December 2021 at our institution. Diagnosis was established by the Department for Endocrinology or transferring centers. All tumors were confirmed by EUS at the endoscopy unit of the Department for Gastroenterology as an inclusion criterion for this study.

All patients were diagnosed with insulinoma due to the occurrence of symptomatic hypoglycemia (glucose < 2.22 mmol/l) and concomitant hyperinsulinemia during a 72-h fasting test. Diagnostic visualization of the insulinoma by a single session EUS was successful in all patients. Apart from EUS, the size, location, and appearance were recorded by magnetic resonance imaging with no signs indicating malignancy or metastases. Additionally, hypervascularization was documented by color Doppler mode and if needed and available in the second half of the observation period, by contrast-enhanced EUS (CE-EUS). In the case of atypical EUS morphology, suspicion of local lymph node metastasis or inconsistent clinical presentation, the diagnosis of a neuroendocrine tumor (NET) was confirmed by EUS fine needle aspiration before any therapeutic procedure. In three of four patients with ethanol ablation therapy, EUS FNA confirmed a neuroendocrine tumor. In one patient adjacent lymph node puncture was negative for metastatic disease. Patients who were judged amenable to surgery were advised to have surgical resection by the multidisciplinary NET board.

Surgical resection methods included enucleation, distal pancreatic resection, and pylorus preserving pancreaticoduodenectomy (PPPD), depending on the location of the tumor. Data from 33 consecutive patients with insulinoma who were diagnosed or confirmed in our center during the observation period were collected. Out of these, 9 patients either were judged to have an intolerably increased surgical risk or refused the recommended operation. These 9 patients were treated endoscopically.

First step of endoscopic tumor therapy was EUS-guided puncture of the tumor with a 22G EUS needle, which was prefilled with ethanol 95% after removal of the stylet. The procedure was performed transgastrically in case of corpus and tail insulinoma while the pancreatic head was addressed from the duodenum. Special attention was paid to limit the number of needle penetrations through the tumor capsule to a minimum for its protection. Alcohol injection was performed stepwise in 0.1 ml steps beginning in distant parts of the tumor with stepwise retracting of the needle to the proximal tumor parts. The injection led to white bloating of the tumor directly surrounding the needle tip, partially spread by shortly visible vessels and with loss of effect in the periphery when observed after injection with the needle still in unmodified position. The needle was repositioned several times at different angles until the whole tumor was highly echogenic. Injection was immediately stopped if whitening occurred outside the tumor as a sign of leakage. EUS-guided ethanol injections in all 9 insulinomas are shown in [Fig. 1]. Additionally, for visualization purposes, a video of the injection procedure is provided in [Video 1].

Fig. 1 EUS-guided ethanol injection of all injected insulinomas. White arrows indicate the tumor and yellow arrows show the ethanol within the tumor during injection. Magnification may vary due to illustration purposes.

Demonstration of the alcohol injection procedure into an insulinoma.Video 1

Finally, the needle was removed, and successful termination of perfusion was documented by color Doppler mode and, if available, contrast injection. Intravenous glucose therapy and oral medication were stopped immediately, and the patients’ glucose levels were monitored closely during and for several hours after the procedure. Lipase activity was monitored before the intervention and on the following day for biochemical monitoring of postprocedural pancreatitis. Clinical symptoms were recorded continuously.

To assess the clinical risk for pancreatic surgery, we calculated the ASA and the PREPARE score. The PREPARE (Preoperative Pancreatic Resection) score was calculated using systolic blood pressure, heart rate, hemoglobin level, albumin level, ASA (American Society of Anesthesiologists) score, surgical procedure, elective surgery or not, and disease of pancreatic origin or not, as has been previously described [13]. The ASA score is a common score to assess the physical condition of patients with regard to anesthesia risk [14].

Follow-up was ensured by documented data in internal hospital records or telephone contact with the patients. The follow-up period varied greatly in terms of time interval and data availability as shown in the results section. Distribution of numeric variables was tested using the Shapiro-Wilk test. Normally distributed variables were tested with a T-test for group comparison, and Wilcoxon Exact tests were used for skewed variables. Complication occurrence rates were compared with a Chi-squared test. Correlations were tested using Spearman’s rank correlation coefficient. Statistical analyses were conducted using R 4.0.3 (GUI 1.73 Catalina build). This retrospective study was approved on April 15, 2021 under registration number by the local ethics committee.

Results

An overview of the study cohort characteristics is given in [Table 1]. Overall, the majority of patients were female and most of the patients underwent pancreatic surgery. Surgical procedures included enucleation (n=13, two in combination with Roux-en-Y-anastomosis and pancreaticojejunostomy), distal pancreatectomy (n=9), and pancreatic head resection (n=2). Details on the decision for endoscopic treatment are given in [Table 2]. Patients undergoing EUS-guided ethanol injection were older and tended to have a higher body mass index (BMI), a higher clinical risk for pancreatic surgery (PREPARE score), and a significantly lower physical condition with regard to anesthesia risk (higher ASA score). The tumor location in the ethanol injection group was predominantly the body and the tail, while tumors in the surgery group were more evenly distributed. Tumor diameter, insulin-glucose ratio, and chromogranin A levels were not different between groups.

Table 1 Study cohort characteristics.

Parameter	Ethanol	Surgery	P-value
*The tumor location of two patients was between the body and tail (allocated to body). † The tumor location of one patient was between the body and tail (allocated to body). One patient had 3 distinct tumors (insulinomatosis), one in the body and two in the neck. P-values were calculated with the non-parametric non-paired Wilcoxon Exact Test. Data are presented as mean±standard error or as medians (PREPARE score, ASA, Charlson comorbidity score) with 25th and 75th percentiles.
N	9	24	/
Sex (m/f)	1/8	11/13	/
Age (years)	67.8 ± 11.2	52.3 ± 15.7	0.014
Weight (kg)	109.4 ± 63.6	81.9±18.0	0.856
Body mass index (kg/m2)	34.5 ± 13.6	27.1 ± 5.5	0.059
PREPARE score	2.0 (1.0; 2.0)	0.0 (0.0; 2.0)	0.096
ASA	2.5 (2.0; 3.0)	2.0 (1.0; 2.0)	0.008
Charlson comorbidity index	3.0 (1.0; 4.0)	1.0 (0.0; 2.0)	0.008
Chromogranin A (ng/ml)	399 ± 690	79 ± 59	0.333
Insulin-glucose ratio	0.6 ± 0.3	0.5 ± 0.2	0.178
Tumor diameter (mm)	13.9 ± 7.3	12.4 ± 4.2	0.984
Tumor location (head/body/tail)	1/5/3†	8/6/7*	/

Table 2 Reasons for refusal of surgical resection.

Patient	(Medical) reasons to refuse surgery
I06	Age >70 years, COPD (chronic obstructive pulmonary disease)
I07	Age >70 years, history of stroke (epilepsy, immobility)
I12	Age >70 years, hypopharyngeal cancer, history of stroke, atrial fibrillation, chronic kidney insufficiency
I19	Age >80 years, atrial fibrillation
I22	Risk of complete pancreatectomy after left-sided pancreas resection 3 years earlier
I30	Morbid obesity (BMI 66.5 (kg/m2))
I31	Insulinomatosis with multiple tumors, history of insulinoma enucleation in the past
I32	Age>70 years, cardio-pulmonary comorbidities (COPD, atrial fibrillation) with limited resilience
I33	Cardio-pulmonary comorbidities (COPD, atrial fibrillation)

Due to tumor recurrence, 2 patients had to undergo ethanol injection twice. Pancreatic fistulas, wound healing disorders, and anastomosis insufficiency were the main complications in the surgery group and one patient with previous pancreatic resection several years ago died from postoperative complications ([Table 3]). As shown in [Table 4] and Table S1, one patient in the ethanol group developed mild pancreatitis as evidenced by transiently elevated lipase and mild abdominal pain for 12 h (Grade 1 according to AGREE classification) [15]. In patients I07 and I12, ethanol injection had to be repeated after approx. 2 years and approx. 1 month, respectively, due to recurring hyperinsulinemic hypoglycemia without indication for a second tumor location. No complications occurred during the second intervention and hospitalization time was 2 and 4 days, respectively. Overall, upon last follow-up contact, patients I12 and I19 in the ethanol group died 24 and 892 days after the respective interventions due to multiple comorbidities unrelated to the insulinoma. In the surgery group, all patients were still alive and asymptomatic upon last follow-up contact, except patient I23, who died due to complications of the surgery ([Table 3]). This patient had been treated with left-sided pancreatic resection and enucleation since the initial diagnosis several years earlier. Due to the proximity of the recurrent insulinoma to the aorta and poor endosonographic image quality related to severe steatosis, an endosonographic approach was declined.

Table 3 Complications of patients undergoing pancreatic surgery.

Pat ID	Complications	Assessment	Outcome
I04	Pancreatic fistula	Conservative by intraoperative drainage	Lost to follow-up
I05	Anastomosis insufficiency,	Re-OP with distal pancreatic resection and intraoperative drainage	Asymptomatic
	Peritonitis and shock
	Pancreatic fistula	ERP with stenting
I08	Pancreatic fistula	CT-guided drainage	Lost to follow-up
I10	Pancreatic fistula	Conservative Local cleavage	Asymptomatic
	Abscess at OP scar
	Wound healing disorder
I11	Small pancreas tail necrosis	Conservative	Asymptomatic
I13	Local seroma	Conservative	Asymptomatic
I14	Pancreatic fistula	Conservative by intraoperative drainage	Asymptomatic
I15	Pancreatitis	Conservative	Asymptomatic
I17	Pancreatic fistula	EUS-guided drainage	Asymptomatic
	Necrosis of the spleen with acute abdomen und erosive splenic artery bleeding	Splenectomy and retroperitoneal necrosectomy
	Wound healing disorder	Conservative
I18	Pancreatic fistula	Re-laparotomy, lavage, drainage, subcutaneous VAC therapy	Asymptomatic
	Inguinal abscess	Local cleavage
	Incisional hernia	Laparoscopic repair in IPOM-technique
I20	Pancreatic fistula	Relaparotomy, drainage, pancreatic sphincterotomy, EUS-guided pigtail drainage	Asymptomatic
	Wound healing disorder, local skin burst	Secondary suture
I21	Pancreatic fistula with necrosis	CT-guided drainage and EUS-guided drainage	Asymptomatic
I23	Insufficiency of biliodigestive anastomosis with common bile duct leakage and biliary peritonitis	BDA renewal	Death
	Secondary insufficiency of biliodigestive anastomosis
	Pneumonia	Antibiotic treatment, mechanical ventilation, reanimation

Table 4 Complications and need for repetitive intervention of patients undergoing ethanol injection.

Pat ID	Complications	Assessment	Outcome	Comment
*FT: fasting test
I07	Relapse after 688 days	Re-injection	Asymptomatic	/
I12	Relapse after 160 days	Re-injection	Asymptomatic	Patient died 24 days after second intervention, due to severe comorbidities unrelated to the tumor
I31	Mild pancreatitis with abdominal pain for 12 h. Max. lipase 1141 U/l	Conservative	Asymptomatic but lowest glucose in FT* 2.2 mmol/l	First tumor was operatively enucleated 5 years earlier and pathology revealed insulinomatosis. 2 of 3 recurring tumors were injected with ethanol. Injection of the remaining tumor is planned.

The hospitalization time was about 4 times as long in the surgery group compared to the ethanol group ([Table 5]). Additionally, complication rates differed between groups (surgery: 54% vs. ethanol: 11 %, p=0.026), and complications in the surgery group were more severe compared to the ethanol group (see [Table 3] and [Table 4]), also shown in the Clavien-Dindo classification [16] (see Table S1). The Comprehensive Complication Index was numerically higher in the surgery group but was statistically not different between groups ([Table 5]).

Table 5 Outcome parameters.

Parameter	Ethanol	Surgery	P-value
P-values are derived from T-tests when variables where normally distributed and from Wilcoxon Exact tests when distribution was skewed. Distribution of data was tested using the Shapiro Wilk test. Complications were compared between groups using the Chi-squared test. Data are presented as mean ± standard error or medians and 25th and 75th percentiles (hospitalization, Clavien-Dindo classification, Comprehensive Complication Index)
Follow-up (days)	500 ± 563	810 ± 871	0.289
Lowest glucose before intervention (mmol/l)	2.09 ± 0.17	1.81 ± 0.08	0.158
Lowest glucose after intervention (mmol/l)	4.95 ± 0.74	5.41 ± 0.28	0.581
Complications (%)	11.1	54.2	0.026
Hospitalization (days)	4.0 (2.0; 5.0)	13.5 (7.0; 24.3)	<0.001
Clavien-Dindo classification	0.0 (0.0; 0.0)	1.0 (0.0; 3.0)	0.076
AGREE classification	0.0 (0.0; 1.0)	n.a.	n.a.
Comprehensive Complication Index	0.0 (0.0; 0.0)	8.7 (0.0; 26.6)	0.117

The lowest glucose concentrations were similar in the surgery group compared to the ethanol group before and also after the respective treatments ([Table 4]), indicating similar treatment efficacy of ethanol injection and surgery in the post-intervention period. In both groups, the lowest measured glucose concentrations increased significantly after the treatment (surgery: p<0.001: ethanol: p=0.02, [Table 5]).

Finally, we assessed the injected ethanol volume with regard to the tumor volume. [Fig. 2]a shows a positive correlation of the injected ethanol volume and tumor size (p=0.008). Additionally, the injection volume exceeded the tumor volume by magnitudes and increased drastically with smaller tumor size ([Fig. 2]b).

Fig. 2 Correlations of tumor volume and injected ethanol volume. Panel a shows the correlation between tumor volume and total injected ethanol volume. Panel b shows the relation between tumor volume and ethanol volume per tumor volume, indicating that relative ethanol injection volume increases drastically with lower tumor volume. Significance was tested using Spearman’s rank correlation coefficient due to the skewed data distribution.

Discussion

In this retrospective study, clinical outcomes and group characteristics are compared between insulinoma patients undergoing pancreatic surgery or ethanol injection at our institution between 2010 and 2021. Despite the ethanol group being older, numerically more obese, and having more comorbidities, complications in the surgery group occurred more frequently, were more severe, and hospitalization time was significantly longer. Post-treatment glucose concentrations increased in both groups significantly and were comparable between groups. The low complication rate was confirmed by a study by So, who treated 97 patients with neuroendocrine tumors and compared them with 188 surgically resected patients. They found significantly lower major complications in the endoscopic group [17]. The treatment of small and asymptomatic neuroendocrine tumors raises the question of overtreatment in patients, who would have been subjected to a wait and see strategy elsewhere [18] [19] _. Underlining the advantage of less invasive treatment, a recent multicenter study comparing 89 patients with insulinomas treated by EUS-guided radiofrequency ablation with a surgical cohort found EUS-RFA to be safer than surgery and highly effective [20].

Limitations of our study are the small cohort size, the retrospective nature, and the differences in baseline characteristics such as age, adiposity, and comorbidity scores. However, considering the rarity of the disease, this cohort includes – to the best of our knowledge – the largest group of insulinoma patients undergoing local ablation by EUS-guided ethanol injection reported so far and allows comparison with surgical treatment during the observation period. The data indicate that ethanol injection is an overall safe and effective procedure for the treatment of insulinomas with a low risk of malignancy. Surgery showed comparable efficacy but led to more severe complications, such as pancreatic fistulas which are common surgical complications and are reported to occur at high rates after pancreatic surgery in patients with insulinomas [20] [21] [22]. Additionally, mortality rates following pancreatic surgery in Germany still range from 6.5% in high-volume centers to 11.5% in low-volume centers [23], while a cohort study from the US reported a mortality rate of 2.7% [24]. Also, in our cohort one patient with reoperation died from surgical complications but would not have been eligible for ethanol treatment due to the proximity to the aorta and poor endosonographic image quality. In contrast, ethanol injection or other minimally invasive procedures to target insulinomas may harbor a much lower complication risk [25]. In this study, no patient in the ethanol group had a post-interventional complication higher than Clavien-Dindo I (if re-interventions are excluded). Interestingly, one patient of this cohort who had previously undergone tumor enucleation presented with multiple tumors and was diagnosed via histology with insulinomatosis (MEN was excluded). Of the three tumors, two were ablated with ethanol and the patient became asymptomatic. In a following fasting test, glucose values dropped to 2.22 mmol/l and ablation of the third tumor will take place in the near future. This is the first case showing that ethanol injection can be effective also in multiple tumor manifestations. However, specifically when multiple tumors are present, pancreatic surgery has the advantage that potentially malignant tumors can be resected including peripancreatic lymphadenectomy with curative intent. Nonetheless, this may even lead to total pancreatectomy with insulinoma-dependent diabetes. Malignant tumors are very rare but should be excluded by histology and staging.

Given the high effectiveness of ethanol injection in hypoglycemia resolution (even if multiple injections become necessary) and the less severe complications, ethanol injection appears to represent a valuable alternative in selected patients [25]. Even though two of nine patients had to be injected twice, both were free of symptoms after the second injection. One of them was already diagnosed with a residual tumor directly after the procedure by contrast-enhanced percutaneous sonography but refused another ethanol instillation due to a lack of symptoms at that time. This tumor was injected with ethanol 22 months later because of recurrent symptoms. This example demonstrates the effectiveness of repeated interventions for symptom control.

In contrast to radiofrequency ablation, where cost-intensive equipment is mandatory, the needed technical equipment consists of a 22G needle and ethanol as already established in centers performing EUS. Furthermore, the potential effects on local biliary and vascular structures seem of less concern due to the protection of the tumor capsule. Importantly, our data show that in the case of injection guided by white bloating a multitude of alcohol volume compared to tumor volume was used for most patients. Additionally, insulinomas have a solid structure which theoretically limits injectable fluid amounts. Since injection was immediately stopped in the case of a loss of alcohol alongside the needle, this discrepancy is indicative of high vascularization, spread, and loss via vessels. The idea of loss via vessels is supported by loss of white bloating in the periphery over time when watching the effects of alcohol injection with the needle in place. This hypothesis could explain the restriction of the bloating effect to the tumor and may be a potentially relevant difference to ablation by RFA, where the ablation effect is mostly related to the distance of the probe. However, there is no data on direct comparison of these methods and RFA has been shown to be effective in neuroendocrine and other tumors [20] [26].

The growing evidence on the safety and efficacy of ethanol injection for insulinoma ablation and this report showing treatment outcomes in comparison to pancreatic surgery support the need for long-term prospective randomized trials. Particularly the post-procedural quality of life compared to surgery should be focused on. In addition, considering long-term effects of pancreatic surgery-associated comorbidities, patients that are diagnosed with insulinoma at a young age may also be evaluated for EUS-guided ethanol ablation.

In conclusion, ethanol injection is similarly effective with regard to resolving insulinoma-associated hypoglycemia compared with pancreatic surgery but is linked to less severe comorbidities. Nonetheless, patients need to be evaluated by an interdisciplinary tumor board with experienced pancreatic surgeons, endocrinologists, and endoscopists. Due to the low number of individuals treated with ethanol injection, the data are of exploratory nature and may not be generalizable to all insulinomas. Whether less invasive procedures such as EUS-guided ethanol injection can replace pancreatic surgery in insulinoma patients with a low risk of malignancy, remains to be addressed by prospective randomized trials.

• References

• 1 Service FJ, McMahon MM, O’Brien PC. et al. Functioning insulinoma--incidence, recurrence, and long-term survival of patients: a 60-year study. Mayo Clin Proc 1991; 66: 711-719
• 2 Stojadinovic A, Brooks A, Hoos A. et al. An evidence-based approach to the surgical management of resectable pancreatic adenocarcinoma. J Am Col Surg 2003; 196: 954-964
• 3 Büchler MW, Wagner M, Schmied BM. et al. Changes in morbidity after pancreatic resection: toward the end of completion pancreatectomy. Arch Surg 2003; 138: 1310-1314
• 4 Strasberg SM, Drebin JA, Soper NJ. Evolution and current status of the Whipple procedure: an update for gastroenterologists. Gastroenterology 1997; 113: 983-994
• 5 Herrera-Cabezón FJ, Sánchez-Acedo P, Zazpe-Ripa C. et al. Quality standards in 480 pancreatic resections: a prospective observational study. Rev Esp Enferm Dig 2015; 107: 143-151
• 6 Jilesen APJ, van Eijck CH, Busch OR. et al. Postoperative Outcomes of Enucleation and Standard Resections in Patients with a Pancreatic Neuroendocrine Tumor. World J Surg 2016; 40: 715-728
• 7 Clavien PA, Barkun J, de Oliveira ML. et al. The Clavien-Dindo classification of surgical complications: five-year experience. Ann Surg 2009; 250: 187-196
• 8 Cauley CE, Pitt HA, Ziegler KM. et al. Pancreatic enucleation: improved outcomes compared to resection. J Gastrointest Surg 2012; 16: 1347-1353
• 9 van Beek DJ, Nell S, Verkooijen HM. et al. Surgery for multiple endocrine neoplasia type 1-related insulinoma: long-term outcomes in a large international cohort. Br J Surg 2020; 107: 1489-1499
• 10 Jürgensen C, Schuppan D, Neser F. et al. EUS-guided alcohol ablation of an insulinoma. Gastrointest Endosc 2006; 63: 1059-1062
• 11 Armellini E, Crinò SF, Ballarè M. et al. Endoscopic ultrasound-guided ethanol ablation of pancreatic neuroendocrine tumours: A case study and literature review. World J Gastrointest Endosc 2016; 8: 192-197
• 12 Falconi M, Eriksson B, Kaltsas G. et al. Consensus guidelines update for the management of functional p-NETs (F-p-NETs) and non-functional p-NETs (NF-p-NETs). Neuroendocrinology 2016; 103: 153-171
• 13 Uzunoglu FG, Reeh M, Vettorazzi E. et al. Preoperative Pancreatic Resection (PREPARE) score: a prospective multicenter-based morbidity risk score. Ann Surg 2014; 260: 857-863
• 14 American Society of Anesthesiologists. New classification of physical status. Anaesthesiology 1963; 24: 111
• 15 Nass KJ, Zwager LW, van der Vlugt M. et al. Novel classification for adverse events in GI endoscopy: the AGREE classification. Gastrointest Endosc 2021; 95: 1078-1085
• 16 Dindo D, Demartines N, Clavien PA. Classification of surgical complications: a new proposal with evaluation in a cohort of 6336 patients and results of a survey. Ann Surg 2004; 240: 205-213
• 17 So H, Ko SW, Shin SH. et al. Comparison of EUS-guided ablation and surgical resection for nonfunctioning small pancreatic neuroendocrine tumors: a propensity score-matching study. Gastrointest Endosc 2023; 97: 741-751
• 18 Larghi A, Rizzatti G, Rimbaş M. et al. EUS-guided radiofrequency ablation as an alternative to surgery for pancreatic neuroendocrine neoplasms: Who should we treat?. Endosc Ultrasound 2019; 8: 220-226
• 19 España-Gómez MN, Velázquez-Fernández D, Bezaury P. et al. Pancreatic insulinoma: a surgical experience. World J Surg 2009; 33: 1966-1970
• 20 Crinò SC, Napoleon B, Facciorusso A. et al. Endoscopic ultrasound-guided radiofrequency ablation versus surgical resection for treatment of pancreatic insulinoma. Clin Gastroenterol Hepatol 2023; 2834-2843.e2
• 21 Wei J, Liu X, Wu J. et al. Diagnosis and surgical management of insulinomas in 33 consecutive patients at a single institution. Langenbecks Arch Surg 2009; 401: 1019-1025
• 22 Xu Q, Xie Q, Ge C. et al. Risk factors and prevention of postoperative pancreatic fistula after insulinoma enucleation: a retrospective study from a high-volume center. Pancreatology 2021; 21: 1208-1215
• 23 Krautz C, Nimptsch U, Weber GF. et al. Effect of Hospital Volume on In-hospital Morbidity and Mortality Following Pancreatic Surgery in Germany. Ann Surg 2018; 267: 411-417
• 24 Kneuertz PJ, Pitt HA, Bilimoria KY. et al. Risk of morbidity and mortality following hepato-pancreato-biliary surgery. Gastrointest Surg 2012; 16: 1727
• 25 El Sayed G, Frim L, Franklin J. et al. Endoscopic ultrasound-guided ethanol and radiofrequency ablation of pancreatic insulinomas: a systematic literature review. Ther Adv Gastroenterol 2021; 14
• 26 Armellini E, Facciorusso A, Crinò SF. Efficacy and Safety of Endoscopic Ultrasound-Guided Radiofrequency Ablation for Pancreatic Neuroendocrine Tumors: A Systematic Review and Metanalysis. Medicina (Kaunas) 2023; 59: 359