Introduction
Pancreatic neuroendocrine tumors (P-NETs) are uncommon, although in recent years, an increased incidence has been reported [1]. While functional P-NETs (i. e. insulinomas, glucagonomas, vipomas, somatostatinomas) produce metabolic disorders and, therefore, are diagnosed early when still small [2], diagnosis of non-functional P-NETs is often delayed because symptoms only occur due to large tumor volume. For this reason, the former are usually treated with surgery and the latter are largely judged as unfit for surgery.
In recent years, enhanced imaging techniques such as endoscopic ultrasound (EUS) have resulted in earlier diagnosis of P-NETs [2]. Once they are discovered, therapeutic decision-making remains quite complicated because surgery for small, non-functional P-NETs may be considered “overtreatment” burdened by high risk in a tumor with a relatively benign natural history [3].
Recently, EUS-guided radiofrequency ablation (RFA) has been described as a new therapeutic approach with significant effectiveness and safety when used for treating P-NETs. Currently, pancreatic EUS-RFA can be performed by using two different devices: the STARmed (Taewoong Medical, Gyeonggi-do, South Korea), an 19 G RFA needle connected to a specific RF monopolar VIVA RFgenerator associated with a pump cooling the needle internally with chilled saline solution; and the Habib EUS RFA (EMcision Ltd, London, UK) device, which is a 1 Fr wire monopolar electrode that can be inserted inside a standard 22 G EUS fine-needle aspiration (FNA) needle, connected to a regular electrosurgical generator (ERBE Medizin VIO 200 or 300 series), and has no cooling system.
To date, only a few case reports and small case series (1 – 12 patients) have described the feasibility, effectiveness and safety of EUS-guided RFA treatment of P-NET [4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]. We report 11 cases of P-NET managed in a single tertiary center, with the aim of exploring the effectiveness and safety of EUS-RFA in treatment of P-NET.
Patients and methods
Study population
The patients described here were treated consecutively in the Department of Gastroenterology and Digestive Endoscopy, Salvini Hospital, Garbagnate Milanese, Milan, Italy, between June 2016 and October 2018 and were followed up for at least 12 months. All patients included had histologically proven P-NET, ≤ 2 cm in size, and either were judged to be inoperable or refused surgery. Indications for endoscopic or surgical management were discussed in a multidisciplinary setting (gastroenterologists, surgeons, oncologists and pathologists). EUS was performed with a linear array echoendoscope (Pentax EG-3870UTK; Pentax, Tokyo, Japan), while tissue acquisition was done through the gastric or duodenal wall depending on the tumor location, with the 19-gauge (G) flex (Expectflex, Boston Scientific Corp., Marlborough, Massachusetts, United States) for core and immunohistochemistry staining. All the subjects underwent total-body contrast-enhanced computed tomography (CT) and Gallium 68-based positron emission tomography (68GA PET) to detect pathological nodes or distant metastases.
As for the endoscopic RFA technique, all procedures were performed within 3 weeks after EUS/CT/68GA PET evaluation by using a linear-array therapeutic echoendoscope under deep sedation. Indomethacin (100-mg suppository) was administered prior to RFA for acute pancreatitis prophylaxis.
The EUS-RFA device consisted of a 19-gauge needle electrode with an active 5 – or 10-mm tip and an internal cooling system (EUSRA system Taewoong, Combo VIVA™ generator STARmed). Under EUS guidance, a monopolar electrode needle with high-frequency alternating current was inserted into the target lesion while avoiding the pancreatic and bile ducts or maintaining a distance of at least 2 mm fron them to avoid damage or duct strictures. Injury to blood vessels was avoided with the help of Doppler evaluation.
The needle tip was positioned at the far end of the lesion. Energy was delivered when the needle tip of the electrode was visualized within the margin of the lesion on EUS and was stopped when impedance quickly increased, reaching at least 500 to 600 Ohms, as per the manufacturer’s protocol.
Using 20 Watt for 10 to 25 sec, a 5 – or 10-mm exposed active tip (according to the size of the tumor) produced an ablation area of about 10 to 25 mm after each application, recognized as a slowly increasing hyperechoic zone within the tumor. If needed, the electrode was repositioned under EUS guidance to ablate another area along the same trajectory.
The completeness of the treatment was assessed endosonographically with the administration of contrast medium Sonovue (Bracco, Milan, Italy) to evaluate for residual tissue to ablate, about 20 minutes after the RFA treatment. In case of active residual tissue, another treatment was performed during the same session.
After the procedure, all patients were fasted for at least 12 hours; analgesic (paracetamole 1 g ev bis in die) and antibiotic therapy (cephalosporin or fluoroquinolone) were administered; massive intravenous hydration (30 – 40 ml/kg/die considering the comorbidities) and blood exam control 6 – 24 – 48 hours after RFA were performed. Finally, abdominal CT scans to control the complete ablation were performed 1 month, 6 and 12 months after procedure. Positive response was defined by absence of enhancing tissue at the tumor site on 6 – and 12-month CT scans.
For patients with functioning insulinomas, endocrinological clinical controls every 4 weeks for the first 3 months and every 8 weeks for the following 9 months were prescribed.
Adverse events (AEs) were recorded in accordance with the following classification [9]: mild AEs: abdominal pain, mild increase of amylases (< 3 days hospital stay post-endoscopy), self-limiting fever; moderate AEs: acute pancreatitis, late onset of hematoma and ulceration in duodenal wall (medical control, no surgical intervention required); severe AEs: pancreatic necrotic lesion requiring surgical intervention by laparoscopy or endoscopic intervention (e. g. endoscopic drainage of intra-abdominal post-procedure collections).
Statistical analysis
Data were analyzed using the Statistical Package for Social Sciences (SPSS software v.15.0, Chicago Illinois, United States) for Windows. The descriptive statistics used included determination of mean values and standard deviation (SD) of the continuous variables, and of percentages and proportions of the categorical variables.
Results
Patients
We prospectively enrolled 10 patients (5 males, mean age 78.6 years, mean BMI 28.2) who had a total of 11 P-NETs with a mean size of 14.5 mm (range 9 – 20 mm) localized in the pancreatic head (3 lesions), pancreatic body (5 lesions), and tail (3 lesions), respectively. The details of enrolled patients are reported in [Table 1].
Table 1
Baseline characteristics and outcomes of the 10 patients who underwent radiofrequency ablation (RFA) of a pancreatic neuroendocrine tumor.
|
Patient number
|
Age (Years)
|
Sex
|
Clinical presentation
|
Diagnosis
|
Location
|
Size of tumor, mm
|
Electrode caliber, G
|
Response
|
Sessions, n
|
Treatment applications per session, n
|
Adverse events
|
|
1
|
71
|
M
|
None
|
Net G1 ki67 < 4
|
Head
|
20
|
19 G 10 mm
|
Yes
|
1
|
2
|
Mild abdominal pain
|
|
2
|
76
|
F
|
None
|
Net G1 ki67 < 1
|
Body
|
17
|
19 G 10 mm
|
Yes
|
1
|
3
|
0
|
|
3
|
72
|
M
|
Hypoglycaemia
|
Net G1 ki67 < 1
|
Tail
|
9
|
19 G 5 mm
|
Yes
|
1
|
2
|
0
|
|
4
|
81
|
M
|
Hypoglycaemia
|
Net G1 ki67 < 1
|
Body
|
11
|
19 G 10 mm
|
Yes
|
1
|
3
|
0
|
|
5
|
78
|
M
|
None
|
Net G1 ki67 < 4
|
Head
|
12
|
19 G 10 mm
|
Yes
|
1
|
2
|
Mild abdominal pain
|
|
6
|
84
|
F
|
None
|
Net G1 ki67 < 4
|
Tail
|
20
|
19 G 10 mm
|
Yes
|
1
|
3
|
0
|
|
7
|
86
|
F
|
Hypoglycaemia
|
Net G1 ki67 < 1
|
Tail
|
15
|
19 G 10 mm
|
Yes
|
1
|
2
|
0
|
|
8
|
80
|
F
|
Hypoglycaemia
|
Net G1 ki67 < 1
|
Body
|
16
|
19 G 10 mm
|
Yes
|
1
|
2
|
0
|
|
9
|
77
|
M
|
None
|
Net G1 ki67 < 1
|
Head + Body
|
15, 12
|
19 G 10 mm
|
Yes
|
1
|
2
|
0
|
|
10
|
81
|
F
|
Hypoglycaemia
|
Net G1 ki67 < 1
|
Body
|
13
|
19 G 10 mm
|
Yes
|
1
|
2
|
0
|
N, number; NET, neuroendocrine tumor; RFA, radiofrequency ablation, M, male; F, female; G, gauge.
Five patients had symptomatic insulinomas (hypoglycemia with elevated fasting insulin and C-peptide levels was documented before EUS-RFA); 3 subjects had non functional intermediate P-NET (Ki 67 < 4 %) and 2 patients had low-grade P-NET (Ki 67 < 1 %). None pathological nodes/distance metastases were detected with total body CT scan and 68GA PET.
Seven patients were judged unfit for surgery for severe comorbidities (2 patients with severe cardiomyopathy and vascular illness and 2 with severe obesity) or advanced age (3 subjects), while the remaining three refused surgery.
Effectiveness
The procedure was technically successful (ability to target the RFA needle) in all patients and tumors. Complete ablation of P-NET was achieved using a single-session RFA with a mean of 2.3 treatment applications per session ([Fig. 1]). In patients with insulinomas, hypoglycemia relief was almost immediate, with normalization of blood sugar levels within 24 hours. These patients remained euglycemic and asymptomatic at all visits through 12 months after insulinoma ablation.
Fig. 1 Endoscopic ultrasound-guided radiofrequency ablation (EUS-RFA) for pancreatic neuroendocrine tumors (P-NETs). a EUS views showing a NET located in the body of the pancreas. b,c RFA with visible white bubbles within the NET.
At 6-month CT scan, all patients had complete disappearance of the lesions with radiological normalization. These findings were also confirmed at 12-month CT scan.
Safety
All RFA sessions ended with no early or late complications related to the procedure and no cases of bleeding were recorded. Only two cases of mild abdominal pain were recorded in two subjects with pancreatic head lesions, which were effectively treated with analgesics ([Table 1]) (slight hyperamylasemia – < 3× ≤ 3× upper limited of normal – without abdominal pain was also observed in 3 subjects). The mean duration of hospital stay was 4 days (range 3 – 7 days).
Discussion
Thanks to enhanced use of imaging techniques, the incidence of pancreatic neuroendocrine tumors has increased in recent years [15]
[16], significantly increasing diagnosis of small, asymptomatic P-NETs [17].
To date, surgery is considered the only effective treatment for P-NET, and while it dramatically increases survival, it is burdened by a poor safety profile, both in the short and long term [18]. About this topic, Jilesen et al [19] in their systematic review of 62 studies, reported significant AEs after pancreatic surgery, which included pancreatic fistula (up to 58 %), delayed gastric emptying (up to 18 %) and postoperative bleeding (up to 6 %), and an overall in-hospital mortality of 6 %. Furthermore, longer hospital stays have been described with both laparoscopic and open pancreatic surgery [20].
On the basis of these well-known limitations of surgery in treating pancreatic neoplasms, non-surgical, safer, effective, and less invasive therapies are warranted [21].
EUS-RFA recently has been described as a new, non-surgical, effective, and safe therapeutic option for patients with P-NETs [4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14].
In the present study, we reported 11 cases of P-NETs prospectively followed up for 12 months and demonstrated complete ablation using a single-session RFA procedure without severe early or late complications.
About effectiveness, at 6-month CT scan, all patients had complete disappearance of lesions with radiological normalization. These outcomes were also confirmed at 12-month CT scan. These results are similar to those of Barthet et al [13], which showed complete disappearance or necrosis of the lesion in 92.3 % of P-NETs at 1-year follow-up evaluation. It is important to emphasize that in functional P-NETs (mostly insulinomas), given their very low malignancy, there is theoretically no need to achieve complete ablation, except when reducing the cell mass is required to decrease insulin production [20].
About safety, only two cases of mild abdominal pain, responsive to analgesic, occurred. Furthermore, slight hyperamylasemia – < 3× ≤ 3× the upper limit of normal – without abdominal pain was also observed in three subjects. The mean duration of hospital stay was 4 days. In patients with insulinomas, hypoglycemia relief was quick, and they remained euglycemic and asymptomatic at the 12-month follow up. Our results are similar to those reported by Barthet et al [13], and better than those of Choi et al [10], probably due to the more stringent inclusion criteria we used. Furthermore, we administered nonsteroidal anti-inflammatory drugs to all our patients in suppositories to prevent acute pancreatitis, although this treatment is not standard. In previous case reports, only one case of necrotizing pancreatitis was recorded [13], while others reported only a few mild AEs, confirming our results.
On the basis of our data, selection of patients for EUS-RFA remains challenging, because it should be based on patient profile (comorbidities, surgical risk, life expectancy), tumor features (location and mass size) and endoscopist experience.
Our study presents some limitations, which deserve a discussion. First, presumptive tumor eradication has been based only on imaging of P-NETs while histology is lacking; this limitation is, however, shared by all studies available in the literature and included in our analysis, a fact that makes imaging-based follow-up the current standard of care for management of P-NETs with EUS-RFA. Second, the follow-up period in our series was no longer than 12 months, but that is comparable to previous studies about this topic. EUS-RFA is a novel technique and long-term data on it thus are lacking. Larger studies with longer follow-up are needed to evaluate the long-term effectiveness of EUS-RFA. Finally, the setting for the radiofrequency generator has not been standardized. We decided to use 20 Watts to treat the lesions, a value that differs from that used in other studies (10 – 50 Watts) [9]
[10]
[13], but which slowly cooks the lesion, likely optimizing effectiveness, controlling extension of the treated area, and minimizing risk of pancreatic tissue damage and AE.
Conclusion
In conclusion, EUS-RFA represents an effective and safe treatment for management of P-NETs and can be considered as a therapeutic option for small P-NETs. Larger series with longer follow-up are needed to better define which patients with P-NETs would benefit from endoscopic therapy and which would be better treated surgically.
