Keywords
Endoscopic ultrasonography - RFA and ablative methods - Pancreas - Intervention EUS
- GI surgery
Introduction
Pancreatic surgery is complex, with a non-negligible risk of complications [1]. Although pancreatic surgery centralization in high-volume centers have helped reduced
morbidity and mortality rates, up to 35% of these patients develop postoperative complications
[1]. Because pancreatic resection offers the only curative option, its indications remain
important. Peripancreatic vessel involvement has prompted vascular resection and reconstruction
techniques, but with increased postoperative morbidity/mortality [2].
Patients with borderline resectable pancreatic cancer (BRPC) and locally advanced
pancreatic cancer (LAPC) (i.e., unresectable non-metastatic pancreatic cancer), initially
undergo chemotherapy ± radiotherapy for attempted downstaging and potential subsequent
radical
surgery. Nevertheless, only 12% of them proceed to surgery, and a R0 resection is
achieved in
70% of these cases [3]. The main impediment is increasing the clean resection margin, particularly for
arteries. Therefore, there is growing interest in complementary minimally invasive
downstaging
therapies.
Endoscopic ultrasound (EUS) enables high precision guidance of interventional therapies,
and it can enhance radiofrequency ablation (RFA) by allowing real-time visualization
for localized, controlled ablation while preserving surrounding structures. RFA causes
coagulative necrosis and fibrotic changes. Its current indications include functional
neuroendocrine tumors (NET) and percutaneous debulking in LAPC of pancreatic body
after failed chemotherapy. Because the use of sucrose in isotonic solutions combined
with RFA was shown to reduce conductivity and increase heating rates, we have tested
the feasibility and safety of adding hydroxyethyl starch (HES) to pancreatic parenchyma
RFA (EUS-sugar-RFA) in an animal model (n = 4), followed by pancreatic biopsies and
resection [4]
[5].
The present study assessed the effects of EUS-sugar-RFA applied to the perivascular
space of the splenic vessels before pancreatectomy, with the objective of causing
targeted necrosis while maintaining vascular integrity.
The primary outcomes were safety of EUS-sugar-RFA applied to the perivascular pancreatic
space, capacity to create a controllable margin of perivascular necrosis to facilitate
R0 resection, best treatment timing before pancreatectomy, and histopathological effect
on pancreas and perivascular space. The secondary outcome was effect visibility in
post-interventional imaging studies.
Methods
The study was approved by the institutional Animal Care and Ethics Committee (reference
#28599–2020121012122760 v2). Seven pigs (sus scrofa domesticus) were included (interventional
group n = 3; control group n = 4, 2x normal pancreas and 2x standard RFA), and managed
according to French regulations, European Community Council directives (2010/63/EU),
and ARRIVE guidelines [6].
Under anesthesia, a blood collection (hemogram, creatinine, amylase, lipase) and a
contrast-enhanced triphasic thoraco-abdominal CT scan were done.
EUS-sugar-RFA
The three porcine pancreatic segments (duodenal lobe [DL], connecting lobe [CL], and
splenic lobe [SL]) and vascular landmarks were identified using a EUS therapeutic
linear
scope (EG38-J10UT, Pentax, Japan; processor Arietta V70, Hitachi, Japan) [7] ([Fig. 1]
a). Two to three target zones (TZs) according to the
individual anatomy were defined adjacent to the splenic artery (SA) and to the portal
vein
(PV) and splenic vein (SV), respectively. The first site was 15 mm distal to the
spleno-porto-mesenteric confluence (SPMC). The second and third TZs were determined
in a
caudal direction, leaving a 10- to 15-mm distance between them ([Fig. 1]
b and [Fig. 1]
c).
Fig. 1
a Porcine pancreas specimen with the duodenal (DL), connecting (CL) and splenic lobe
(SL) as well as the portal vein (PV) section. b Latero-anterior vision of 3D CT reconstruction showing the planned EUS-sugar-RFA ablation
strategy. Orange transparent: Porcine pancreas; red: arterial system, blue: venous
system; green points: Sites of injection that were both planned and performed; red
point: Site injection planned, not performed after considering the individual anatomy
of the pig. c Schematic design of the hypothesized EUS-sugar-RFA’s effect. Superior image: Pancreatic
tumor compromising the adjacent vessel. Purple arrows: EUS-sugar-RFA application in
the perivascular space. Inferior image: Necrotic effect with vascular wall preservation,
allowing vessel-preserving dissection and tumor resection.
A 22G needle (Expect Slimline, Boston Scientific Corporation, United States; SonoTip
ProFlex, MediGlobe GmbH, Germany) was used to inject 1 to 1.5 cc of HES 130/0.4. Then,
50 watts were applied through a 19G EUSRA needle (Taewoong Medical, United States)
placed in the TZs for 6 seconds (VIVA COMBO RF Generator System). Color Doppler was
routinely used to determine the TZ and to check for bleeding. Finally, an adapted
GAPS-EUS assessment tool was completed [8].
Pancreatectomy
Under anesthesia, the blood sample was repeated. A thoraco-abdominal computed tomography
(CT) scan and a diagnostic EUS were done to document changes.
Open pancreatectomy was performed en bloc with vascular axes, using a vessel-sealing
device (Ligasure, Covidien, Ireland). Then, an Objective Structured Assessment of
Technical Skills (OSATS) and a questionnaire created by us evaluating the subjective
perception regarding the difficulty of pancreatectomy between interventional and control
groups were completed (Annex 1) [9].
The control pancreatectomy group included two normal specimens and two after RFA alone
(50 watts applied in the splenic vessel perivascular space, 10 to 15 seconds), obtained
from educational courses.
Statistics
Due to the pilot character of the study with purposely low sample size and variable
survival period durations, no statistical analysis was performed. Descriptive results
are provided as mean standard deviation.
Results
All procedures were successfully completed. Pigs 1 and 2 had a survival of 4 and 1
days, respectively, between both procedures. They had appetite and tolerated a liquid
diet. No vital sign alterations occurred. On the second follow-up day, pig 1 showed
mild abdominal tenderness during palpation, with a soft abdomen, which resolved within
1 day. No abnormalities were found in the blood samples. Pig 3 underwent a non-survival
protocol (both procedures on the same day).
EUS-sugar-RFA
The first TZ was 15 mm distal to the SMPC, and the following at 10 to 15 mm in the
direction of the SL. With the SA as the landmark, three TZs were defined for pigs
1 & 3, and 2 TZs for pig 2, varying according to the pancreas length; and three TZs
adjacent to the SV. Consequently, five to six injections/RFAs were performed in each
pig. The mean procedure duration was 48.3 ± 10.89 minutes. Classic hyperechogenic
bubbles were observed during RFA. No bleeding was observed under color Doppler control.
Two EUS experts performed the procedures, one with extensive (pig 1) and one with
less experience on animal models (pigs 2 and 3). Completion of EUS-sugar-RFA was represented
by a GAPS-EUS overall score of 71 of 75 for the interventional group ([Fig. 2]).
Fig. 2 GAPS-EUS assessment tool adapted from [8]. Ratings are shown as orange bars representing the absolute numbers for the three
interventional group procedures.
Pancreatectomy
Normal pancreas control
For the 30- and 45-minute procedures: In the longest, metal stent gastrojejunostomies
(EUS-GJ) had been placed during an EUS course, which reduced maneuverability.
Surgical difficulty (Annex 1) was normal complexity for one (score 0), harder exposure
for the EUS-GJ sample (score 3).
RFA alone control
For the 20- and 55-minute procedures: In the longer one, splenic vessels dissection
was more difficult due to previous educational coil + glue treatment, without entailing
complications. Surgical difficulty was normal complexity (score 0), slightly increased
difficulty with the coiled vessel (score 1).
The control OSATS [9] score was 35 of 35 for all but one pig (EUS-GJ: 28/35).
Interventional group
Mean duration of the pancreatectomies = 54 ± 27 minutes. There were no signs of bleeding
or peritonitis during exploration. The pancreas had a regular consistency. There were
no visible signs of inflammation, neovascularization or tissue scarring in pig 3 (acute
study), which allowed the selective use of vessel-sealing device. Pigs 2 and 1 (1
and 4-day survival respectively) had neovascularization around the TZ.
OSATS score: 35/35 (all cases). Surgical difficulty was normal for two pigs (score
0); harder dissection of TZ for pig 1(score 1). The pre-EUS CT scan was normal. On
the presurgical scan, hypodense areas were visible adjacent to the splenic vessels,
corresponding to the TZ ([Fig. 3]). On the post-interventional EUS, a Doppler-negative hypoechogenic zone was visible
adjacent to the splenic vessels, which were slightly larger when compared to the initial
EUS control after EUS-sugar-RFA ([Fig. 3]).
Fig. 3 Treatment sites, visualized in EUS and contrast-enhanced CT before (a,d), during (b),
and after (c,e,f) the EUS-sugar-RFA procedure in pig 1. a EUS assessment prior to the EUS-sugar-RFA with identification of the course of the
splenic vessels’ course, and choice of the target zones. b First part of the EUS-sugar-RFA treatment. Injection in the target zone adjacent to
the splenic vein. c Second step of the EUS-sugar-RFA treatment: After needle retrieval, the RFA probe
is inserted in the target zone and 50 watts are applied for 6 seconds. d Contrast-enhanced CT scan prior to EUS-sugar-RFA: The pancreas is normal. e,f Contrast-enhanced CT scan prior to pancreatectomy (4 days after the EUS-sugar-RFA).
The hypodense area adjacent to the splenic vessels is indicated by the yellow arrows.
P, pancreas; SV, splenic vein; SA, splenic artery; a, aorta; N, needle; SS, sugar
solution; RFA-N, radiofrequency ablation needle; RFA, radiofrequency ablation effect;
HES, hydroxyethyl starch.
Pathology
Normal pancreas specimens
Signs of peripancreatic adiponecrosis and slight coagulation necrosis at the pancreatic
margins, consistent with the use of monopolar cautery and a vessel-sealing device
during dissection. There were minimal foci of acute lymphadenitis. Overall, the pancreatic
parenchyma was homogeneous and served as a reference for comparison with the study
group and RFA controls.
RFA-alone specimens
The vascular coil + glue treatment sample showed a SL hematoma and local peritonitis
consistent with the splenic vessel injury.
Interventional group
Specimen #1 had a 3 × 4 cm yellowish, necrotic zone in the posterior part of the pancreas.
Specimen #2 had a 9 × 3 mm congestive zone in perivascular pancreatic tissue. Specimen
#3 had 1-cm pancreatic haematoma adjacent to treated vessels, and a 1–2 mm necrotic
area ([Fig. 4]). The vessel walls revealed mild mesenterico-portal phlebitis in specimen #1; in
the other pigs, the walls were normal. Wall integrity was maintained in all specimens
([Fig. 5]).
Fig. 4 Histopathological pancreatic findings of the interventional group. Macroscopy: Hematomas
in the perivascular treatment zones (a, pig 3), and necrotic area (yellow arrows) developed in the 4-day period between procedures
1 and 2 (b, pig 1). Microscopy (H&E x 100): Acute pancreatitis with necrosis of pancreatic and
peripancreatic tissues (c, pig 1). Omentum with acute inflammatory reaction (yellow arrow) and adipose tissue
necrosis (blue arrow) adjacent to the pancreas (P) (d, pig 2); early perivascular pancreatic cell necrosis with microscopic foci of perivascular
hematomas (yellow circles) (e, pig 3).
Fig. 5 Microscopic vascular findings in the interventional group (pig 1). Transversal cut
of the portal vein (PV) (a, H&E x 40); mesenterico-portal phlebitis with thickening
of the PV showing polymorphonuclear infiltration and angiogenetic foci (B, H&E x 100).;
and polymorphonuclear infiltration extending up to the tunica intima of the PV (C,
H&E x 200).
Microscopic specimens from pigs 1 and 2 showed acute pancreatitis and peripancreatic
fat necrosis. The mesenterico-portal phlebitis (pig 1) revealed polymorphonuclear
infiltration up to the tunica intima and foci of neoangiogenesis. Pig 2 had signs
of acute peritonitis and Pig 3 presented a perivascular hematoma as well as early
perivascular pancreatic and fat cell necrosis ([Fig. 4] and [Fig. 5]).
Detailed histopathological results are shown in [Table 1].
Table 1 Histopathological findings from the study and control groups.
|
|
Specimen
|
Procedure done
|
Macroscopic description
|
Microscopic evaluation
|
|
D0, day of procedure 1; D1, day after procedure 1; D4, 4 days after procedure 1; EUS-GJ,
endoscopic ultrasound-guided gastrojejunostomy; RFA, radiofrequency ablation.
|
|
Control group
|
Non-treated pancreas 1
|
No particularities
|
25 × 3 × 4 cm specimen, no visible lesion. 1.5-cm lymph node
|
Slight pancreatic coagulation necrosis (specimen margins), rest normal
|
|
Peri-pancreatic adiponecrosis
|
|
Minimal acute lymphadenitis
|
|
Non-treated pancreas 2
|
Modified anatomy (lumen-apposing metal stents with electrocautery enhanced system
for EUS-GJ)
|
15 × 4 × 3 cm specimen, with 1,5 cm² pancreatic gray/white lesion + focal congestion.
1.7-cm lymph node
|
Slight pancreatic coagulation necrosis (specimen margins) + focus of isolated coagulative
necrosis (consequence of EUS-GJ)
|
|
Peri-pancreatic adiponecrosis
|
|
Minimal acute lymphadenitis
|
|
RFA 1
|
RFA and pancreatectomy (D0); Modified anatomy (artificial attached fluid-filled collections,
and a coil + glue treatment in the splenic vessel)
|
19 × 4 × 3 cm specimen
2 × 1 cm gray pancreatic lesion
6 mm hematoma in the tail Coil near splenic vessel
|
|
|
|
|
|
|
|
|
|
|
RFA 2
|
RFA and pancreatectomy (D0) Modified anatomy (artificial attached fluid-filled collections)
|
13 × 5,5 × 3 cm specimen, no visible lesion
|
|
|
Acute lymphadenitis
|
|
Acute peritonitis
|
|
Interventional group
|
EUS-sugar-RFA 1
|
EUS-sugar-RFA (D0); Pancreatectomy (D4)
|
15 × 5 × 4 cm specimen.
3 × 4 cm yellowish necrotic zone
|
Pancreatic & peri-pancreatic coagulation necrosis, nerval and fat tissue necrosis
|
|
Pancreatitis and peripancreatitis
|
|
Focal mesenterico-portal phlebitis
|
|
EUS-sugar-RFA 2
|
EUS-sugar-RFA (D0); Pancreatectomy (D1)
|
Congestive perivascular pancreatic tissue. Gastric submucosal hematoma
|
|
|
Pancreatic and peri-pancreatic adiponecrosis
|
|
Acute pancreatitis
|
|
Acute peritonitis
|
|
EUS-sugar-RFA 3
|
EUS-sugar-RFA + pancreatectomy (D0)
|
17 × 5 × 4 cm specimen
1 cm hematoma around treated vessels, with no other pancreatic lesion
|
Focal early perivascular pancreatic necrosis, minimal adiponecrosis
|
|
Hematoma
|
Discussion
This study assessed the impact of adding a starch solution to RFA (EUS-sugar-RFA)
applied to the perivascular space of the splenic vessels before pancreatectomy. After
proving feasibility and safety along a 4-day survival period (pig 1), the following
survival periods were shortened to minimize local inflammatory response. The best
timeframe for EUS-sugar-RFA was within 24 hours prior to pancreatectomy, where the
reduced inflammatory response and neovascularization limited the use of a vessel-sealing
device during dissection. Vascular wall integrity was maintained for all specimens.
Our previous study suggested that the interaction between starch and RFA generated
a demarcated necrosis that allowed a clear separation of necrotic from normal tissue,
but that study was performed within the parenchyma [5]. The present study focused on assessing the effect when applied to the perivascular
space, specifically on the vascular axes adjacent to the pancreas, for potential application
in pancreatic cancer with vascular compromise.
Because the present study targeted the perivascular space, in contrast to pancreatic
parenchyma or neoplasia, the energy was applied for a shorter time to avoid potential
vascular complications. The result was a perivascular 5-mm charred layer composed
by fibrin and granulation tissue.
The interaction between sugar solutions and RFA has not been extensively explored
in in vivo, but an ex vivo study using a porcine vascular model concluded that the
addition of carbohydrates to a solution enabled a selective higher cell death rate
and lower conductivity when exposing a tissue to RF energy [4]. Therefore, we have subsequently assessed sugar-boosted RFA, targeting a specific
zone in which a circumscribed, augmented effect is desired.
Perivascular space injection may benefit from the local fluid spread adjacent to the
initial injection site, dissecting the space, and thereby supporting energy transmission
to the perivascular space. Moreover, the sugar/RF interaction allows an augmented
ablation while remaining limited to the TZ.
As observed in our previous study, the addition of starch allowed the delivery of
a lower amount of energy to achieve the desired effect.
The present study is limited by the absence of pancreatic neoplasia. The proof of
concept, therefore, was achieved without assessing its capacity to downstage pancreatic
tumors. Survival after pancreatectomy was omitted due to the expected complex management
of insufficiencies in accordance with ethical considerations.
A detailed assessment of the vascular area is essential prior to treatment and avoiding
areas close to the pancreatic ducts. The maintenance of a stable position during HES
injection, needle retrieval, and RFA catheter insertion is also fundamental. Such
technical precision requires a high level of EUS expertise and the assistance of a
second operator, which is a disadvantage. However, the high overall GAPS-EUS score
obtained by the second operator indicates that the procedure can be quickly learned
by expert endoscopists.
The small number of animals is also a limitation, as it does not allow taking significant
conclusions as to the ideal timeframe of application. Also, the control animals were
taken from educational courses for ethic reasons, but still hinder comparison. However,
feasibility, safety and histopathological findings are consistent along the previous
and present study, with an overall of seven animals treated with EUS-sugar-RFA. A
multicentric study with several experts and higher number of procedures is required
to assess generalisability of the procedure and further biological aspects before
clinical translation.
Conclusions
In conclusion, perivascular EUS-sugar-RFA of the pancreas is an emerging neoadjuvant
supportive technique. Although not conclusive, the best observed time to perform it
was on the day of surgery, because it efficiently induced perivascular necrosis without
complicating inflammation/hemorrhage. Potential applications are preoperative treatment
before distal/partial pancreatectomies for NETs, in selected patients with BRPC and
LAPC, and for treatment of metastases in the body/tail. Larger studies, with follow-up
periods after EUS-sugar-RFA and surgery, are needed to evaluate the impact on the
residual parenchyma. Only then can further clinical protocols be planned. If this
approach is also shown to be safe and feasible in clinical contexts, it may become
part of the multidisciplinary treatment of pancreatic disease in the future.
