Introduction
Acute pancreatitis (AP) is amongst the most frequent causes of gastrointestinal tract
diseases that requires acute hospitalization and its incidence continues to rise [1 ]
[2 ]. Around 20 % of patients with acute pancreatitis develop necrotizing pancreatitis
with about a third of them progressing to infected necrosis which is associated with
mortality rates reported between 15 % and 30 % [3 ]
[4 ]
[5 ]. Since infected necrosis rarely responds to conservative treatment alone, virtually
always some form of invasive and interventional treatment is necessary.
Over recent decades, the treatment of infected necrotizing pancreatitis has changed
dramatically. Early open surgery is associated with a very high mortality rate and
is largely avoided nowadays [6 ]. A shift towards less invasive techniques has become the standard of care. Minimally
invasive techniques, either by percutaneous drainage, if necessary followed by video-assisted
retroperitoneal drainage or endoscopic ultrasound (EUS)-guided transluminal drainage,
if necessary followed by direct endoscopic necrosectomy (DEN), have been shown to
improve outcomes for patients with regard to a combined endpoint consisting of mortality,
multi-organ failure, external fistula, and endo- and exocrine insufficiency [5 ]
[7 ]
[8 ]. Several studies have reported on the potential and efficacy of direct endoscopic
necrosectomy [9 ]
[10 ].
If signs of infection persevere or worsen after EUS-guided transgastric or transduodenal
drainage, the cyst cavity can be entered by a regular forward viewing endoscope to
perform DEN. This can be achieved by balloon dilation of the transgastric fistula
(up to 20 mm) when plastic double pigtail stents were placed initially, or directly
through the stent opening when a large bore fully covered metal lumen apposing stent
was placed. Usually, several sessions are required for complete removal of the necrosis;
the mean number of DEN sessions varied from 1 to 15 in a meta-analysis by Puli et
al. [11 ] with a weighted mean of 4.09 procedures.
One of the main limitations of endoscopic necrosectomy is the lack of dedicated and
effective instruments to remove the necrotic tissue. For this purpose, various instruments,
originally designed for other indications, are used. These devices, such as lithotripsy
baskets, grasping forceps, retrieval nets, and polypectomy snares, are able to grasp
and hold material but often lack sufficient grip making the procedure cumbersome,
time consuming, and often marginally effective with only small chunks of necrosis
being pulled into the gastrointestinal lumen per pass. Also, opening these devices
in areas of necrosis is largely visually uncontrolled as is the amount of tissue that
is caught. Pure suction can be helpful to pull out tissue chunks but often results
in clogging of the working channel of the endoscope.
The aim of our study was to evaluate the technical feasibility, safety, and clinical
outcome of the EndoRotor, a novel automated mechanical endoscopic resection system
to suck, cut, and remove small pieces of tissue in patients with necrotizing pancreatitis.
Materials and methods
This prospective study took place at the Department of Gastroenterology and Hepatology
of the Erasmus MC, University Medical Center in Rotterdam, a tertiary referral center
in the Netherlands and at the Medizinische Klinik II, Sana Klinikum Offenbach in Offenbach,
Germany. We recorded data on patient demographics, clinical presentation, etiologies
of acute pancreatitis, American Society of Anesthesiologists’ (ASA) Physical Status
Classification score [12 ], radiologically defined size of the necrotic collection in a transverse computed
axial tomographic image displaying the largest diameter of the necrotic cavity, procedural
details, and adverse events during and after endoscopic necrosectomy. All patients
with symptomatic walled-off pancreatic necrosis (WOPN) were considered eligible for
this study.
The EndoRotor
The EndoRotor (Interscope Medical, Inc., Worcester, MA, United States) is a novel
automated mechanical endoscopic resection system designed for use in the gastrointestinal
tract for tissue dissection and resection with a single device. The EndoRotor system
can be advanced through the working channel of a therapeutic endoscope with a working
channel of at least 3.2mm in diameter. The EndoRotor can be used to suck, cut, and
remove small pieces of tissue through the catheter, consisting of a fixed outer cannula
with a hollow inner cannula. A motorized, rotating, cutting tool driven by an electronically
controlled console performs tissue resection and rotates at either 1000 or 1700 revolutions
per minute. The catheter shaft is flexible and can tolerate endoscope bending of manipulation
up to greater than 160 degrees. If greater manipulation is required, a longer catheter
can be used to facilitate less torsional stress on the device. The necrotic tissue
is sucked into the catheter using negative pressure and cut by the rotating blade
from the inner cannula. Tissue is transported to a standard vacuum container. Both
the cutting tool and suction are controlled by the endoscopist using two separate
foot pedals. During this study the EndoRotor catheter was upgraded to potentially
resect necrotic tissue more effectively. All procedures with both versions of the
EndoRotor are reported in this study. The originally designed EndoRotor has a 3.0 mm2 opening at the tip, in which the rotator blade is located, and teeth on the inner
cutter. The novel design is adjusted to facilitate the resection of necrotic tissue;
the tip has a 50 % wider opening of 4.4 mm2 , the teeth on the inner cutter are smaller, and this design additionally has teeth
on the outer cutter. No changes were made in the material or available rotation speed
([Fig. 1 ]).
Fig. 1 The EndoRotor.
Procedure
Procedures were performed as per protocol under conscious or propofol sedation with
close monitoring by the anesthesia team and by six senior endoscopists with a broad
experience in advanced endoscopic procedures. Initially, all patients underwent EUS-guided
transgastric drainage, creating a fistula from the stomach to the adjacent WOPN. The
choice of placement of one or more plastic stents or a lumen apposing metal stent
(LAMS) was at the discretion of the endoscopist. In some patients, a nasocystic irrigation
catheter was placed. In the absence of clinical improvement following initial endoscopic
transgastric drainage, we proceeded to perform endoscopic necrosectomy. For this,
a therapeutic gastroscope was advanced into the collection cavity, if necessary after
balloon dilation to 18 or 20 mm with a controlled radial expansion (CRE) balloon.
The EndoRotor was inserted through the working channel of a therapeutic endoscope
and advanced into the collection cavity. Rotation speed of the EndoRotor catheter
was recorded as well as changes in setting. Suction was set at between 500 and 620 mmHg,
the maximum achievable negative pressure level.
Questionnaire
Endoscopists were asked to rate their experiences with the EndoRotor in a short questionnaire.
Appreciation was expressed on a 10-point Likert scale [13 ], varying from very negative appreciation (1) towards very positive appreciation
(10). Questions were asked about the ease of use of the EndoRotor, handling of the
device, the safety, and their appreciation on the additional value of the EndoRotor
in the treatment of patients with pancreatic necrosis.
Statistics
All statistical analyses were performed using IBM SPSS Statistics, Version 24.0 (IBM
Corp., Armonk, NY, United States). Data were expressed as mean ± standard deviation,
median, and range.
Results
Patient characteristics
Twelve patients with a mean age of 60.6 ± 11.4 years underwent endoscopic necrosectomy
using the EndoRotor. Of these 12 patients, nine were male (75 %). Two patients (16.7 %)
had a class IV score on the ASA physical status classification system. Four patients
(33.3 %) scored ASA class III, five patients (41.7%) scored class II, and one patient
(8.3 %) scored ASA class I. Nine patients were diagnosed with acute necrotizing pancreatitis,
which had developed into infected WOPN. The time from the onset of acute complicated
pancreatitis to necrosectomy was a median of 48 days (range 13 to 368). Three patients
were initially drained because of mechanical complaints from large fluid collections.
As a result of these procedures, the necrotic debris became infected necessitating
necrosectomy. The mean necrotic collection size was 117.5 ± 51.9 mm. The etiology
of the acute pancreatitis was biliary in four patients (33.3 %), alcoholic in three
patients (25 %), endoscopic retrograde cholangiopancreatography (ERCP) with sphincterotomy
in one patient (8.3 %), and in four patients (33.3 %), the etiology was unknown.
Transgastric endoscopic drainage was performed in all patients; eight patients (66.6 %)
received two or three plastic stents, and four patients (33.3 %) received a LAMS to
achieve transluminal drainage of necrotic debris. In all patients, an aspirate of
the WOPN was obtained and sent for Gram staining and culture. Culture-proven infected
necrosis was present in 10 out of 12 patients (83.3 %). The predominant microorganisms
found were Streptococcus species.
Three patients (25 %) were previously treated unsuccessfully with conventional instruments
before being treated with the EndoRotor. Patient characteristics are included in [Table 1 ]. [Fig. 2 ] illustrates a typical case of infected WOPN and the steps during DEN.
Table 1
Patient characteristics.
Patient
Age, years
Sex
Etiology
Infected necrosis proven by culture
Size of the collection, mm
Stent placement
Previous necrosectomy[1 ]
1
56
Female
Biliary
Yes
100
3 Pigtails
2
2
65
Male
Unknown
Yes
167
2 Pigtails
3
3
68
Male
Unknown
Yes
182
LAMS
1
4
43
Male
Biliary
Yes
141
2 Pigtails
0
5
67
Male
Biliary
Yes
130
2 Pigtails
0
6
71
Male
Biliary
Yes
78
LAMS
0
7
76
Female
Alcoholic
Yes
124
2 Pigtails
0
8
58
Male
Iatrogenic
No
220
3 Pigtails
0
9
51
Male
Alcoholic
Yes
84
LAMS
0
10
67
Female
Unknown
Yes
45
LAMS
0
11
66
Male
Unknown
Unknown
100
2 Pigtails
0
12
39
Male
Alcoholic
Yes
90
1 Pigtail
0
LAMS, lumen apposing metal stent.
1 Number of endoscopic necrosectomy procedures previously performed with conventional
instruments.
Fig. 2 Typical case of infected walled-off pancreatic necrosis (WOPN). a Pre-intervention computed tomography scan illustrating WOPN with air bubbles. b Endoscopic view of necrosis after direct access into the cavity. c Necrosectomy using the EndoRotor. d Post-necrosectomy result.
Endoscopic procedure
In the 12 patients described, a total of 27 endoscopic necrosectomy procedures were
performed using the EndoRotor to achieve complete removal of necrotic tissue. The
median procedure time was 38 minutes (IQR 28.9). To achieve complete removal of pancreatic
necrosis, the median number of procedures required was two per patient (range 1 to
7). The first version of the EndoRotor was used in 19 procedures, and the median procedure
time was 45.8 minutes (IQR 28.1), with a median number of procedures required of 2.0
(range 1 to 7). The second version was used in eight procedures with a median procedure
duration of 33 minutes (IQR 31.3); the median number of procedures required was 1.5
(range 1 to 3) to achieve complete removal of necrotic tissue.
Adverse events
No adverse events occurred during the necrosectomy procedures or within the next 24
hours. Three patients (27.2 %) experienced adverse events within the course of their
infected pancreatic necrosis. One patient died eight days after the last endoscopic
necrosectomy as a result of ongoing multi-organ failure caused by massive collections
of infected pancreatic necrosis which, despite multiple sessions, could not be completely
removed. One patient eventually died three months after discharge due to an underlying
pancreatic carcinoma after having undergone two successful endoscopic necrosectomy
procedures for infected necrotizing obstructive pancreatitis using the EndoRotor.
In one patient, a gastrointestinal bleed occurred two days after the procedure necessitating
coiling of the splenic artery. During the procedure, there was no evidence of bleeding
or damage to any exposed vessel.
Questionnaire
Endoscopists rated the EndoRotor easy in its use (mean 10-point Likert scale score,
8.3; range, 8 to 9) and an effective tool to remove necrotic tissue (mean 10-point
Likert scale score, 8.3; range, 8 to 9). They were especially satisfied by the ability
to manage the removal of necrotic tissue in a controlled way (mean 10-point Likert
scale score, 8.6; range, 8 to 9). The risk of causing complications was estimated
to be low (mean 10-point Likert scale score, 1.9; range, 1 to 2). Overall, the device
was judged to be of substantial additional value in the management of pancreatic necrosis
(mean 10-point Likert scale score, 8.6; range, 8 to 9), and respondents were very
willing to use the device in subsequent cases with necrotizing pancreatitis (mean
10-point Likert scale score, 9.3; range, 9 to 10).
Discussion
Direct endoscopic necrosectomy has proven to be safe and effective in the treatment
of patients with infected pancreatic necrosis; however, up until now, no dedicated
instruments were available for treating these patients. Recently, we published our
preliminary experience with the EndoRotor [14 ]. This multicenter prospective cohort study describes the results in the first 12
patients with infected pancreatic necrosis who underwent a combined total of 27 DEN
procedures using the EndoRotor. We have demonstrated the efficacy of the EndoRotor
and good clinical outcomes, without any directly device-related adverse events.
For years, open necrosectomy has been considered as the gold standard treatment for
management of pancreatic necrosis; however, it was accompanied by high morbidity and
mortality rates [6 ]. In 2000, Carter et al. [15 ] demonstrated a new minimally invasive approach indicating that adequate necrosectomy
can be achieved by either percutaneous or endoscopic techniques. These results encouraged
several research groups to further investigate minimally invasive techniques, with
the first randomized controlled trial published in 2010 by Van Santvoort et al. [5 ] confirming the benefit of a minimally invasive approach versus open necrosectomy
in terms of major complications or death [6 ]
[15 ]
[16 ]
[17 ]. The potential and efficacy of endoscopic necrosectomy in terms of overall outcome
is undisputed, but the proper tools available for adequate endoscopic debridement
are still unavailable. Several alternatives have been described as additional treatment
after initial endoscopic debridement to optimize clinical results, such as the use
of a high-flow waterjet system [18 ]
[19 ]
[20 ]
[21 ]
[22 ], the use of hydrogen peroxide [23 ]
[24 ], and a vacuum-assisted closure system [25 ]
[26 ]
[27 ]. These techniques seem promising, but to date, only small case series have been
published.
All 12 patients in this study underwent minimally invasive DEN using the EndoRotor.
Of these 12 patients, three were treated after initial failure using conventional
instruments. During the procedures, the rotation speed of the EndoRotor catheter was
set at 1000 or 1700 revolutions per minute at the discretion of the treating endoscopist,
with suction set at 620 mmHg negative pressure, the maximum achievable level. For
optimal removal of tissue, the angle of the device relative to the necrotic tissue
plane is important. The cutter opening should be directed to face the necrosis with
direct contact. The high vacuum setting alone was not always able to suck in all of
the tissue, and the best results were achieved by “trapping” the necrotic tissue between
the cavity wall and the cutter opening of the catheter. This resulted in relatively
fast and highly effective removal of necrotic tissue. There was no need for “blind”
grabbing into the necrosis as is often unavoidable with snare-based instruments. We
were able to clear the walls of the necrotic cavities of any necrosis left without
damaging the wall itself. The tip of the EndoRotor cutter remains visible at all times
making it a very safe procedure, and providing good control of what is being cut and
what is not.
A median number of two procedures was required to achieve complete removal of necrotic
tissue in this series. Several studies have reported on the mean number of interventions
necessary to completely remove necrotic tissue using conventional instruments, with
a weighted mean of four endoscopic necrosectomy procedures per patient [11 ]
[28 ]. Two large studies, published by Papachristou et al. [29 ] and Seifert et al., described a mean of four and even six procedures, respectively,
with a maximum of 10 and even 35 endoscopic procedures. A more recently published
study by Van Brunschot et al. [30 ] showed that 41 % of patients undergoing transluminal endoscopic necrosectomy required
at least three procedures to achieve complete removal. Our study was not powered to
detect a difference in the number of procedures compared to historical series or between
the two technical iterations of the EndoRotor. Nevertheless, taking into consideration
the large sizes of the necrotic collections in our series, the current data indicate
that this device and, in particular, the adapted design of the EndoRotor, is even
more effective to achieve complete clearance of the pancreatic necrosis in terms of
the number of procedures required and the time spent.
Despite the reduction in overall mortality over recent years, acute necrotizing pancreatitis
is still associated with high morbidity and mortality rates [3 ]
[4 ]
[5 ]. In patients treated endoscopically, a complication rate of 36 % was reported in
a recently published meta-analysis, with bleeding as the most prevalent complication
(22 %) [30 ]. In the current study, no device-related complications occurred during the procedures
or within the first 24 hours. One patient died as a result of ongoing organ failure
eight days after the last necrosectomy procedure; the necrotic cavity was very comprehensive
and complete removal of necrotic tissue was therefore not achieved. Another patient
died as a result of pancreatic carcinoma, which became apparent two months after successful
endoscopic treatment of pancreatic necrosis using the EndoRotor, in the absence of
complications. One patient suffered from a gastrointestinal bleed occurring two days
after the necrosectomy and necessitating coiling of the splenic artery. During the
necrosectomy, no bleeding or damage to any exposed vessel was observed. We believe
that the bleeding was a direct result of ongoing inflammation within the remaining
necrosis leading to pseudoaneurysm formation and eventually bleeding. Based on our
experience, we believe that the risk of bleeding using this device is low, because
removal of necrosis occurs in a very controlled way under direct endoscopic vision.
The general opinion of the endoscopists on the use of the EndoRotor for pancreatic
necrosectomy was encouraging in the way that this novel tool was judged to be easy
to use, effective, having a low risk of complications, and being of additional value
in the treatment of patients with acute necrotizing pancreatitis. A limitation of
the current study is the number of patients who were included. This was partly compensated
by the fact that, in total, 27 necrosectomy procedures were carried out. This study
was set up as a prospective cohort study testing initial feasibility and safety. These
results should be confirmed by others and in comparative series.
The EndoRotor is the first instrument specifically designed to facilitate easy, safe,
and rapid removal of necrotic tissue in patients with (infected) WOPN under direct
endoscopic vision. Its unique design overcomes some of the inherent problems and shortcomings
that are associated with conventional instruments currently used for endoscopic necrosectomy.
Prospective comparative evaluation of the EndoRotor in a larger series of patients
is required to confirm these favorable observations and to further evaluate its safety
profile and clinical efficacy.
