Keywords Endoscopic ultrasonography - Pancreas - Interventional EUS - Fine-needle aspiration/biopsy
Introduction
Pancreatic cancer has a poor prognosis, with over half of cases diagnosed at distant
stage [1 ]. Fewer than 20% of patients have resectable disease [2 ]. Delivery of high-dose ablative radiation therapy (RT) is effective for some cancers,
but for the pancreas, the amount of radiation possible is limited by surrounding anatomy
and the risk for significant duodenal toxicity [3 ].
The concept of creating space proximal to a primary tumor has shown promise in prostate
cancer populations undergoing RT [4 ]
[5 ], prompting an interest in its application for other cancers. Polyethylene glycol
hydrogel [6 ]
[7 ]
[8 ] addresses some limitations of previously tested biomaterials used for spacing [9 ]. Hydrogel can be distributed evenly for precise and durable spacing, and the
water-soluble molecules reliably degrade and excrete through renal filtration months
after
application [10 ].
There is evidence from porcine and cadaver studies of the feasibility of endoscopic
ultrasound (EUS)-guided injection of PEG hydrogel into the space between the head
of the pancreas (HOP) and the duodenum, increasing space for RT administration [6 ]
[11 ]. A first-in-human single-site pilot study showed feasibility of administering PEG
hydrogel and creating space in patients with pancreatic cancer undergoing ultra-hypofractionated
stereotactic body radiotherapy (SBRT) [7 ]. In this multi-site study of a similar population, we sought to evaluate with greater
generalizability the feasibility, RT benefits, and safety of using an injectable PEG
hydrogel to create space between the pancreas and duodenum. Here we report on technical
and safety outcomes of PEG hydrogel administration; RT findings have been reported
separately [12 ].
Patients and methods
Study design
This was a multicenter prospective, single-arm early feasibility study (April 24,
2019-May 1, 2021) of a PEG hydrogel used as a spacer in six patients with localized
(resectable, borderline resectable or locally advanced) pancreatic cancer for whom
a course of RT was indicated. The trial was registered at ClinicalTrials.gov (NCT03998566).
Patients
Inclusion criteria were: age ≥ 18 years; biopsy-confirmed localized pancreatic cancer
in the head or neck of the pancreas visualized via CT or other imaging modality with
no evidence of distant metastasis as defined by National Comprehensive Cancer Network
guidelines; tumor clearly delineable from duodenum and no clear evidence of invasion
of the duodenum; RT was indicated; medically fit to undergo endoscopy; screening/baseline
laboratory testing met established laboratory value criteria; and life expectancy
of at least 9 months.
Patients were excluded for any of the following: RT was contraindicated; history of
previous thoracic or abdominal RT; presence of tumor invasion of the duodenum detected
on EUS at time of biopsy; previous Whipple procedure or other resection of pancreatic
tumor prior to screening; active gastroduodenal ulcer or uncontrolled watery diarrhea;
history of chronic renal failure; history of uncontrolled diabetes; enrolled in another
investigational drug or device trial that would clinically interfere with this study;
or unable to comply with the study requirements or follow-up schedule.
Patients provided written consent before study-specific procedures were performed.
The study was approved by the Institutional Review Board of the respective clinical
sites (July 3–31, 2018).
Index procedure
TraceIT Tissue marker (“TraceIT,” Augmenix, Waltham, Massachusetts, United States)
is an absorbable radiopaque PEG hydrogel material, currently FDA-cleared to radiographically
mark soft tissue during a surgical procedure or for future surgical procedures for
at least three months after injection. In this study, TraceIT was administered for
an off-label use, soft tissue spacing in the duodenal pancreatic groove, to temporarily
position the duodenum away from the pancreas in patients undergoing RT for treatment
of pancreatic cancer.
Enrolled patients underwent placement of fiducial markers into the pancreatic head
tumor and peri-duodenal administration of PEG hydrogel within the same EUS procedure.
Patients were positioned in the left lateral position and a linear EUS was used to
identify the duodenal wall, pancreas head tumor, and HOP interface. Under EUS guidance,
a 22G fine needle aspiration was advanced into the potential space between the duodenum
and pancreas. PEG hydrogel was prepared per the instructions for use. The single-use
kit consists of a pre-filled glass syringe containing the absorbable radiopaque cross-linked
PEG hydrogel spacer and a delivery system (syringe and needle) ([Fig. 1 ]
a ). Immediately before injection, the PEG hydrogel was mixed five times between the
two syringes and placed in a plastic receiving syringe. Once the needle was confirmed
to be in the proper position, hydrogel was injected in 1- to 2-mL increments. This
process was repeated as the needle was repositioned around the target ([Fig. 1 ]
b ).
Fig. 1 PEG hydrogel. a Injectable PEG hydrogel used in the study.
b PEG hydrogel injection (blue) into the peri-duodenal space
(white outline) between the duodenum and pancreas tumor using endoscopic ultrasound
guidance.
After each procedure, the following was collected: ability to access injection site
and inject PEG hydrogel; average duodenal space measurements on CT measured at three
points along the HOP; injection procedure duration; ease of device use; device malfunctions;
and adverse events (AEs) per NCI Common Terminology Criteria for Adverse Events (CTCAE
v4). Any event precipitating an intestinal acute CTCAE score of 2 or higher was documented
as an AE.
RT simulation planning was performed prior to and following PEG hydrogel placement
for evaluation and comparison of duodenal dose/dose distribution and to assess differences
in RT dosing parameters. RT was to be initiated no later than 28 days following PEG
hydrogel administration.
Within 2 to 6 weeks after RT completion, patients were restaged to determine whether
they could progress to surgery. If surgical resection was successful, pathological
data were recorded. All patients were evaluated at minimum 3 and 6 months post-index
procedure. MRI was performed at the six-month visit to evaluate for PEG hydrogel presence
in unresected patients. Throughout the study, patients were assessed for duodenal
AEs. Additional follow-up clinic visits were performed per standard care at 12 and
18 months at minimum.
Study endpoints
Feasibility was defined as technical success, i.e., the ability to administer and
visualize PEG hydrogel and create space between the duodenum and HOP; the technical
success rate was calculated as the proportion of patients who achieved technical success
in the intention-to-treat (ITT) population. RT benefits were assessed via comparison
of pre- and post-administration RT plans and have been reported separately [12 ].
Patients were monitored for AEs and RT toxicity (using CTCAE v4) and in particular,
for PEG hydrogel administration procedure-related events resulting in a delay in initiation
of RT, as reviewed and adjudicated by a Clinical Events Committee.
Additional data collection included: incidence of resection; histology of duodenal
tissues when resection was performed; incidence of acute (≤ 3 months) and late (>
3 months) duodenal toxicity for unresected patients; theoretical dose escalation from
post-injection treatment plan (reported separately); PEG hydrogel persistence (6 months
post-treatment in unresected patients); and progression-free and overall survival
through follow-up. Histology of the duodenal tissues was assessed when resection was
performed. The pathological duodenum damage score was rated by a local board-certified
gastrointestinal surgical pathologist using the methods outlined by Verma et al. (1
= no/minimal signs of mucosal damage, 2 = moderate damage, 3 = severe damage) [13 ]. Pathologic response was graded according to the College of American Pathologists
(CAP) Protocol for pancreatic cancer [14 ].
Screening/baseline data collection
Information collected at screening and baseline included: demographics; disease documentation
(tumor location, initial resectability status, tumor staging, pretreatment tumor dimension);
medical/surgical history and status (concomitant medical conditions, prior surgeries,
prior therapies); physical examination; assessment of baseline duodenal symptoms;
and baseline concomitant medications.
Results
Enrollment in analysis population
A total of eight patients were consented. One was consented but did not meet the eligibility
criteria and was deemed a screen failure. Another was consented but withdrew consent
before eligibility criteria was verified. This left six patients in the ITT population.
There were no major protocol deviations with the potential to affect the study.
Baseline characteristics
The median and age range of the patients enrolled were 69.5 and 60 to 80 years,
respectively ([Table 1 ]). The majority were male (66.7%) and 50.0% were White. All subjects received
neoadjuvant therapy for pancreatic cancer. There were four patients (66.7%) who were
in
borderline resectable stage while two patients (33.3%) were in the locally advanced
stage.
RT planning and treatment characteristics have been reported separately; briefly,
five of
six patients were treated with SBRT and one with intensity-modulated RT.
Table 1 Patient characteristics.
Baseline demographics and clinical characteristics
SD, standard deviation; BMI, body mass index.
Patients
(N = 6)
Age (n = 6 patients)
Mean ± SD
69.5 ± 7.4
Median
69.5
Range
60.0 – 80.0
Sex
Male
4 (66.7%)
Female
2 (33.3%)
Ethnicity
Hispanic or Latino
1 (16.7%)
Not Hispanic or Latino
5 (83.3%)
Race
White
3 (50.0%)
Black/African American
2 (33.3%)
Asian
1 (16.7%)
American Indian/Alaska Native
0 (0.0%)
Native
Hawaiian/Pacific
Islander
0 (0.0%)
Smoking history
Current smoker
0 (0.0%)
Past smoker
2 (33.3%)
Never smoked
4 (66.7%)
BMI (n = 4 patients)
Mean ± SD
29.9 ± 5.4
Median
32.1
Min, max
22.0 – 33.5
Pancreatic cancer
medical history
Neoadjuvant therapy for pancreatic cancer
No
0 (0.0%)
Yes
6 (100.0%)
Largest pretreatment dimension of tumor (cm)
(n = 6 patients)
Mean (SD)
3.0 ± 0.6
Median
3.2
Min, max
2.2–3.5
Initial resection status
Borderline resectable
4 (66.7%)
Locally advanced
2 (33.3%)
Initial tumor anatomic stage
Stage 0
0 (0.0%)
Stage IA
0 (0.0%)
Stage IB
3 (50.0%)
Stage IIA
1 (16.7%)
Feasibility/technical success
The ability to place and visualize PEG hydrogel, creating space between the duodenum
and the HOP was successful in all patients (100%, N = 6).
PEG hydrogel administration and space created
The median time between hydrogel injection and initiation of RT was 18 days (range:
9–19
days). PEG hydrogel was distributed in small volumes (approximately 1 to 2 mL per
injection
for up to a total of 10 mL) at several areas along the proximal portion of the duodenum
in
the areas closest to the HOP. The precise location of each injection varied according
the
anatomy and tumor characteristics of each case. [Table 2 ] presents injection characteristics by patient. [Fig. 2 ] shows the space created in two patients on EUS. [Fig. 3 ] shows the hydrogel between the duodenum and pancreatic head tumor on CT scan.
Table 2 Injection characteristics by patient.
Patient
22 gauge dilution ratio
Number of injections
Total volume
Pre-injection space (mm)
Post-injection space 1 (mm)
Post-injection space 2 (mm)
Post-injection space 3 (mm)
1
1:1
3
3 mL
1
6.84
3.96
8.95
2
3:1
3
2.5 mL
1
5.56
7.89
5.73
3
3:1
4
3.75 mL
1
10.78
7.99
6.76
4
3:1
3
3 mL
1
5.85
3.19
3.89
5
No dilution
12
10 mL
1
12.35
10.52
9.16
6
No dilution
6
6 mL
1
10.16
8.78
10.83
Fig. 2 Injection and post-injection EUS images. a Transduodenal PEG hydrogel injection with a 22 G EUS FNA needle (needle tip circled
in red) with a hypoechoic nodule (margins indicated by arrows) developing around the
needle tip after the injection. b Injection of PEG hydrogel with 22 G EUS FNA needle between the duodenum and the pancreatic
head mass with a hypoechoic nodule forming around the tip of the needle. c After injection of PEG hydrogel in the same patient as in Fig. 2b , there is a hypoechoic nodule (arrow) separating the pancreatic head tumor from the
duodenal wall by 7.6 mm.
Fig. 3 PEG hydrogel between the duodenum and pancreatic head tumor on CT scan. a Contrast-enhanced CT scan showing pancreatic head tumor and metal biliary stent prior
to PEG hydrogel injection. b , c , d Contrast-enhanced CT scan post-injection showing the PEG hydrogel (in blue outline)
clearly visible and separating the duodenal wall from the pancreatic head for stereotactic
radiation therapy (axial, coronal, and sagittal planes in same patient as shown in
Fig. 3a ).
Of the three unresected patients able to be assessed for persistence at two to six
weeks post-RT, PEG hydrogel was still present in all. At 6 months post-RT, in the
two patients who received imaging and were assessed for persistence, PEG hydrogel
was not detected. Stability was similar from the time of placement to 2 to 6 weeks
afterward, based on the distance from a fiducial marker (8.2 ± 5.9 mm vs 8.5 ± 10.2
mm, N = 3).
Safety/adverse events
No procedure-related AEs resulting in a delay in RT initiation were reported. Two
patients (33.3%) had procedure-related AEs; these took place at two sites. One of
these patients was reported to have sinus bradycardia; the other had nausea, chills,
dehydration, and stomach pain. All of these AEs were noted as resolved. There were
no device-related AEs. There were no reports of pancreatitis, perforation, bleeding,
or infection.
Progression-free and overall survival
Mean progression-free survival for all patients was 17.6 ± 4.7 months, ranging from
8.5 to 21.2 months ([Table 3 ]). Progression-free survival was longer for the three patients with resected tumors
compared to the three who did not undergo resection (18.9 ± 2.3 months vs 16.2 ± 6.8
months). Overall mean survival (to death or end of study) was 21.6 ± 6.9 months (12.0
to 31.3 months).
Table 3 Progression-free and overall survival.
Total
(N = 6)
Resected
(N = 3)
Not Resected
(N = 3)
SD, standard deviation.
Progression-free survival (months)
(in patients with tumor progression
event)
Mean ± SD
17.6 ± 4.7
18.9 ± 2.3
16.2 ± 6.8
Median
18.9
18.8
19.0
Min-max
8.5–21.2
16.7–21.2
8.5–21.1
Overall survival (months)
Mean ± SD
21.6 ± 6.9
20.0 ± 3.4
23.3 ± 10.0
Median
21.7
20.0
26.5
Min-max
12.0–31.3
16.7–23.4
12.0–31.3
Acute and late duodenal toxicity
All six patients had one or more grade 1 duodenal AEs, with 100% (6/6) having acute
and 66.7% (4/6) having late duodenal AEs. Four patients (66.7%) had grade 2 duodenal
AEs. Of these four patients, three had acute and three had late duodenal AEs. Two
patients (33.3%) had grade 3 duodenal AEs occurring after three months. No patients
had grade 4 or 5 duodenal AEs. (See Supplementary Material for tabular summary.)
Pathology
Of the three patients that qualified for tumor resection, 100% showed no/minimal signs
of mucosal damage. One subject (33%) had complete pathologic response (no cancer cells
detected on pathology specimen), and two (66.7%) had near complete response. Pathology
is included in tables in the Supplementary Material .
Discussion
High-dose ablative radiation is an important advancement in RT, enabling precise delivery
of high-dose radiation to a small tumor volume [15 ]
[16 ]. SBRT (the approach used in the majority of patients in this study) has shown promise
for use in pancreatic cancer [17 ]
[18 ], but the tissues adjacent to the pancreas raise substantial concerns about late
gastrointestinal toxicity [19 ]
[20 ]. The duodenum, a radiosensitive organ, limits RT dosing of the HOP [15 ]
[21 ].
This study adds to findings from animal and cadaver studies and a single-site pilot
study [6 ]
[7 ]
[11 ], all of which showed the feasibility of administering and creating space with PEG
hydrogel between the HOP and the duodenum. The current study provides evidence for
a lack of AEs that would delay RT.
A leading concern following injection of a spacing agent would be causing acute pancreatitis.
In this study, there were no reports of pancreatitis following the index procedure.
Another potential concern is injection into the duodenal wall and resulting toxicity.
The patients in this study who were resected did not have any significant toxicity
to the duodenal wall. In the previous porcine study, the same PEG hydrogel was injected
directly into duodenal wall and no necrosis was observed [11 ].
Conclusions
Although the multi-site design of this study builds on the previous evidence of feasibility,
conclusions about the clinical relevance of PEG hydrogel for use in borderline resectable
or locally advanced pancreatic cancer are limited by small sample size and the high
morbidity and mortality rate. Poor survival outcomes are typical of a pancreatic cancer
population, but none of the serious AEs that led to death were attributed to PEG hydrogel
or RT. The concept of enhancing RT for pancreatic cancer by using PEG hydrogel to
create peri-duodenal space warrants further investigation.
