A phase 2 intervention study:
Detection of early esophageal neopastic lesions by quantified fluorescence molecular endoscopy using oral and topical administration of bevacizumab-800CW and cetuximab-800CW

Patients with Barrett*s esophagus (BE) have an increased risk of developing
esophageal adenocarcinoma (EAC). Esophageal adenocarcinoma (EAC) causes six
percent of cancer related deaths worldwide, with studies predicting a rise in
the incidence of EAC. Early diagnosis is challenged by asymptomatic disease
progression. Consequently, late-stage detection translates to a five-year
survival rate of 15 - 20%. Current endoscopic surveillance with random biopsies
according to the Seattle protocol is costly and time-consuming with a potential
high miss rate. This stresses the great need for better endoscopic
visualization and thereby the ability to take targeted biopsies to improve
detection of esophageal (pre)malignant lesions during surveillance endoscopy of
patients at risk of developing these esophageal malignancies. Optical molecular
imaging of neoplasia associated biomarkers seems a promising technique to
accommodate this need. It is known that several biomarkers are overexpressed in
dysplastic and neoplastic tissue and these biomarkers can be a valid target for
quantified fluorescence molecular endoscopy (qFME).

One of these biomarkers is vascular endothelial growth factor A (VEGFA). The
phase I study, named VICE, completed within the UMCG, showed that synchronal
use of VEGFA-guided quantified fluorescence molecular endoscopy and
high-definition white light endoscopy (HD-WLE) with narrow-band imaging (NBI)
could improve early lesion detection by 33% using the topically applied tracer
approach compared to HD-WLE/NBI endoscopy. Within the phase II study, the
ESCEND study, we further evaluated topical administration of bevacizumab-800CW
and qFME in sixty patients (unpublished data). qFME with bevacizumab-800CW
detected 71 lesions compared to 56 lesions detected by HD-WLE/NBI performed by
the BE expert endoscopist (+27%) and 33 lesions detected by the non-BE expert
endoscopist (+115%). Additionally, two other tracers were investigated in a
smaller subset of patients, cetuximab-800CW, targeting epidermal growth factor
receptor (EGFR) and the nonspecific near-infrared fluorescent dye indocyanine
green (ICG). qFME with cetuximab-800CW detected 17 lesions compared to 12
lesions detected by HD-WLE/NBI performed by the BE expert endoscopist (+42%)
and 9 lesions detected by the non-BE expert endoscopist (+89%). The ESCEND
study has confirmed the great potential of qFME looking at additional lesion
detection initially missed by HD-WLE, especially compared to non-BE expert
centers.

However, we hypothesized, that we can potentially identify more additional
lesions by simultaneous use of two targeted tracers because of variable
expression of VEGFA and EGFR within EAC. Until now, solely intravenous and
topical administration of the tracers has been investigated. However,
optimization of tracer administration and shortened incubation is necessary for
clinical translation and implementation of this new technique from BE expert
centers to regional non-expert centers. BE surveillance procedures normally
takes up to 15 minutes at regional hospitals, of which most of the procedural
time is needed to take biopsies according to the Seattle protocol. Introducing
qFME into these hospitals would elongate the procedure time with at least 10 -
15 minutes. This would increase healthcare costs and put increased pressure on
BE healthcare. Ideally, the gastroenterologist can immediately start with the
qFME procedure without any incubation time while maintaining the best TBR
possible. Oral administration by drinking the tracer prior to the procedure
would eliminate incubation time and its consequences. Quantified qFME with oral
tracer administration and targeted biopsies could potentially replace the
time-consuming, high miss rate Seattle protocol, improve lesion detection and
decrease global healthcare costs associated with BE.

Therefore, we propose to investigate the following:
1. Can qFME procedural time be diminished by oral administration of the
fluorescent tracer(s)?
2. Does combining bevacizumab-800CW and cetuximab-800CW improve lesion
detection compared to single tracer administration?

Primary objectives:
The primary objectives are to evaluate the feasibility of oral administration
of bevacizumab-800CW and cetuximab-800CW for detection of neoplasia in BE
patients compared to HD-WLE to shorten the qFME procedure and test whether
combined tracer administration increases lesion detection.

Secondary objectives:
1. To collect safety data on oral administration of (combined)
bevacizumab-800CW and cetuximab-800CW by spraying.
2. To (semi)quantify and evaluate the in vivo NIR fluorescent signal of
bevacizumab-800CW and cetuximab-800CW combined by using the spectroscopy probe
and compare this to the ex vivo VEGFA/EGFR levels in the resected mucosal
lesions and/or biopsies taken.
3. Eventually further specify and objectify the improvement of qFME by
standardization and tracer administration, by dose optimization, leading to
reduction of unnecessary biopsies.

The study design consists of three study arms. An overview of the study design
is shown in figure 1. Arm 3 will only be performed if oral administration is
not feasible. We want to include non-dysplastic BE as control group to confirm
that our fluorescent tracers do not bind to Barrett tissue without a lesion to
show their specificity.
1. Oral bevacizumab-800CW in dysplastic BE
2. Oral cetuximab-800CW and combined oral bevacizumab-800CW/cetuximab-800CW in
dysplastic BE or oral bevacizumab-800CW in non-dysplastic BE
3. Topical bevacizumab-800CW and combined topical
bevacizumab-800CW/cetuximab-800CW in dysplastic and non-dysplastic BE

After signing of the informed consent by both the patient and the study
investigator, the patient will undergo the combined procedure (qFME and HD-WLE
with biopsies or EMR/ESD). In case of EMR/ESD, the procedure will be performed
under propofol sedation and therefore the patient*s vital parameters will be
monitored closely by an employee of anesthesiology department (standard
clinical care).

In arm 1 and 2, patients will receive oral tracer administration and are asked
to drink Infacol to clean the esophagus. This is standard care before endoscopy
procedures. The tracer bevacizumab-800CW and/or cetuximab-800CW will by orally
administered by drinking two cups of 15 ml in an upright position around ten
minutes before the study procedure. Endoscopic procedures will be performed by
a specialized BE endoscopist. The procedure will start with HD-WLE inspection.
Subsequently, NIR-imaging will take place, by using a fiber-bundle for
fluorescence molecular endoscopy and spectroscopy probe, which can both be
inserted subsequently through the working channel of the HD-WLE. The HD-WLE
suspected (pre)malignant lesion, normal esophageal epithelium, gastric mucosa
and, when present, non-dysplastic BE segment will be evaluated according to
their fluorescent intensity. Biopsies will be taken according to the Seattle
protocol. All endoscopically visible lesions will be biopsied (maximum of 6),
followed by four-quadrant biopsies with 2 cm interval from the lower esophageal
sphincter to the squamocolumnar junction (10,11). All biopsies will be formalin
fixed paraffin embedded (FFPE).

We will perform an EMR/ESD during the therapeutic procedure instead of taking
biopsies from additional fluorescent lesions detected with qFME that were
initially missed by HD-WLE. qFME shows a high positive predictive value of 96%.
The positive predictive value might even be higher because of sampling error
where qFME can detect fluorescence while biopsies do not reach the depth of the
dysplasia. We think that this is sufficient evidence to support that a
therapeutic EMR/ESD of invisible fluorescent lesions being beneficial for the
patient. An EMR/ESD is associated with < 1 % risk of perforation and bleeding.
Fluorescence endoscopy procedures will be digitally documented. The complete
procedure will be recorded.

The oral administration will be suspended if no accumulation of
bevacizumab-800CW in (pre)malignant lesions can be demonstrated in the dose
finding stage. We will then continue directly to arm 3 of the study design
where we topically administer bevacizumab-800CW and compare this to combined
topical tracer administration with bevacizumab-800CW and cetuximab-800CW.

In arm 3 of the study design, that will only be performed when oral tracer
administration is not feasible, patients will receive topical tracer
administration and the esophagus will be cleaned with acetyl cysteine 0.1% at
the beginning of the procedure. After standard HD-WLE, patients will be
topically administered by spraying the tracer(s) with subsequent standard 5
minutes incubation time. Followed by rinsing the esophagus with water to wash
away unbound tracer after which qFME can be performed.

Biopsies will be taken according to the Seattle protocol. The biopsies will be
processed in the same manner as standard clinical biopsies. If necessary, the
procedure is followed by routine EMR/ESD performed by the same physician
performing the standard procedure. EMR/ESD specimens and biopsies will be
analyzed ex vivo and as a fresh specimen analyzed FFPE, according to the
standard procedure. Elderly patients with relevant comorbidity will be
hospitalized the evening prior to the procedure when there is a scheduled
EMR/ESD, according to standard EMR/ESD-protocol. In some cases, treated
patients will stay hospitalized a subsequent night.

Moreover, the study will be suspended immediately if any severe adverse event
(SAE) related to the administration of the tracers occurs in any of the
patients. The design of this study warrants maximal data collection, while
risks and burden for patients are minimized.

Patients will receive oral administration of the tracer bevacizumab-800CW
and/or cetuximab-800CW prior to the endoscopy procedure During the procedure
standard HD-WLE together with fluorescence endoscopy and MDSFR/SFF spectroscopy
will be performed. Furthermore we will collect biopsies according to standard
Seattle protocol. We will perform an EMR/ESD during the therapeutic procedure
instead of taking biopsies from additional fluorescent lesions detected with
qFME that were initially missed by HD-WLE because fluorescence endoscopy shows
a high positive predictive value of 96%. Patients will be sedated during the
endoscopic procedure following standard clinical care procedures.

The administration risks of bevacizumab-800CW and cetuximab-800CW are reported
in the investigational medicinal product dossier (IMPD) (bevacizumab-800CW:
version 6.0, June 2020, section 2.4, cetuximab-800CW: version 2.0, June 2020,
section 2.4). No adverse events were reported from previous administrations
with bevacizumab-800CW with more than 120 patients included. Bevacizumab-800CW
will be orally administered in 4.5 mg and 9 mg, which is a 50 - 100 times lower
dose than therapeutic bevacizumab. Therefore, no (serious) adverse events are
expected. Topical administration of cetuximab-800CW did not show any adverse
events in the ESCEND study. Bevacizumab-800CW and cetuximab-800CW are both
human monoclonal antibodies and are not expected to interact with anything else
than their own target. It is not expected that the intact antibody is taken up
after esophageal passage as it is expected to be dissolved by gastric acids.

Oral administration can increase the risk of choking after swallowing the
tracer, especially in patients with reflux disease associated with a higher
prevalence of swallowing impairment. However, administration of the tracer will
be performed in a controlled hospital environment where the patient is prepared
for the endoscopy procedure according to standard care. Patients normally drink
Invacol from an open cup before a procedure. The staff is well trained and can
react to adverse events when adverse events related to swallowing occur. We
will ask the patient at the start whether there were problems with drinking the
tracer from an open cup.

The endoscopic procedure will be prolonged with 15 minutes to enable
fluorescence imaging (qFME and MDSFR/SFF). The fiber of qFME or MDSFR/SFF can
be inserted through the working channel of HD-WLE endoscope and therefore the
risks of fluorescence imaging are comparable to standard HD-WLE endoscopy and
assumed negligible. The biopsy procedure may cause some bleeding which resolves
spontaneously and does only need intervention in very rare cases (1 per 5000
endoscopies per year).

We have confirmed the potential of qFME in the phase II ESCEND study. Based on
these results, we want to perform an EMR/ESD during the therapeutic procedure
instead of taking biopsies from additional fluorescent lesions detected with
qFME. qFME shows a high positive predictive value of 96% when confirming
fluorescent lesions as dysplasia by pathology. Three lesions were marked
positive by qFME while biopsies could not confirm dysplasia. The positive
predictive value might even be higher because of sampling error where qFME can
detect fluorescence while biopsies do not reach the depth of the dysplasia. We
think that this is sufficient evidence to support a therapeutic EMR/ESD of
invisible fluorescent lesions being beneficial for the patient. An EMR/ESD is
associated with < 1 % risk of perforation and bleeding.