Autofluorescence endoscopy and spectroscopy for detection of early Barrett's neoplasia: the AFI-III study

Endoscopic surveillance of Barrett oesophagus (BO) patients is recommended to
detect high-grade intraepithelial neoplasia (HGIN) or early cancer (EC) at a
curable stage. With standard endoscopy, however, it is difficult to distinguish
areas with HGIN/EC. In the absence of visible lesions, random biopsies are
obtained for histological assessment of neosplasia, but these random biopsies
may miss dysplastic lesions (sampling error). The endoscopic detection of early
neoplasia may be improved by the use of endoscopic tri-modal imaging (ETMI); a
system that incorporates white light endoscopy (WLE) and autofluorescence
imaging (AFI) for primary detection of early neoplasia and allows for targeted
imaging of suspicious areas with narrow-band imaging (NBI). In a recent
international multicenter study, AFI increased the sensitivity for detecting
early neoplasia from 53% to 90% compared to WLE. Subsequent inspection with NBI
of AFI-positive areas reduced the false-positive rate of AFI from 81% to 26%.
Preliminary results of an international multicenter randomized cross-over trial
also show that AFI increases the detection of early neoplasia with 40%, but
again with a high false positive rate (i.e. do not contain neoplasia). The
false positive rate was reduced from 72% to 47% with NBI, but at the expense of
misclassifying 8 neoplastic lesions as unsuspicious. For the AFI III study the
current AFI system is replaced for a new AFI III algorithm. Inspection with AFI
III may be a better approach to detect early neoplastic lesions in the Barrett
Oesophagus and to reduce the current high AFI false positive rate.
AFI and NBI are based on fluorescence spectroscopy, which measures the
interaction between tissue and light. Every tissue has distinct fluorescence
properties, thus by meticulously measuring the fluorescence spectra, one could
possibly distuinguish between neoplastic and non-neoplastic tissue. However,
the challenge is to uncouple the unspecific reemitted light from physiological
fluctuation, from the specific scattering signal due to pathological processes.
One of the possible solutions to this problem is a tunable, single wavelength
laser. By tuning the wavelength, the specific fluorescence spectra
corresponding to tissue characteristics can be optimized and thus specifically
recognised. Adding in-vivo fluorescence to the AFI III system, may reduce the
high AFI false positive rate compared to NBI inspection.

Study phase 1: In patients with BO undergoing work-up endoscopy for early
neoplasia, we aim to develop a classification to evaluate fluorescence
spectroscopy of neoplastic and non-neoplastic lesions in Barrett epithelium, as
used in phase 2.
Study phase 2: In patients with BO undergoing surveillance or work-up
endoscopy for early neoplasia, we aim to evaluate if AFI III increases the
accurary of detecting early neoplasia. Furthermore we aim to determine whether
the addition of in-vivo fluorescence spectroscopy to the AFI III system further
reduces the high false positive rate compared to NBI inspection.

For this prospective study a total of 50 patients will be included.

In phase 1 a total of 10 patients with proven HGIN or EC in a visual lesion
will be included. During endoscopy the BO is examined with WLE to detect
suspicious lesions, all findings are recorded. Then, the BO is inspected with
AFI and the location, size and macroscopic appearance for additionally detected
lesions are recorded. Visual lesions will be removed by endoscopic mucosal
resection, according to current common therapeutic practice. Subsequently, the
EMR specimen is placed on the ex-vivo spectroscopy set-up and measured,
followed by corresponding biopsies of the measured areas. The biopsies and EMR
specimens are then evaluated by the same expert gastro-intestinal pathologist
to assess the presence of neoplasia. The histological data will be correlated
to the fluorescence spectra. The measurement of fluorescence spectra does not
influence the histological assessment of the EMR specimen by the pathologist.

In phase 2 of this study, a total of 40 patients will be included. 10 patients
referred with HGIN.EC in a visual lesion, 10 patients with HGIN/EC without
visual lesions, 10 patients with LGIN and 10 patients under surveillance for
NDBO. During endoscopy the BO is examined with WLE to detect suspicious
lesions, all findings are recorded. Then, the BO is inspected with AFI and the
location, size and macroscopic appearance for additionally detected lesions are
recorded, followed by inspection with NBI. Subsequently, the endoscope is
replaced by the AFI III endoscope and the oesophagus is inspected with all
modalities, followed by in-vivo fluorescence spectroscopy. A laserconducting
fiberoptic probe, connected to the blue laser and a spectrometer, is passed
through the working channel of the endoscope and positioned above the mucosa.
Biopsies are then taken from all measured areas for histological correlation.
The biopsies will be evaluated by the same expert gastro-intestinal pathologist
to assess the presence of neoplasia.

Next to the general risks associated with upper endoscopy such as irritation of
the throat by introduction of the endoscope, difficult swallowing and
retrosternal pain, the use of ETMI for neoplasia detection and the addition of
in-vivo fluorescence spectroscopy do not increase endoscopy risk. During
endoscopic resection, delayed bleeding may occur in 3.3% of cases, usually
easily manageable with endoscopic hemostatic techniques. Also, a perforation
may occur in 2.4% of cases, usually manageable with endoscopic and conservative
management.
In phase 1, there will be no extra burden for the patient. In phase 2, the
extra burden for the patient will be the switch of endoscopes halfway the
endoscopic procedure.