Clinical Trial Details
— Status: Completed
Administrative data
NCT number |
NCT04821349 |
Other study ID # |
ArtiC STUDY-NP4350 v.05 140920 |
Secondary ID |
|
Status |
Completed |
Phase |
N/A
|
First received |
|
Last updated |
|
Start date |
February 16, 2021 |
Est. completion date |
October 1, 2022 |
Study information
Verified date |
November 2022 |
Source |
Fondazione Poliambulanza Istituto Ospedaliero |
Contact |
n/a |
Is FDA regulated |
No |
Health authority |
|
Study type |
Interventional
|
Clinical Trial Summary
Capsule Endoscopy (CE) is a safe, patient friendly and easy procedure performed for the
evaluation of gastrointestinal tract unable to be explored via conventional endoscopy. The
most common indication to perform SBCE is represented by Suspected Small Bowel Bleeding
(SSBB). According to the widest meta-analysis available in literature, SBCE shows a
diagnostic yield in SSBB of about 60%, and angiodysplasias are the most relevant findings,
accounting for 50% of patients undergoing SBCE for SSBB. Accordingly, it represents the first
line examination in SSBB investigation for determining the source of bleeding, if primary
endoscopy results negative. Despite its high clinical feasibility, the evaluation of
CE-video-captures is one of the main drawbacks since it is time consuming and requests the
reader to concentrate to not miss any lesion. In order to reduce reading time, several
software have been developed with the aim to cut similar images and select relevant images.
For example, automated fast reading software have demonstrated to significantly reduce
reading time without impacting the miss rate in pathological conditions affecting diffusely
the mucosa (as IBD lesions do). Not the same assumption can be taken for isolated lesions
since several studies reported an unacceptable miss rate for such a detection modality. New
advancements such as artificial intelligence made their appearance in recent years. Deep
convolutional neural networks (CNNs) have demonstrated to recognize specific images among a
large variety up to exceed human performance in visual tasks. A Deep Learning model has been
recently validated in the field of Small Bowel CE by Ding et al. According to their data
collected on 5000 patients, the CNN-based auxiliary model identify abnormalities with 99.88%
sensitivity in the per patient analysis and 99.90% sensitivity in the per-lesion analysis.
With this perspective, it is believable that AI applied to SBCE can significantly shorten the
reading time and support physicians to detect available lesions without losing significant
lesions, further improving the diagnostic yield of the procedure.
Description:
This is a multicenter, multinational, blinded prospective trial, involving a consecutive
series of patients recruited by 12 European centers based on the indication of OGIB:
- after negative upper and lower endoscopy
- France: after negative pregnancy test
- Hb cut-off male: <13, female: <12 Capsule endoscopy will be performed in each site
according to local rules and requirements, and the study protocol will concern only the
post-procedure analysis on reading modalities for each patient, as specified below.1.
Regimen of preparation AI depends on the possibility of the software to "see" images. An
inadequate cleansing level precludes a proper visualization and the impact of
technology. In order to have homogenous results, a standard regimen of preparation is
advisable. The regimen includes a split dose of PEG-based solution (PlenVu, Moviprep) as
recommended by ESGE guideline - technical report. Dose 1 will be administered at 7 pm of
the day before and dose 2 in the morning of the procedure in order to be completed at
least 1 hour before capsule ingestion. On the day before patients can have breakfast and
a light meal. After lunch they should be fasting. Two and 4 hours after capsule
ingestion patient are allowed to drink clear liquids and a light meal, respectively.
2. Standard reading Each center will review the images collected according to the
"normal reading" as recommended by ESGE guidelines. To compare reading time correctly,
readers must read the video without including annotation of images. All lesions should
be considered independently of their relevance. Findings are not marked on every image
if they appear repeatedly on every consecutive image, but if they appear repeatedly with
normal images in between. Annotations should be done after the reading, each lesion
shall be labelled if it belongs to P1/2 category according to the Saurin Classification.
Definition of P1, P2: red spots on the intestinal mucosa or small or isolated erosions
or angioectasia, ulcers, tumors or varices or any other bleeding abnormality.
3. AI-assisted reading Each center will receive from another center anonymized patient
videos that have been uploaded on an encrypted USB key. Indeed, expert
gastroenterologists from each center will proceed with AI-assisted video reading without
knowing which center the video comes from nor the results of the normal reading. Reading
conditions are the same as for standard reading.
4. Consensus reading Results from both normal and AI-assisted reading will be compared.
In the case of significant disagreement between the results from conventional capsule
endoscopy reading and AI-assisted reading (e.g. missed lesions), a consensus review for
each center will be done. Any lesion shall be considered.