Keywords
Endoscopy Small Bowel - Capsule endoscopy - Training - Quality and logistical aspects
- Image and data processing, documentatiton
Introduction
Since video capsule endoscopy (VCE) was introduced in the early 2000s, it has been
found to be a safe and effective tool for investigating gastroenterological symptoms
and disease worldwide [1]. The range of indications for its use has expanded in recent years due to improvements
in both hardware and software [2]
[3], and in certain clinical circumstances, it has replaced endoscopy as investigative
or monitoring tools [4]
[5]. European guidelines have standardized VCE use for both small bowel and colonic
disease. They include technical guidelines and quality performance measures [6]
[7]
[8].
More recent advances in capsule technology have included the addition of machine learning
and artificial intelligence (AI)-assisted reading [9]; however, despite this, reporting continues to be a time-consuming process, which
is confined to centers with expert readers. The time to read each VCE and availability
of expert readers is often seen as a barrier to the development or expansion of appropriate
capsule services globally. While commercial reading services are available, the cost
can be prohibitive for many centers.
Similarly, new legal focus on data protection at a European and global level (General
Data Protection Regulation) makes sharing of videos among centers and expert readers
for the purposes of quality assurance or report validation challenging.
While AI-assisted VCE is very likely to improve reading speed in the very near future,
and therefore, is likely to expand VCE capacity in centers with expert readers, it
does not allow for sharing videos between hospital centers. This will continue to
be a hindrance for the development of VCE in new and smaller centers.
Medtronic (Medtronic Limited, Watford, UK) recently launched a commercial cloud-based
platform with two-way encryption, which in theory allows for secure remote reading
of VCEs by a selection of pre-approved expert readers located anywhere in the world.
This system could improve access to VCE by enabling centers without trained readers
to initiate a service for their patients, and would also allow readers to share images
and videos with colleagues on a secure platform as a quality assurance process. In
addition, it could promote training, a hybrid-working model, and the development of
after-hours emergency reading service similar to that used for radiology services.
The aim of this study was to assess the performance of a cloud-based VCE reading platform
in a busy tertiary referral service.
Patients and methods
The system
PillCam (Medtronic Limited, Watford, UK) remote reading platform (RRP) is a secure
encrypted, cloud-based reading environment allowing for remote reading, analysis,
and report creation which incorporates the latest RAPID Reader software. PillCam Sync
is a component of the platform installed locally at hospitals and enables uploading
of de-identified studies for review and downloading of the relevant, locally re-identified
reports. The video remains available on the local machine, even while the de-identified
copy is in the cloud ([Fig. 1]). Authorized users can access, analyze, and generate reports about these studies
via the online RRP, which is identical to the standard Medtronic PillCam Desktop reader
([Fig. 2]).
Fig. 1 Medtronic PillCam Cloud user homepage [rerif]. Source: Medtronic
Fig. 2 PillCam Remote Reading Platform (RRP). [rerif]Source: Medtronic
Remote readers are assigned and given access to a specific archive by a Medtronic
system administrator. Any remote reading clinician with access to that specific archive
can choose any study that has been uploaded to that archive to read; the study is
then locked to this clinician and it can only be unassigned by the same clinician
or by the Medtronic system administrator. Once PillCam studies are uploaded, they
are available in the cloud for 30 days, even after they have been assigned to a reader,
and can be edited during that time. Once the final report is approved and the analyzed
video is downloaded, the video is removed from the platform and reader access.
Study population
Following installation of the system in our unit, and an internal pilot study, all
routine VCE studies performed over an 8-month period were automatically uploaded and
read from the cloud-based platform. PillCam Colon2 and PillCam SB3 capsules were used
as appropriate following standard triage and procedure protocols. There were no changes
to procedure protocols (small bowel, colon, pan-intestinal); the only difference was
that videos were uploaded and read on the cloud rather than on the local hospital
server.
Patient demographic and procedure specifics were recorded from our local VCE database,
while remote reader technical information was collected from the PillCam Capsule Endoscopy
Platform. Technical outcomes included: video upload/report download success rates
and speeds, video analysis and technical success, encryption/decryption rates, and
overall procedure success. Reader-reported outcomes were assessed using an online
survey including direct questions and invited comments.
Results
Technical data
A total of 377 VCEs were completed during the timeframe (318 SBCEs, 59 CCEs). Feedback
was collected from seven expert users (2 consultants, 3 NCHDs, and 2 nurse specialists).
Video upload, video analysis, and report download success rates were 100%. The only
technical issue was an upload delay in two studies, both < 24 hours. There were no
encryption/decryption errors and no breaches in General Data Protection Regulation
protocols. The overall procedure success rate was 100%. In addition, the RRP has been
successfully used to support remote access to multiple participants from different
sites to review video clips and images at our quality assurance and training meeting.
User-reported outcomes
In all, seven of seven respondents felt the cloud-reading platform was “very easy”
to access and use. In contrast, six of seven found it “difficult” or “extremely difficult”
to access and use the old desktop hospital server-based system ([Fig. 3]). In addition, six of seven felt the RRP increased department efficiency and would
“definitely” incorporate the RRP into future practice.
Fig. 3 Comparison of ease of access to reading software.
Self-reported additional benefits included off-site/after-hours reading (n = 7), hybrid-working
opportunities (n = 6), enabling virtual conferences (n = 4), and higher-definition
screens on personal laptops (n = 2). User-reported issues included a lack of access
to other hospital system while off-site (n = 3).
Discussion
The performance of the RRP in our center was excellent with 100% technical success
and excellent user acceptance. In particular, the improved access to RRP and its potential
to increase capacity were underscored by all readers.
Being RRP-enabled, remote reading could support the development of VCE services in
new centers and enable these and established smaller units to have access to the expertise
in remote tertiary referral VCE centers. Our experience to date suggests this is achievable
with this new system. While VCE in our center is not currently available as an after-hours
service, the adoption of the RRP could be used to develop a 7-day-a-week, 24-hour
emergency VCE service, which would be a significant diagnostic advantage, particularly
for patients with obscure gastrointestinal bleeding. If the use of this technology
were to spread globally, an international certification system for readers would be
necessary to support inter-institution or international reading, and governing bodies
should consider the creation of such awards in the future. In addition, a successful
RRP is a key component of any future near-patient or at-home colon capsule testing
service that may be in development and thought to improve patient acceptance.
The RRP-enabled system has improved access to our VCE quality improvement meeting,
allowing multiple readers to review and sign off on video clips and images from multiple
locations/institutions, while also allowing trainees the opportunity to participate
and have one-on-one feedback and supervision with expert readers. In the future, the
RRP could increase access to trainees not currently working at a VCE center to gain
skills and expertise in this technology and would have the potential to collect anonymized
image clips and reports from the cloud as teaching materials for both young trainees,
creating online learning and testing resources, and as input data to help train future
AI-assisted reading software.
Finally, this technology provides ease of access to the videos for readers. It has
the potential to promote a hybrid-working environment that will undoubtedly see wider
adoption of VCE among healthcare professionals in the coming years. This, coupled
with increased quality testing capacity at new centers, including near-patient services,
will further enhance the future role of VCE as a pivotal diagnostic test.
Conclusions
In summary, PillCam remote reader is a reliable, secure and effective capsule analysis
platform and should be incorporated into any capsule service development plan.
