Discovery: Pentax* Computer-aided Detection to Improve Adenoma Detection in a Real-time Setting - The Discovery II Study. A randomized clinical trial.

Colorectal cancer (CRC) is the third most commonly diagnosed malignancy and the
fourth leading cause of death in the Western world. CRC usually develops from
focal changes within benign, precancerous polyps. Colonoscopy aims to early
detect and remove these precancerous polyps. Although colonoscopy is generally
considered to be the most accurate screening modality, a substantial number of
polyps are still missed. Two meta-analysis showed pooled miss rates for polyps
of any size of 22-26%, an adenoma miss rate (AMR) of 26% for diminutive polyps
(1-5mm) and an AMR of 27% for serrated polyps. Missed lesions may have the
possibility to develop into cancer and it is thought that at least 50% of all
interval carcinomas (iCRCs; defined as a cancer diagnosed between screening and
post-screening surveillance colonoscopies) arise from missed lesions during
colonoscopy. In recent years, a new solution to human error in detecting polyps
has been developed; computer-aided detection (CADe) systems. CADe systems use
deep learning to improve polyp and adenoma detection in a more consistent and
reliable way. In the past years, several CADe systems have been developed.
Albeit performance of these systems on offline videos and images seems
promising, evidence on the ability during real-time clinical practice is
lacking. Recently Pentax Medical has developed a CADe system, named
*Discovery*. Pre-clinical studies have shown a 90% and 75.2% sensitivity and
specificity, respectively, with an area under curve (AUC) of 91% for polyp
detection (unpublished data). We hypothesize that the use of the Pentax
Discovery system is feasible and safe.

The primary objective of the study is to compare the ADR between PCSC and CC in
patients referred for diagnostic, screening (non-iFOBT based) or surveillance
colonoscopy.

Secondary objectives are the following:
- To compare the polyp detection rate (PDR);
- To compare the mean number of adenomas per patient detected;
- To compare the mean number of polyps per patient detected;
- To compare the number of sessile serrated lesions (SSLs) detected;
- To compare the number of advanced adenomas (i.e. adenomas >= 10 mm and/or
harbouring a villous component and/or containing high grade dysplasia (HGD))
detected;
- To compare detected colon lesions for size, location (per colon segment and
in relation to colonic folds), morphology (using the Paris classification) and
histopathological characteristics (using the Vienna classification);
- To compare the ADR of the first 20% of patients scoped by each endoscopist
with the last 20% of patients in each arm to identify any changes in ADR
throughout the trial (i.e. as part of a learning effect);
- To compare bowel cleansing levels using the Boston Bowel Preparation Scale
(BBPS);
- To compare procedure times for both techniques (i.e. total procedure time,
mean polypectomy time and withdrawal time in each colonic segment);
- To compare (severe) adverse events (S(AE)s) up to 30 days post-procedure;
- To compare post-colonoscopy surveillance intervals applying European and US
surveillance guidelines;
- To assess the number of false positives by the PCS (defined as the number of
times the system highlights an unsuspected area (as assessed by the
endoscopist) >3 seconds)
- To assess the reason why a notification by the PCS is assessed as false
positive by the endoscopist (i.e. due to bubbles, fecal material, etc.);
- To assess the number of false negatives by the PCS (defined as the number of
times the system does not highlight an unsuspected area (as assessed by the
endoscopist) <3 seconds)

Randomized, two arm colonoscopy trial, including 560 patients.

Patients will be enrolled for a period of 30 days, starting at the day of the
procedure and ending after 30 days of follow up. It is likely that the Pentax
CADe system will lead to the detection of more (adenomatous) polyps and thereby
resulting in more polypectomies, therefore participation in the study might
lead to a longer procedure time and more adverse events, especially the risk of
intraprocedural or delayed bleeding. Nonetheless, the risk of intraprocedural
or delayed bleeding is estimated to be low, i.e. 1.8% and <=0.1%, respectively.
Most bleedings can be treated during the same, or an additional colonoscopy,
procedure. The removal of additional polyps that are detected by the CADe
system might have a beneficial effect on the morbidity and mortality resulting
from colorectal carcinoma, depending on the type of polyp that is removed
during the procedure. The follow-up of the procedure (e.g. the number of
hospital visits) will take place according to local guidelines.