Evaluation of Obscure Gastrointestinal Bleeding Patients With Conventional Capsule Endoscopy and Panoramic Side View Capsule Endoscopy

Materials and Methods Patients 30 Patients with occult gastrointestinal bleeding aged 20-85 years old will be enrolled from Endoscopy Center for Diagnosis and Treatment, Taipei Veterans General Hospital. Patients needed to have negative esophagogastroduodenoscopy and colonoscopy for potential bleeding source within 12 months prior to the capsule examination. Patient with suspected or confirmed gastrointestinal tract obstruction, pacemaker use or unwilling to swallow the capsule endoscope will be excluded. Female at child bearing age will receive urine HCG test for pregnancy evaluation. Pregnant patients will be excluded. All patients will be given informed consent before inclusion.

Capsule Endoscopy procedure The Endocapsule (EC-10) and CapsoCam Plus (Capsovision) capsule will be used. The Endocapsule is 26mm x 11mm in size and 3.3g in weight. It has 160° angle of view with 12 hours of battery life, 2 frame per second sampling rate. One antenna unit will be packed on the patient's waist during examination adjacent to a recorder. [17] The CapsoCam Plus capsule is 31mm x 11mm in size and 4g in weight. It has a field of view of 360° with at least 15 hours of battery life, takes 20 images per second maximum, (4 frames per second from each of the four cameras). No other accessories are needed during the procedure.

Small-bowel preparation will be performed in patients before capsule ingestion as routine practice and previously published guidelines. [19] The patients will take one pack of GI-Klean powder(polyethylene glycol) in 1000ml water followed by 1000ml water at 6-8 pm the night before capsule endoscopy. Then they will take another pack of GI-Klean powder(polyethylene glycol) in 1000ml water followed by 1000ml water at 5AM on the morning on the day of capsule endoscopy.

On the day of examination, the patients will take 10ml of simethicone fluid first. Then they'll ingest the two capsules, with 1.5 hour between and in a randomized order (using sealed envelopes). After capsule endoscopy entered small bowel, the patient will take a light meal at least 4 hours after the 2nd capsule endoscope ingestion. They'll take dinner as usual. The capsule endoscope will be collected by the patient using a specified capsule collecting system during defecation. If no capsule endoscope was noted in the stool in 3 days. The patient will received through 20mg sennosides on the 3rd night. The patients will return back the capsule endoscope, antenna and recorder after getting the capsule endoscope. For patient who didn't collect capsule endoscope in 15 days, abdominal x-ray plain film will be taken to evaluate the status of capsule endoscope.

Capsule endoscopy review For each case, the two capsule films will read in a randomized order by three experienced physicians (one >250 VCE examinations, two >100 VCE examinations). None of the physicians had experience of panoramic side view capsule endoscopy(<5) before this study. The two readings will be done blindly. The reading frames will be were available to each reader, and suggested not faster than a rate of 15 images per second. The cases will be classified into concordant positive, concordant negative, and discordant cases. For discordant cases, a third reviewer will conduct the clarification of the CE images for confirmation and confirmed during a conference by all three experienced readers. An image per image comparison will be performed and the complete film will be read a second time when necessary.

The primary outcome of the study will be (1) ampulla of the vater detection rate and (2) the concordance between the two capsule examinations with a kappa value>0.5. The secondary outcomes will be the small bowel cleansing quality, technical failure rate, transit and operation time, capsule endoscopy completion rate, video record reliability, diagnostic yield, safety profile and the physicians' and patients' satisfaction with the capsule endoscopy.

The definition of the outcome measurement will be defined as following:

1. Ampulla of Vater detection rate: the number of examinations in which the ampulla was clearly identified over the number of the overall number of patients receiving VCE
2. Data interpretation: All endoscopic findings will be systemically recorded according to location, endoscopic features and clinical significance, and in relation to procedure indications, as P0 (low probability), P1 (intermediate pcccrobability), or P2 (high probability) as in previous descriptions [18]. Nonsignificant (P0) lesions and images located outside the small intestine were not considered. For per patient analysis, the final diagnosis of P1 or P2 lesions will be used according to the description of the most important or the most relevant lesion or group of lesions. The sensitivity of each capsule examination will be assessed, with true-positive cases being calculated as the sum of positive cases obtained from the Endocapsule 10 or CapsoCam plus readings, including expert review of discordant cases. In a per lesion analysis, each lesion (up to three main lesions found from capsule reading) or group of lesions will be analyzed independently.
3. Diagnostic Yield: VCE will be defined as 'positive' when at least one clinically significant finding (P2) was identified; the rate of "positive" VCE (ie, DY) will be calculated.
4. Small bowel cleansing quality: The small-bowel cleansing will be scored by estimating the amount of clearly visible mucosa using the scale proposed by Esaki et al. [19]. The images of the small intestine will equally be separated into 3 segments according to SBTT. The cleansing quality will be graded according to fluid transparency and mucosal invisibility. The fluid transparency will be graded into grade 1, Clear fluid without obscuring vision, grade 2 Slightly dark fluid minimally obscuring vision, grade 3 Opaque fluid partly obscuring vision and grade 4 Turbid fluid severely obscuring vision. The mucosal invisibility will be graded as Grade 1, <5% in duration of 50% bubbles or residues, grade 2 5%-15%, grade 3 15%-25% and grade 4 >25%. The overall image quality will be categorized into Grade A 3-5, Grade B 6-8 and Grade C 9-12.
5. Technical failure: it will be defined as any encountered technical problem due to one or more system components (such as capsule device, downloading system, and/or workstation) influencing the generation of the video.
6. Transit and operating times: Gastric transit time, will be defined as the time between the first and the last gastric image. Small-bowel transit time (SBTT), will be defined as the timeframe between the first and the last small-bowel image. Total capsule endoscope time will be defined as the time between the first and the last image captured by the capsule. Reading time will be defined as the time doctors read the full capsule endoscopic images.
7. Capsule endoscopy completion rate: the number of examinations in which the colon was reached during the operating time over the overall number of patients receiving VCE.
8. Video record reliability: the video record reliability will be evaluated according to a quantitative scoring system as described before, the images of the small intestine will equally be separated into 3 segments according to SBTT. The scoring system is defined as following: 3 points - excellent: when less than 5 total frames and no consecutive non-evaluable frames were observed. 2 points - good: when ≥5 to <10 total frames or one sequence with <5 consecutive non-evaluable frames were observed. 1 point - fair: when ≥10 to <15 total frames, or one sequence with <10 consecutive non-evaluable frames were observed. 0 points - poor: when ≥15 total frames, or multiple sequences, or one sequence with ≥10 consecutive non-evaluable frames were observed.[13] The video record reliability will be evaluated only in completed capsule endoscopy cases.
9. Safety profile: As reported in previous studies on front-viewing capsules [1,21], we will focus our attention on capsule aspiration, at time of ingestion, and capsule retention. Capsule retention will be defined as the persistence of the capsule within the patient's intestinal tract, evaluated by an abdominal x-ray plain film, 15 days after capsule ingestion. There is 1 % chance of capsule endoscope retention in gastrointestinal tract and patients in whom the capsule was retrieved by elective surgery and/or by endoscopic procedures will also be considered as capsule retention. Any other incident and/or adverse event, reported as potentially related to VCE by the local investigator, will be documented and classified according to previously published international guidelines.[20]
10. Patients' Satisfaction Questionnaire: For evaluation of the convenience and satisfaction of the patient about two capsule endoscopy, the patient will be asked to fill a questionnaire about the capsule endoscopy procedure after the capsule endoscopy procedure.
11. Physician's satisfaction evaluation: For evaluation and further recommendation of the physician on two different capsule endoscopy about their convenience and efficacy, physician will be asked to fill up a satisfaction sheet after capsule endoscopy procedure and image analysis.
12. Statistical analysis In a recent study, the detection rate of ampulla vater is 81.7 % in panoramic view capsule endoscopy and 9.6% in conventional capsule endoscopy. 18 patients are needed to give the study 80% power atα=0.05%.[21] For evaluation of interobserver agreement, the calculated sample size for detection of was 27 patients in order to obtain a satisfying estimation of concordance, with an 80% power, of the kappa interobserver agreement coefficient between paired (Endocapsule 10(Olympus) and CapsoCam Plus(Capsovision)) capsule readings, taking into account the positive diagnostic yield of VCE in OGIB reaches about 60%. With an estimation of 10% dropout rate, 30 patients will be needed for this study.

The median with inter-quartile range (IQR) and range will be provided for non-normally distributed variables, whereas the mean ± standard deviation (SD) for normally distributed variables. The paired T test will be used to compare technical outcomes among two capsule endoscopies (ie, operative time and video records reliability). The 2-tailed Fisher exact test will be used to compare the diagnostic yield. To compare the sensitivity of Endocapsule 10and CapsoCam Plus, a Mac-Nemar test will be used with the exact test of the binomial law.

For statistical analysis, SPSS software version 26.0 (SPSS Inc., Chicago, Illinois, USA) in a Windows XP (Microsoft, Seattle, Washington,USA) environment will be used. A P value of <0.05 was considered significant.