Development and Validation of a Deep Learning Algorithm for Bowel Preparation Quality Scoring

The purpose of this study is to develop and validate the performance of an artificial intelligence(AI) assisted Boston Bowel preparation Scoring(BBPS) system for evaluation of bowel cleanness, then testify whether this new scoring system can help physicians to improve the quality control parameters of colonoscopy in clinic practice.

Study Overview

• Status: Unknown
• Conditions: Bowel Preparation
• Intervention / Treatment: Device: Artificial intelligence assisted bowel preparation quality scoring system; Device: Conventional human scoring

Detailed Description

Colonoscopy is recommended as a routine examination for colorectal cancer screening. Adequate bowel preparation is indispensable to ensure a clear vision of colonic mucosa，complete inspection of all colon segments, and furthermore improves the detection rates of small adenomas. Thus, the adequacy of bowel preparation should be accurately evaluated and documented. However, the accuracy of current bowel preparation quality scales greatly relies on intra-observer and inter-observer consistency for lack of objective measurements. Recently, deep learning based on central neural networks (CNN) has shown multiple potential in computer-aided detection and computer-aided diagnose of gastrointestinal lesions. While, no studies have been conducted to evaluate the performance of deep learning algorithm in bowel preparation quality scoring. This study aims to train an algorithm to assess bowel preparation quality using the BBPS, and testify whether the engagement of AI can improve the quality control parameters of colonoscopy.

• Study Type: Interventional
• Enrollment (Anticipated): 100
• Phase: Not Applicable

Contacts and Locations

Study Locations

• China
  • Shandong
    • Jinan, Shandong, China, 257000
      • Qilu hosipital

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: 18 years to 70 years (Adult, Older Adult)
• Accepts Healthy Volunteers: No
• Genders Eligible for Study: All

Description

Inclusion Criteria:

• Patients aged 18-70 years undergoing afternoon colonoscopy

Exclusion Criteria:

• Known or suspected bowel obstruction, stricture or perforation
• Compromised swallowing reflex or mental status
• Severe chronic renal failure(creatinine clearance < 30 ml/min)
• Severe congestive heart failure (New York Heart Association class III or IV)
• Uncontrolled hypertension (systolic blood pressure > 170 mm Hg, diastolic blood pressure > 100 mm Hg)
• Dehydration
• Disturbance of electrolytes
• Pregnancy or lactation
• Hemodynamically unstable
• Unable to give informed consent

Study Plan

How is the study designed?

Design Details

• Primary Purpose: Health Services Research
• Allocation: Randomized
• Interventional Model: Parallel Assignment
• Masking: Single

Number of Arms

2

Arms and Interventions

Participant Group / Arm	Intervention / Treatment
Experimental: Artificial Intelligence assisted Scoring Group; Patients in this group go through colonoscopy under the AI monitoring device.	Device: Artificial intelligence assisted bowel preparation quality scoring system; After receiving standard bowel preparation regimen, patients go through colonoscopy under the AI monitoring device. During the withdrawal process, bowel preparation quality is monitored by AI-associated scoring system. Whenever a sub-score below 2 points is detected, endoscopist will be alarmed up to three times to wash and suck the colonic contents. Videos will be recorded and re-evaluated by experts to determine the final BBPS score. The withdrawal time is targeted at least 6min in accordance with colonoscopy quality practice. All detected polyps will be removed and obtained for histological assessment, with the possible exception of diminutive(less than 5mm) rectal polyps.
Active Comparator: Conventional Human Scoring Group; Patients in this group go through conventional colonoscopy without AI monitoring device.	Device: Conventional human scoring; After receiving standard bowel preparation regimen, patients go through conventional colonoscopy without the AI monitoring device. During the withdrawal process, after washing and sucking the colonic contents according to endoscopist's personal experience, bowel preparation quality is evaluated by human. Videos will be recorded and re-evaluated by experts to determine the final BBPS score. The withdrawal time is targeted at least 6min in accordance with colonoscopy quality practice. All detected polyps will be removed and obtained for histological assessment, with the possible exception of diminutive(less than 5mm) rectal polyps.

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
The rate of patients achieving adequate bowel preparation in each group.	Bowel preparation quality was measured by BBPS. After fully washing or suctioning of colonic contents, three segments including right colon (containing cecum and ascending colon), transvers colon (containing hepatic and splenic flexures) and left colon (containing descending and sigmoid colon) were individually scored from 0 to 3. Point 0 refers to unprepared colon segment with obscured solid stool making mucosa cannot be seen; Point 1 refers to part of mucosa can be seen, but some areas are covered by staining, residual stool, and/or opaque liquid; Point 2 refers to entire mucosa is well-seen; Point 3 refers to clean colon segment without staining, fecal materials or liquids. A sub-score of each colon segment was used, ranging from minimum 0 to maximum 3. The highest score means the excellent bowel preparation. Adequate bowel preparation was defined as a total BBPS≥6 and sub-BBPS≥2 per segment.	6 months

Secondary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Adenoma Detection Rate	The proportion of patients from whom at least one adenoma can be detected.	6 months
Polyp Detection Rate	The proportion of patients from whom at least one polyp can be detected.	6 months

Collaborators and Investigators

• Sponsor: Shandong University

Study record dates

Study Major Dates

• Study Start (Actual): December 15, 2018
• Primary Completion (Anticipated): December 15, 2019
• Study Completion (Anticipated): April 15, 2020

Study Registration Dates

• First Submitted: April 8, 2019
• First Submitted That Met QC Criteria: April 8, 2019
• First Posted (Actual): April 9, 2019

Study Record Updates

• Last Update Posted (Actual): April 9, 2019
• Last Update Submitted That Met QC Criteria: April 8, 2019
• Last Verified: April 1, 2019

More Information

Terms related to this study

• Keywords: Bowel Preparation, Deep Learning, Central Neural Networks
• Other Study ID Numbers: 2019SDU-QILU-G001

Drug and device information, study documents

• Studies a U.S. FDA-regulated drug product: No
• Studies a U.S. FDA-regulated device product: No