Colonoscopy Quality Research Articles

Quality in colonoscopy: time to ensure national standards are implemented?

BackgroundHigh-quality colonoscopy is crucial to ensure complete mucosal visualisation and to maximise detection of pathology. Previous audits showing variable quality have prompted national and international colonoscopy improvement programmes, including the...

The Efficacy of AI-Assisted CADe Systems in Colorectal Polyps and Adenomas during Colonoscopies: A Systematic Review

Colorectal polyps and adenomas are a significant health concern that can lead to colorectal cancer. Detecting polyps and adenomas early is important to remove them in a timely manner, reducing the risk of cancer. Artificial intelligence (AI) could assist in identifying polyps and adenomas more efficiently. This study will investigate the potential benefits of AI technology in detecting colorectal adenomas, using AI-assisted colonoscopy diagnostic systems. The goal is to determine if there are differences in detection rates using AI compared to standard colonoscopy methods. PRISMA Statement 2020 guidelines were adhered to in this systematic review. Three databases were searched including PubMed/MEDLINE, Embase and Web of Science through February 23, 2023. A combination of the following keywords was used: AI, Artificial, Intelligence, Colorectal, Polyp, Adenoma. The adenoma detection rate (ADR) and polyp detection rate (PDR) were additionally meta-analyzed for proportions, and reported as odds ratio (OR), applying 95% confidence intervals (CI). A total of 9 randomized clinical trials were included in this review, pooling in 6981 participants. The findings were presented systematically, based on the design, technology used, effects on ADR and PDR, and key conclusions. The meta-analytical findings were reported as follows. On applying a random-effects model, the odds ratio (OR) for ADR was 1.532 (95% CI=1.125-2.087), with significant detection favoring CADe systems (P=0.0068). Similarly, when assessing the chance of diagnosing colorectal polyps with CADe systems as an add-on compared to standard colonoscopy alone, the odds ratio (OR) was 1.893 (95% CI=1.687-2.125, P<0.0001). This systematic review found that real-time computer-aided detection during colonoscopy improved the detection of adenomas and polyps. The use of the system is feasible and safe, and future studies should focus on improving colonoscopy quality indicators including the most widely used– ADR. Larger randomized controlled studies are needed to evaluate the impact of this system on adenoma and serrated polyp detection and its effect on endoscopists with lower detection rates.

High quality colonoscopy: using textbook process as a composite quality measure.

BACKGROUND : High quality colonoscopy is fundamental to good patient outcomes. "Textbook outcome" has proven to be a feasible multidimensional measure for quality assurance between surgical centers. In this study, we sought to establish the "textbook process" (TP) as a new composite measure for the optimal colonoscopy process and assessed how frequently TP was attained in clinical practice and the variation in TP between endoscopists. METHODS : To reach consensus on the definition of TP, international expert endoscopists completed a modified Delphi consensus process. The achievement of TP was then applied to clinical practice. Prospectively collected data in two endoscopy services were retrospectively evaluated. Data on colonoscopies performed for symptoms or surveillance between 1 January 2018 and 1 August 2021 were analyzed. RESULTS : The Delphi consensus process was completed by 20 of 27 invited experts (74.1 %). TP was defined as a colonoscopy fulfilling the following items: explicit colonoscopy indication; successful cecal intubation; adequate bowel preparation; adequate withdrawal time; acceptable patient comfort score; provision of post-polypectomy surveillance recommendations in line with guidelines; and the absence of the use of reversal agents, early adverse events, readmission, and mortality. In the two endoscopy services studied, TP was achieved in 5962/8227 colonoscopies (72.5 %). Of 48 endoscopists performing colonoscopy, attainment of TP varied significantly, ranging per endoscopist from 41.0 % to 89.1 %. CONCLUSION : This study proposes a new composite measure for colonoscopy, namely "textbook process." TP gives a comprehensive summary of performance and demonstrates significant variation between endoscopists, illustrating the potential benefit of TP as a measure in future quality assessment programs.

Gastroenterologist focus of clinical practice affects adenoma detection in screening colonoscopy.

Our objective was to determine whether the clinical focus of gastroenterology practice would affect screening colonoscopy quality metrics, specifically adenoma detection (AD). In a retrospective study of screening colonoscopies, gastroenterologists were categorized based on their clinical subspecialty focus into general/motility, hepatology, inflammatory bowel disease (IBD), and interventional endoscopy. The primary outcome was AD with a secondary outcome of adenoma and/or sessile serrated polyp (SSP) detection (AD + SSP). A total of 5271 (male: 49.1%) complete colonoscopies were performed between 2010 and 2020 by 16 gastroenterologists (male: 62.5%, general/motility specialists: 3, hepatologists: 3, IBD specialists: 4, interventional endoscopists: 6). The AD and AD + SSP rate between each specialty focus were 27.5% and 31.0% for general/motility, 31.4% and 35.5% for hepatology, 38.4% and 43.6% for IBD, and 37.5% and 43.2% for interventional endoscopy. In regression analysis, patient's male gender (odds ratios [OR]: 1.81, 95% CI: 1.60-2.05, P < .001), longer withdrawal time (OR: 1.16, 95% CI: 1.14-1.18, P < .001), hepatologist (OR: 1.25, 95% CI: 1.02-1.53, P = .029), IBD subspecialist (OR: 1.60, 95% CI: 1.30-1.98, P < .001), and interventional endoscopist (OR: 1.36, 95% CI: 1.13-1.64, P < .001) were independently associated with AD. Moreover, patient's male gender (OR: 1.64, 95% CI: 1.45-1.85, P < .001), acceptable bowel preparation (OR: 1.29, 95% CI: 1.06-1.56, P = .010), withdrawal time (1.20, 95% CI: 1.18-1.22, P < .001), hepatologist (OR: 1.30, 95% CI: 1.07-1.59, P = .008), IBD subspecialist (OR: 1.72, 95% CI: 1.39-2.12, P < .001), interventional endoscopist (OR: 1.44, 95% CI: 1.20-1.72, P < .001) were independent factors that improved detection of AD + SSP. Subspecialty focus of practice was an important factor in AD rate along with the male gender of the patient, bowel preparation, and withdrawal time.

Unlocking quality in endoscopic mucosal resection.

A review of the development of the key performance metrics of endoscopic mucosal resection (EMR), learning from the experience of the establishment of widespread colonoscopy quality measurements. Potential future performance markers for both colonoscopy and EMR are also evaluated to ensure continued high quality performance is maintained with a focus service framework and predictors of patient outcome.

Evaluating the Safety and Quality of Diagnostic Colonoscopies Performed by General Surgeons: A Retrospective Study.

Introduction Colonoscopy, which is a challenging procedure and requires adequate time to master the skill, is the procedure of choice to visualize colonic mucosa to rule out many colonic pathologies. There is a dearth of published information from real clinical experiences regarding successful procedures and limitations. The end point of colonoscopy is the visualization of the cecal pole by intubating the cecum. Many Europeans and English health organizations recommend that the procedure should have a completion rate of around or above 90%. Gut preparation is an important determinant for a successful procedure and obviates the need for further invasive and/or expensive procedures such as imaging. The majority of colonoscopies are being performed by gastroenterologists (GI) throughout the world, and the role of a surgeon as an endoscopist is in debate. Before this study, neither a retrospective nor a prospective evaluation of the general surgeon's (GS) endoscopy's quality and safety had been evaluatedin our institution. Material and method This retrospective observational study was carried out from 1January2022 to 31August2022 in the Department of Surgery at Mayo Hospital, Lahore, to evaluate colonoscopy completion rates, reason for failure, and complications in terms of bleeding and perforation. All patients undergoing lower gastrointestinal endoscopy (LGiE), both elective and emergency, were included. Patients under 15 years of age and patients known to be hepatitis B-positive or hepatitis C-positive were excluded from the study. All relevant data were entered into a data sheet. Qualitative variables such as gender, cecal intubation, adjusted cecal intubation, gut preparation, reasons for failed colonoscopy, analgesia use, and complications (bleeding and perforation) were calculated as frequency and percentage. Quantitative data such as age and pain score were reported as mean and standard deviation (SD). Details obtained were tabulated and analyzed via the Statistical Package for Social Sciences (SPSS) version 29.0 (IBM SPSS Statistics, Armonk, NY). Results A total of 57 patientdata were collected; 35.1% (n=20) were female, and 64.9% (n=37) were males. The cecal intubation rate (CIR) was 49.1% (n=28), and the adjusted rate was 71.9%, excluding incompleteness due to mass obstructing lumen, 8.8% (n=5); planned left colonoscopy, 7% (n=4); sigmoidoscopy, 3.5% (n=2); distal stoma scope, 1.8% (n=1); and colonic stricture, 1.8% (n=1). The prevalent reason for failed colonoscopy was inadequate gut preparation (15.8% {n=9}). Other reasons include patientdiscomfort, 3.5% (n=2); looping of scope, 7% (n=4); and acute colonic angulation, 1.8% (n=1). No complications were recorded. Conclusion This study shows that colonoscopy can be done by general surgeons safely and effectively with adequate training. High rates of cecal intubation emerge during colonoscopies performed under deep sedation and by skilled colonoscopists. Adequate bowel preparatory regimen is compulsory for quality procedure.

Endoscopic surveillance after resection of sessile serrated lesions: so far, so good?

Polyp detection is on the rise! Since the establishment of the adenoma detection rate (ADR) as the primary quality parameter in colonoscopy [1], we have been focusing on improving our ability to detect colorectal adenomas, both through the refinement of our techniques and the growing availability and use of technological means that enhance detection capabilities. This increased awareness of the importance of meticulous mucosal inspection and the incorporation of quality assurance programs, including benchmarking, auditing, and retraining, have led to an increase in mean ADR across endoscopists and have played a decisive role in the global improvement of colonoscopy quality. However, we have also learned that not all that glitters is gold – ADR is not perfect! As a surrogate rather than direct measure of colonoscopy efficacy, it is suspectable to gaming (“one adenoma and done”) and may not reflect the colorectal cancer (CRC) risk for the individual patient or their need for surveillance colonoscopy. Indeed, up to 30 % of all CRCs appear to arise from the serrated neoplasia pathway rather than the conventional adenoma–carcinoma sequence. This raises doubts about the protective effects of a high ADR among this relatively large proportion of patients.

Evaluation of the Effectiveness of Patient’s Position Change during the Withdrawal Phase of Colonoscopy from Left Lateral Decubitus to Supine Position on Increasing Adenoma Detection Rate: A Randomized Controlled Trial

Background: Improvement in adenoma detection rate (ADR) reduces colorectal cancer incidence by increasing the colonoscopy quality. Using dynamic patient position changes during the withdrawal phase has shown promise in increasing ADR. We conducted this study to assess the effectiveness of the supine position on the improvement of ADR to improve its feasibility and avoid frequent patient position changes, particularly in sedated patients. Methods: This was a randomized, single-blind, parallel-group, single-center study implemented in the Mehregan private in Babol. Inclusion criteria were 40 to 85 years old, 4 L application of polyethylene glycol from the day before the procedure, no history of inflammatory bowel disease, bowel surgery, musculoskeletal problems, and negative familial history of colorectal cancer. Patients were allocated in a 1:1 ratio to the supine or left lateral positions during the withdrawal phase. All colonoscopies were performed by a single physician using a Fujifilm colonoscope. A P-value of &lt;0.005 was considered statistically significant. Results: A total of 880 patients were assessed for eligibility, of which 472 patients were included in the final analysis; 53.4% were female, the mean age of participants was 55.86±10.30 years old, 95.1% of patients had adequate bowel preparation, and adenomatous polyps were the most common histopathologic type (63.7%). Despite the intervention group’s higher rate of ADR and PDR (19.5% vs. 17.7% for ADR and 27.2% vs. 26.5% for PDR), no statistically significant difference in ADR or PDR was detected (P=0.613 and 0.866, respectively). Conclusion: No statistical significance was observed despite the increase in ADR when the supine position was used exclusively during the withdrawal phase. As a result, we recommend that the dynamic position change method be used if a position change is required. Nonetheless, additional research is required to determine a more effective alternative to dynamic position change in obese or heavily sedated patients.

Image acquisition as novel colonoscopic quality indicator: a single-center retrospective study.

In order to reduce the incidence and mortality of colorectal cancer, improving the quality of colonoscopy is the top priority. At present, the adenoma detection rate is the most used index to evaluate the quality of colonoscopy. So, we further verified the relevant factors influencing the quality of colonoscopy and found out the novel quality indicators by studying the relationship between the influencing factors and the adenoma detection rate. The study included 3824 cases of colonoscopy from January to December 2020. We retrospectively recorded the age and sex of the subjects; the number, size, and histological features of lesions; withdrawal time and the number of images acquired during colonoscopy. We analyzed the associated factors affecting adenoma and polyp detection, and verified their effectiveness with both univariate and multivariate logistic regression analyses. Logistic regression analyses showed that gender, age, withdrawal time and the number of images acquired during colonoscopy could serve as independent predictors of adenoma/polyp detection rate. In addition, adenoma detection rate (25.36% vs. 14.29%) and polyp detection rate (53.99% vs. 34.42%) showed a marked increase when the number of images taken during colonoscopy was ≥29 (P<0.001). Gender, age, withdrawal time and the number of images acquired during colonoscopy are influencing factors for the detection of colorectal adenomas and polyps. And we can gain higher adenoma/polyp detection rate when endoscopists capture more colonoscopic images.

Automated Detection of Anatomical Landmarks During Colonoscopy Using a Deep Learning Model.

Identification and photo-documentation of the ileocecal valve (ICV) and appendiceal orifice (AO) confirm completeness of colonoscopy examinations. We aimed to develop and test a deep convolutional neural network (DCNN) model that can automatically identify ICV and AO, and differentiate these landmarks from normal mucosa and colorectal polyps. We prospectively collected annotated full-length colonoscopy videos of 318 patients undergoing outpatient colonoscopies. We created three nonoverlapping training, validation, and test data sets with 25,444 unaltered frames extracted from the colonoscopy videos showing four landmarks/image classes (AO, ICV, normal mucosa, and polyps). A DCNN classification model was developed, validated, and tested in separate data sets of images containing the four different landmarks. After training and validation, the DCNN model could identify both AO and ICV in 18 out of 21 patients (85.7%). The accuracy of the model for differentiating AO from normal mucosa, and ICV from normal mucosa were 86.4% (95% CI 84.1% to 88.5%), and 86.4% (95% CI 84.1% to 88.6%), respectively. Furthermore, the accuracy of the model for differentiating polyps from normal mucosa was 88.6% (95% CI 86.6% to 90.3%). This model offers a novel tool to assist endoscopists with automated identification of AO and ICV during colonoscopy. The model can reliably distinguish these anatomical landmarks from normal mucosa and colorectal polyps. It can be implemented into automated colonoscopy report generation, photo-documentation, and quality auditing solutions to improve colonoscopy reporting quality.

Editorial: Optimizing the success of cold snare polypectomy in colonoscopy practice

Editorial: Optimizing the success of cold snare polypectomy in colonoscopy practice

Colorectal cancer and advanced adenoma incidence during post-polypectomy surveillance: a national cohort study in the English Bowel Cancer Screening Programme.

Improved colonoscopy quality has led to debate about whether all post-polypectomy surveillance is justified. We evaluated surveillance within the English Bowel Cancer Screening Programme (BCSP) to determine the yield of surveillance and identify predictive factors for surveillance outcome. We performed a retrospective cohort study of individuals undergoing post-polypectomy surveillance between July 2006 and January 2017. BCSP records were linked to the National Cancer Registration Database to identify interval-type post-colonoscopy colorectal cancers (CRCs). Advanced adenoma and CRC at surveillance were documented. CRC incidence was compared with the general population using standardized incidence ratios (SIRs). Predictors of advanced adenomas at first surveillance (S1), and CRC during follow-up, were identified. 44 151 individuals (23 078 intermediate risk; 21 073 high risk) underwent 64 544 surveillance episodes. Advanced adenoma and CRC yields were, respectively, 10.0 % and 0.5 % at S1, 8.5 % and 0.4 % at S2, and 10.8 % and 0.4 % at S3. S1 yield was lowest in those with one index adenoma ≥ 10 mm (advanced adenoma 6.1 %; CRC 0.3 %). The SIR was 0.76 (95 %CI 0.66-0.88), accounted for by the intermediate risk group (intermediate risk SIR 0.61, 95 %CI 0.49-0.75; high risk SIR 0.95, 95 %CI 0.79-1.15). Adenoma multiplicity, presence of a large nonpedunculated adenoma, and greater villous component were associated with advanced adenoma at S1. Older age and multiplicity were significantly associated with CRC risk. This large, national analysis found low levels of CRC in those undergoing surveillance and low advanced adenoma yield in most subgroups. Less intensive surveillance in some subgroups is warranted, and surveillance may be avoided in those with a single large adenoma.

Artificial Intelligence Applied to Colonoscopy: Is It Time to Take a Step Forward?

Growing evidence indicates that artificial intelligence (AI) applied to medicine is here to stay. In gastroenterology, AI computer vision applications have been stated as a research priority. The two main AI system categories are computer-aided polyp detection (CADe) and computer-assisted diagnosis (CADx). However, other fields of expansion are those related to colonoscopy quality, such as methods to objectively assess colon cleansing during the colonoscopy, as well as devices to automatically predict and improve bowel cleansing before the examination, predict deep submucosal invasion, obtain a reliable measurement of colorectal polyps and accurately locate colorectal lesions in the colon. Although growing evidence indicates that AI systems could improve some of these quality metrics, there are concerns regarding cost-effectiveness, and large and multicentric randomized studies with strong outcomes, such as post-colonoscopy colorectal cancer incidence and mortality, are lacking. The integration of all these tasks into one quality-improvement device could facilitate the incorporation of AI systems in clinical practice. In this manuscript, the current status of the role of AI in colonoscopy is reviewed, as well as its current applications, drawbacks and areas for improvement.

Mortality After Postcolonoscopy Colorectal Cancer in the Veterans Affairs Health Care System

Postcolonoscopy colorectal cancer (PCCRC) refers to colorectal cancer (CRC) diagnosed after a colonoscopy in which no cancer was found and is reflective of colonoscopy quality at the individual and system levels. Colonoscopy is widely performed in the Veterans Affairs (VA) health care system, but the prevalence of PCCRC and its associated mortality are unknown. To examine PCCRC prevalence and its all-cause mortality (ACM) and CRC-specific mortality (CSM) within the VA health care system. This retrospective cohort study used VA-Medicare administrative data to identify 29 877 veterans aged 50 to 85 years with newly diagnosed CRC between January 1, 2003, and December 31, 2013. Patients whose colonoscopy occurred less than 6 months before CRC diagnosis with no other colonoscopy within the previous 36 months were categorized as having detected CRC (DCRC). Those who had a colonoscopy that did not detect CRC between 6 and 36 months before CRC diagnosis were categorized as having postcolonoscopy CRC (PCCRC-3y). A third group included patients with CRC and no colonoscopy within the prior 36 months. The final analysis of the data was performed in September 2022. Prior receipt of colonoscopy. Cox proportional hazards regression (with censoring, last follow-up December 31, 2018) analyses were conducted to compare PCCRC-3y and DCRC for 5-year ACM and CSM after CRC diagnosis. Of 29 877 patients with CRC (median [IQR] age, 67 [60-75] years; 29 353 [98%] male; 5284 [18%] Black, 23 971 [80%] White, and 622 [2%] other), 1785 (6%) were classified as having PCCRC-3y and 21 811 (73%) as having DCRC. The 5-year ACM rates were 46% vs 42% for patients with PCCRC-3y vs patients with DCRC. The 5-year CSM rates were 26% vs 25% for patients with PCCRC-3y vs patients with DCRC. In multivariable Cox proportional hazards regression analysis, there was no significant difference in ACM and CSM between patients with PCCRC-3y (adjusted hazard ratio [aHR], 1.04; 95% CI, 0.98-1.11; P = .18) and patients with DCRC (aHR, 1.04; 95% CI, 0.95-1.13; P = .42). However, compared with patients with DCRC, patients with no prior colonoscopy had significantly higher ACM (aHR, 1.76; 95% CI, 1.70-1.82; P < .001) and CSM (aHR, 2.22; 95% CI, 2.12-2.32; P < .001). Compared with patients with DCRC, patients with PCCRC-3y had significantly lower odds of having undergone colonoscopy performed by a gastroenterologist (odds ratio, 0.48; 95% CI, 0.43-0.53; P < .001). This study found that PCCRC-3y constituted 6% of CRCs in the VA system, which is similar to other settings. Compared with patients with CRC detected by colonoscopy, those with PCCRC-3y have comparable ACM and CSM.

A study to assess the Nurses knowledge &amp; compliance to the protocol of adequate bowel preparation for colonoscopy

Introduction: Cleansing the colon before a colonoscopy is called bowel preparation, or “prep.” It involves taking medication that causes frequent, loose bowel movements to empty the colon. The medication is taken by mouth and typically comes in liquid form. Bowel cleansing is one of the most important parameters included in the evaluation of colonoscopy quality. The available evidence suggests that inadequate bowel preparation reduces the diagnostic yield of colorectal neoplasia and increases post-colonoscopy colorectal cancer risk. Nowadays, up to 30% of patients referred for colonoscopy have poor bowel cleansing. Recently, several studies have analyzed the risk factors for inadequate bowel cleansing as well as the strategies to optimize bowel preparation. Poor bowel preparation has been shown to be associated with low lower-quality dilators of colonoscopy performance. Patients with Personal History of Inadequate Bowel Preparation, Diabetes Mellitus, Chronic Constipation, and a History of Colorectal Resection are at risk of inadequate bowel preparation for colonoscopy. Methodology: In November 2022, there was an increase in consultant complaints about inadequate Bowel preparation for the patient in the ward &amp; ICU which impacted on proper diagnosis of the actual reason the patient took admitted at Apollo Hospitals, Bhubaneswar. So on basis of the complaints, a quasi-experimental group pre-test was conducted &amp; found the less compliance. To look forward to checking the compliances, training &amp; educational materials were prepared on basis of protocol of adequate bowel preparation for colonoscopy (education materials contained agents, methods (single/split) of preparation, dietary patterns, stool consistency assessment) &amp; started implement from December 2022 onwards in both one to one &amp; group teaching. And by end of February 2023, a post-test was conducted to assess the improvement of knowledge of Nurses. To collect the data of their practices on adequate bowel preparation a structured PIP (performance Improvement Project) tool was used &amp; total sample took 100 assigned Register Nurses who all were assigned the In-patients for Colonoscopy. Results: With constant reinforcement, educational materials on adequate Bowel preparation of Colonoscopy &amp; knowledge assessment of assigned nurses made an impact on the effective intervention in adequate Bowel preparation. The impact of compliance to assess the knowledge &amp; interventions of nurses in adherence to the protocol of Adequate Bowel preparation for colonoscopy. The knowledge of Nurses found less compliance in the Pre-Test which was conducted after consultant’s complaints in the month of November 2022 with 55.83% &amp; then after training &amp; educational materials were prepared on basis of protocol of adequate bowel preparation for colonoscopy {education materials contained agents, methods (single/split) of preparation, dietary patterns, stool consistency assessment)} &amp; started implement from December 2022 onwards in both one to one &amp; group teaching. And by end of February 2023, a post-test was conducted to assess the improvement of knowledge of Nurses &amp; found compliance percentage of 78.33%. And Compliance with adherence to the proper method of Bowel Preparation Protocol found compliances month on month, in December 2022 it was 30.29%, in January 2023 it improves to 55.26% &amp; reaches with compliances of 78.94% by the end of February 2023. Conclusion: In patients under the care of nurses who underwent enhanced education on bowel preparation had significantly better bowel preparation quality and higher compliance rates than those in inpatients of the control group. Nurses’ education was identified as an independent risk factor for bowel preparation quality. To make bowel preparation adequate &amp; sufficient frequent updating on knowledge of the recruited nurses need to be done on regular basis.

One-year review of real-time artificial intelligence (AI)-aided endoscopy performance.

Colonoscopies have long been the gold standard for detection of pre-malignant neoplastic lesions of the colon. Our previous study tried real-time artificial intelligence (AI)-aided colonoscopy over a three-month period and found significant improvements in collective and individual endoscopist's adenoma detection rates compared to baseline. As an expansion, this study evaluates the 1-year performance of AI-aided colonoscopy in the same institution. A prospective cohort study was conducted in a single institution in Singapore. The AI software used was GI Genius™ Intelligent Endoscopy Module, US-DG-2000309 © 2021 Medtronic. Between July 2021 and June 2022, polypectomy rates in non-AI-aided colonoscopies and AI-aided colonoscopies were calculated and compared. Some of the AI-aided colonoscopies were recorded and video reviewed. A "hit" was defined as a sustained detection of an area by the AI. If a polypectomy was performed for a "hit," its histology was reviewed. Additional calculations for polyp detection rate (PDR), adenoma detection rate (ADR), and adenoma detection per colonoscopy (ADPC) were performed. Cost analysis was performed to determine cost effectiveness of subscription to the AI program. 2433 AI-aided colonoscopies were performed between July 2021 and June 2022 and compared against 1770 non-AI-aided colonoscopies. AI-aided colonoscopies yielded significantly higher rates of polypectomies (33.6%) as compared with non-AI-aided colonoscopies (28.4%) (p < 0.001). Among the AI-aided colonoscopies, 1050 were reviewed and a final 843 were included for additional analysis. The polypectomy to "hit" ratio was 57.4%, PDR = 45.6%, ADR = 32.4%, and ADPC = 2.08. Histological review showed that 25 polyps (3.13%) were sessile-serrated adenomas. Cost analysis found that the increased polypectomy rates in AI-aided colonoscopes led to an increase in revenue, which covered the subscription cost with an excess of USD 20,000. AI-aided colonoscopy is a cost effective means of improving colonoscopy quality and may help advance colorectal cancer screening in Singapore.

Artificial Intelligence-Aided Endoscopy and Colorectal Cancer Screening

Colorectal cancer (CRC) is the third most common cancer worldwide, with the highest incidence reported in high-income countries. However, because of the slow progression of neoplastic precursors, along with the opportunity for their endoscopic detection and resection, a well-designed endoscopic screening program is expected to strongly decrease colorectal cancer incidence and mortality. In this regard, quality of colonoscopy has been clearly related with the risk of post-colonoscopy colorectal cancer. Recently, the development of artificial intelligence (AI) applications in the medical field has been growing in interest. Through machine learning processes, and, more recently, deep learning, if a very high numbers of learning samples are available, AI systems may automatically extract specific features from endoscopic images/videos without human intervention, helping the endoscopists in different aspects of their daily practice. The aim of this review is to summarize the current knowledge on AI-aided endoscopy, and to outline its potential role in colorectal cancer prevention.

A25 MEASURING THE OBSERVER (HAWTHORNE) EFFECT ON ADENOMA DETECTION RATES: A CASE-CONTROL RETROSPECTIVE STUDY

Abstract Background The effectiveness of colonoscopy screening to prevent colorectal cancer (CRC) is directly linked to its procedural quality. An independent observer (Hawthorne effect) can improve colonoscopy procedural quality metrics, including adenoma detection rate (ADR). However, the results of studies are limited or controversial. Purpose We aimed to evaluate the colonoscopy quality metrics in a group of patients undergoing screening or diagnostic colonoscopies under stringent observer conditions. Method In a single-center, case–control study, consecutive patients undergoing routine screening or diagnostic colonoscopy were prospectively enrolled. In the case group, all procedural steps and quality metrics were observed and documented, and the procedure was video recorded by an independent research assistant. In the control group, colonoscopies were performed without independent observation. Colonoscopy quality metrics such as polyp, adenoma, serrated lesions, and advanced adenoma detection rates (PDR, ADR, SLDR, AADR), the mean number of adenomas detected per patient (MAP), and the mean number of adenomas and serrated lesions detected per patient (MASP) were compared. The probabilities of increased quality metrics were evaluated through regression analyses weighted by the inversed probability of observation during the procedure. Result(s) We included a total of 687 patients (327 cases and 360 controls) in the final analyses. The case group had significantly higher PDRs (62.4% vs. 53.1%) and ADRs (39.4% vs. 28.3%) compared with the control group. The SLDR was also higher in the case group than in the control group, but the difference was not significant (7.3% vs. 4.4%; P = 0.14). The AADR was not significantly increased. After adjusting for potential confounders, the ADR and SLDR were 50% (odds ratio [OR] 1.51; 95%CI 1.05–2.17) and more than twofold (OR 2.17; 95%CI 1.05–4.47) more likely to be higher in the case group than in the control group. The MAP and MASP were significantly increased in the case group compared with the control group (P &amp;lt; 0.001). The regression analyses for both metrics demonstrated the direct and significant association between the Hawthorne effect and elevated MAP/MASP. Conclusion(s) The presence of an independent observer documenting colonoscopy quality metrics and video recording the colonoscopy resulted in a significant increase in ADR and other quality metrics. The Hawthorne effect should be considered an alternative strategy to advanced devices to improve colonoscopy quality in routine practice. Please acknowledge all funding agencies by checking the applicable boxes below None Disclosure of Interest M. Taghiakbari: None Declared, D. Coman: None Declared, M. Takla: None Declared, A. N. Barkun: None Declared, M. Frija-Gruman: None Declared, M. Bouin: None Declared, S. Bouchard: None Declared, E. Deslandres: None Declared, S. Sidani: None Declared, D. von Renteln Grant / Research support from: ERBE, Ventage, Pendopharm, and Pentax, Consultant of: Boston Scientific and Pendopharm

Supply and quality of colonoscopy according to the characteristics of gastroenterologists in the French population-based colorectal-cancer screening program.

Since its complete roll-out in 2009, the French colorectal cancer screening program (CRCSP) experienced 3 major constraints [use of a less efficient Guaiac-test (gFOBT), stopping the supply of Fecal-Immunochemical-Test kits (FIT), and suspension of the program due to the coronavirus disease 2019 (COVID-19)] affecting its effectiveness. To describe the impact of the constraints in terms of changes in the quality of screening-colonoscopy (Quali-Colo). This retrospective cohort study included screening-colonoscopies performed by gastroenterologists between Jan-2010 and Dec-2020 in people aged 50-74 living in Ile-de-France (France). The changes in Quali-colo (Proportion of colonoscopies performed beyond 7 mo (Colo_7 mo), Frequency of serious adverse events (SAE) and Colonoscopy detection rate) were described in a cohort of Gastroenterologists who performed at least one colonoscopy over each of the four periods defined according to the chronology of the constraints [gFOBT: Normal progress of the CRCSP using gFOBT (2010-2014); FIT: Normal progress of the CRCSP using FIT (2015-2018); STOP-FIT: Year (2019) during which the CRCSP experienced the cessation of the supply of test kits; COVID: Program suspension due to the COVID-19 health crisis (2020)]. The link between each dependent variable (Colo_7 mo; SAE occurrence, neoplasm detection rate) and the predictive factors was analyzed in a two-level multivariate hierarchical model. The 533 gastroenterologists (cohort) achieved 21509 screening colonoscopies over gFOBT period, 38352 over FIT, 7342 over STOP-FIT and 7995 over COVID period. The frequency of SAE did not change between periods (gFOBT: 0.3%; FIT: 0.3%; STOP-FIT: 0.3%; and COVID: 0.2%; P = 0.10). The risk of Colo_7 mo doubled between FIT [adjusted odds ratio (aOR): 1.2 (1.1; 1.2)] and STOP-FIT [aOR: 2.4 (2.1; 2.6)]; then, decreased by 40% between STOP-FIT and COVID [aOR: 2.0 (1.8; 2.2)]. Regardless of the period, this Colo_7 mo's risk was twice as high for screening colonoscopy performed in a public hospital [aOR: 2.1 (1.3; 3.6)] compared to screening-colonoscopy performed in a private clinic. The neoplasm detection, which increased by 60% between gFOBT and FIT [aOR: 1.6 (1.5; 1.7)], decreased by 40% between FIT and COVID [aOR: 1.1 (1.0; 1.3)]. The constraints likely affected the time-to-colonoscopy as well as the colonoscopy detection rate without impacting the SAE's occurrence, highlighting the need for a respectable reference time-to-colonoscopy in CRCSP.

A108 AUTOMATED DETECTION OF ILEOCECAL VALVE, APPENDICEAL ORIFICE, AND POLYP DURING COLONOSCOPY USING A DEEP LEARNING MODEL

Abstract Background Identification and photo-documentation of the ileocecal valve (ICV) and appendiceal orifice (AO) confirm completeness of colonoscopy examinations. We hypothesized that an artificial intelligence (AI)-empowered solution could help us automatically differentiate anatomical landmarks such as AO and ICV from polyps and normal colon mucosa. Purpose We aimed to develop and test a deep convolutional neural network (DCNN) model that can automatically identify ICV and AO, and differentiate these landmarks from normal mucosa and colorectal polyps. Method We prospectively collected annotated full-length colonoscopy videos of 318 patients undergoing outpatient colonoscopies. We created three non-overlapping training, validation, and test datasets with 25,444 unaltered frames extracted from the colonoscopy videos showing four landmarks/image classes (AO, ICV, normal mucosa, and polyps). For each landmark, we extracted an average of 30 frames for each time of its appearance. All the extracted frames were reviewed and annotated by a team of three clinicians. Using a quality assessment tool, the clinicians examined a total of 86,754 frames (7982 AO, 8374 ICV, 32,971 polyps, and 37,427 normal mucosa) and verified whether or not the frame contained one unique landmark. For this research, all frames were extracted from the white-light colonoscopies, and all narrow-band imaging frames were excluded. A DCNN classification model was developed, validated, and tested in separate datasets of images. The primary outcome was the proportion of patients in whom the AI model could identify both ICV and AO, and differentiate them from polyps and normal mucosa, with an accuracy of detecting both AO and ICV above a threshold of 40% (representing a value in which reliable identification of the landmarks can be assumed without increasing false-positive alerts). Result(s) We trained a DCNN AI model on 21,503 unaltered frames extracted from the recorded colonoscopy videos of 272 patients, and validated and tested the model on 1,924 (25 patients) and 2,017 (21 patients) unaltered frames, respectively. We applied a transfer learning technique to fine-tune the model parameters to the endoscopic images using a cross-entropy loss function and back-propagation algorithm. After training and validation, the DCNN model could identify both AO and ICV in 18 out of 21 patients (85.71%), if accuracies were above the threshold of 40%. The accuracy of the model for differentiating AO from normal mucosa, and ICV from normal mucosa were 86.37% (95% CI 84.06% to 88.45%), and 86.44% (95% CI 84.06% to 88.59%), respectively. Furthermore, the accuracy of the model for differentiating polyps from normal mucosa was 88.57% (95% CI 86.60% to 90.33%). Conclusion(s) The model can reliably distinguish these anatomical landmarks from normal mucosa and colorectal polyps. It can be implemented into automated colonoscopy report generation, photo-documentation, and quality auditing solutions to improve colonoscopy reporting quality. Please acknowledge all funding agencies by checking the applicable boxes below Other Please indicate your source of funding; MEDTEQ Disclosure of Interest M. Taghiakbari: None Declared, S. Hamidi Ghalehjegh Employee of: Imagia Canexia Health Inc. , E. Jehanno Employee of: Imagia Canexia Health Inc. , T. Berthier Employee of: Imagia Canexia Health Inc. , L. di Jorio Employee of: Imagia Canexia Health Inc. , A. N. Barkun Grant / Research support from: co-awardee in funded research projects with Imagia Canexia Health Inc., Consultant of: Medtronic Inc. and A.I. VALI Inc, E. Deslandres: None Declared, S. Bouchard: None Declared, S. Sidani: None Declared, Y. Bengio: None Declared, D. von Renteln Grant / Research support from: ERBE, Ventage, Pendopharm, and Pentax, Consultant of: Boston Scientific and Pendopharm