AI-assisted capsule endoscopy reading in suspected small bowel bleeding: a multicentre prospective study.

BACKGROUND: Capsule endoscopy reading is time consuming, and readers are required to maintain attention so as not to miss significant findings. Deep convolutional neural networks can recognise relevant findings, possibly exceeding human performances and reducing the reading time of capsule endoscopy. Our primary aim was to assess the non-inferiority of artificial intelligence (AI)-assisted reading versus standard reading for potentially small bowel bleeding lesions (high P2, moderate P1; Saurin classification) at per-patient analysis. The mean reading time in both reading modalities was evaluated among the secondary endpoints. METHODS: Patients aged 18 years or older with suspected small bowel bleeding (with anaemia with or without melena or haematochezia, and negative bidirectional endoscopy) were prospectively enrolled at 14 European centres. Patients underwent small bowel capsule endoscopy with the Navicam SB system (Ankon, China), which is provided with a deep neural network-based AI system (ProScan) for automatic detection of lesions. Initial reading was performed in standard reading mode. Second blinded reading was performed with AI assistance (the AI operated a first-automated reading, and only AI-selected images were assessed by human readers). The primary endpoint was to assess the non-inferiority of AI-assisted reading versus standard reading in the detection (diagnostic yield) of potentially small bowel bleeding P1 and P2 lesions in a per-patient analysis. This study is registered with ClinicalTrials.gov, NCT04821349. FINDINGS: From Feb 17, 2021 to Dec 29, 2021, 137 patients were prospectively enrolled. 133 patients were included in the final analysis (73 [55%] female, mean age 66·5 years [SD 14·4]; 112 [84%] completed capsule endoscopy). At per-patient analysis, the diagnostic yield of P1 and P2 lesions in AI-assisted reading (98 [73·7%] of 133 lesions) was non-inferior (p<0·0001) and superior (p=0·0213) to standard reading (82 [62·4%] of 133; 95% CI 3·6-19·0). Mean small bowel reading time was 33·7 min (SD 22·9) in standard reading and 3·8 min (3·3) in AI-assisted reading (p<0·0001). INTERPRETATION: AI-assisted reading might provide more accurate and faster detection of clinically relevant small bowel bleeding lesions than standard reading. FUNDING: ANKON Technologies, China and AnX Robotica, USA provided the NaviCam SB system.

Small-bowel capsule endoscopy and device-assisted enteroscopy for diagnosis and treatment of small-bowel disorders: European Society of Gastrointestinal Endoscopy (ESGE) Guideline - Update 2022.

MR1: ESGE recommends small-bowel capsule endoscopy as the first-line examination, before consideration of other endoscopic and radiological diagnostic tests for suspected small-bowel bleeding, given the excellent safety profile of capsule endoscopy, its patient tolerability, and its potential to visualize the entire small-bowel mucosa.Strong recommendation, moderate quality evidence. MR2: ESGE recommends small-bowel capsule endoscopy in patients with overt suspected small-bowel bleeding as soon as possible after the bleeding episode, ideally within 48 hours, to maximize the diagnostic and subsequent therapeutic yield.Strong recommendation, high quality evidence. MR3: ESGE does not recommend routine second-look endoscopy prior to small-bowel capsule endoscopy in patients with suspected small-bowel bleeding or iron-deficiency anemia.Strong recommendation, low quality evidence. MR4: ESGE recommends conservative management in those patients with suspected small-bowel bleeding and high quality negative small-bowel capsule endoscopy.Strong recommendation, moderate quality evidence. MR5: ESGE recommends device-assisted enteroscopy to confirm and possibly treat lesions identified by small-bowel capsule endoscopy.Strong recommendation, high quality evidence. MR6: ESGE recommends the performance of small-bowel capsule endoscopy as a first-line examination in patients with iron-deficiency anemia when small bowel evaluation is indicated.Strong recommendation, high quality evidence. MR7: ESGE recommends small-bowel capsule endoscopy in patients with suspected Crohn's disease and negative ileocolonoscopy findings as the initial diagnostic modality for investigating the small bowel, in the absence of obstructive symptoms or known bowel stenosis.Strong recommendation, high quality evidence. MR8: ESGE recommends, in patients with unremarkable or nondiagnostic findings from dedicated small-bowel cross-sectional imaging, small-bowel capsule endoscopy as a subsequent investigation if deemed likely to influence patient management.Strong recommendation, low quality evidence. MR9: ESGE recommends, in patients with established Crohn's disease, the use of a patency capsule before small-bowel capsule endoscopy to decrease the capsule retention rate.Strong recommendation, moderate quality evidence. MR10: ESGE recommends device-assisted enteroscopy (DAE) as an alternative to surgery for foreign bodies retained in the small bowel requiring retrieval in patients without acute intestinal obstruction.Strong recommendation, moderate quality evidence. MR11: ESGE recommends DAE-endoscopic retrograde cholangiopancreatography (DAE-ERCP) as a first-line endoscopic approach to treat pancreaticobiliary diseases in patients with surgically altered anatomy (except for Billroth II patients).Strong recommendation, moderate quality evidence.

Score reproducibility and reliability in differentiating small bowel subepithelial masses from innocent bulges.

AIMS: The primary aim of this study was to assess the reliability, intra- and inter-observer variation of the SPICE, Mucosal protrusion angle (MPA) and SHYUNG scores in differentiating a subepithelial mass (SEM) from a bulge. METHODS: This retrospective multicentre study analysed the 3 scores, radiological studies, enteroscopy and/or surgical findings. RESULTS: 100 patients with a potential SEM (mean age 57.6years) were recruited with 75 patients having pathology. In patients with a SEM the mean SPICE score was 2.04 (95% CI 1.82-2.26) as compared to 1.16 (95% CI 0.81-1.51) without any pathology (AUC 0.74, p<0.001), with a fair intra-observer agreement (Kappa 0.3, p<0.001) and slight inter-observer agreement (Kappa 0.14, p<0.05). SPICE had a 37.3% sensitivity and 92.0% specificity in distinguishing between a SEM and bulge, whereas MPA<90˚ had 58.7% and 76.0% respectively, with poor intra-observer(p = 0.05) and interobserver agreement (p = 0.64). The SHYUNG demonstrated a moderate intra-observer (Kappa 0.44, p<0.001) and slight inter-observer reliability (Kappa 0.18, p<0.001). The sensitivity of an elevated SHYUNG score (≥4) in identifying a SEM was 18.7% with a specificity of 92.0% (AUC 0.71, p = 0.002). CONCLUSIONS: Though these scores are easy to use, they have, at best, slight to moderate intra and inter-observer agreement. Their overall diagnostic performances are limited.

New Generation Express View: An Artificial Intelligence Software Effectively Reduces Capsule Endoscopy Reading Times.

BACKGROUND: Reading capsule endoscopy (CE) is time-consuming. The Express View (EV) (IntroMedic, Seoul, Korea) software was designed to shorten CE video reading. Our primary aim was to evaluate the diagnostic accuracy of EV in detecting significant small-bowel (SB) lesions. We also compared the reading times with EV mode and standard reading (SR). METHODS: 126 patients with suspected SB bleeding and/or suspected neoplasia were prospectively enrolled and underwent SB CE (MiroCam®1200, IntroMedic, Seoul, Korea). CE evaluation was performed in standard and EV mode. In case of discrepancies between SR and EV readings, a consensus was reached after reviewing the video segments and the findings were re-classified. RESULTS: The completion rate of SB CE in our cohort was 86.5% and no retention occurred. The per-patient analysis of sensitivity, specificity, positive predictive value, negative predictive value, and diagnostic accuracy of EV compared to SR were 86%, 86%, 90%, 81%, and 86%, respectively, before consensus. After consensus, they increased to 97%, 100%, 100%, 96%, and 98%, respectively. The median reading time with SR and EV was 71 min (range 26−340) and 13 min (range 3−85), respectively (p < 0.001). CONCLUSIONS: The new-generation EV shows high diagnostic accuracy and significantly reduces CE reading times.

Indications, Detection, Completion and Retention Rates of Capsule Endoscopy in Two Decades of Use: A Systematic Review and Meta-Analysis.

BACKGROUND: Capsule endoscopy (CE) has become a widespread modality for non-invasive evaluation of the gastrointestinal (GI) tract, with several CE models having been developed throughout the years. The aim of this systematic review and meta-analysis is to evaluate performance measures such as completion, detection and retention rates of CE. METHODS: Literature through to August 2021 was screened for articles regarding all capsule types: small bowel, double-headed capsule for the colon or PillCam®Crohn's capsule, magnetically-controlled capsule endoscopy, esophageal capsule and patency capsule. Primary outcomes included detection rate (DR), completion rate (CR) and capsule retention rate (RR). DR, CR and RR were also analyzed in relation to indications such as obscure GI bleeding (OGIB), known/suspected Crohn's disease (CD), celiac disease (CeD), neoplastic lesions (NL) and clinical symptoms (CS). RESULTS: 328 original articles involving 86,930 patients who underwent CE were included. OGIB was the most common indication (n = 44,750), followed by CS (n = 17,897), CD (n = 11,299), NL (n = 4989) and CeD (n = 947). The most used capsule type was small bowel CE in 236 studies. DR, CR and RR for all indications were 59%, 89.6% and 2%, respectively. According to specific indications: DR were 55%, 66%, 63%, 52% and 62%; CR were 90.6%, 86.5%, 78.2%, 94% and 92.8%; and RR were 2%, 4%, 1%, 6% and 2%. CONCLUSIONS: Pooled DR, CR and RR are acceptable for all capsule types. OGIB is the most common indication for CE. Technological advancements have expanded the scope of CE devices in detecting GI pathology with acceptable rates for a complete examination.

Small Bowel Capsule Endoscopy and artificial intelligence: First or second reader?

Several machine learning algorithms have been developed in the past years with the aim to improve SBCE (Small Bowel Capsule Endoscopy) feasibility ensuring at the same time a high diagnostic accuracy. If past algorithms were affected by low performances and unsatisfactory accuracy, deep learning systems raised up the expectancy of effective AI (Artificial Intelligence) application in SBCE reading. Automatic detection and characterization of lesions, such as angioectasias, erosions and ulcers, would significantly shorten reading time other than improve reader attention during SBCE review in routine activity. It is debated whether AI can be used as first or second reader. This issue should be further investigated measuring accuracy and cost-effectiveness of AI systems. Currently, AI has been mostly evaluated as first reader. However, second reading may play an important role in SBCE training as well as for better characterizing lesions for which the first reader was uncertain.

Real-Time Diagnosis of Helicobacter pylori During Endoscopy by Gastric Juice Analysis.

EndoFaster is a high precision device for gastric juice analysis in real time during gastroscopy that enables detection of Helicobacter pylori infection and hypochlorhydric/achlorhydric conditions through the measurement of ammonium concentration and gastric pH. The high accuracy and feasibility of this technology enables a more accurate diagnosis and a reduced number of histologies, focusing the attention of the endoscopist on patients with high risk for cancer progression and limiting or avoiding biopsies on the low-risk ones while also saving costs and time.

The diagnostic yield of colonoscopy in hospitalized patients. An observational multicenter prospective study.

BACKGROUND: Colonoscopy demands a considerable amount of resources, and little is known about its diagnostic yield among inpatients. AIMS: To assess indications, diagnostic yield and findings of colonoscopy for inpatients, and to identify risk factors for relevant findings and cancer. METHODS: Multicentre, prospective, observational study including 12 hospitals. Consecutive adult inpatients undergoing colonoscopy were evaluated from February through November 2019. RESULTS: 1,302 inpatients underwent colonoscopy. Diagnostic yield for relevant findings and cancer was 586 (45%) and 112 (8.6%), respectively. Adequate colon cleansing was achieved in 896 (68.8%) patients. Split-dose/same-day regimen was adopted in 847 (65%) patients. Factors associated to relevant findings were age ≥70 years (RR 1.32), male gender (RR 1.11), blood loss (RR 1.22) and adequate cleansing (RR 1.63). Age ≥70 years (RR 2.08), no previous colonoscopy (RR 2.69) and split-dose/same-day regimen (RR 1.59) significantly increased cancer detection. Implementing adequate cleansing and split-dose/same-day regimen in all patients would increase the diagnostic yield for any relevant findings and cancer from 43% to 70% and from 6% to 10%, respectively. CONCLUSION: Relevant colorectal diseases and cancer were frequent among inpatients. Factors associated with detection of relevant findings were identified. Adequate colon cleansing and split-dose/same-day regimen significantly increased colonoscopy diagnostic yield.

1L- vs. 4L-Polyethylene glycol for bowel preparation before colonoscopy among inpatients: A propensity score-matching analysis.

BACKGROUND: Inpatients are at risk for inadequate colon cleansing. Experts recommend 4L-polyethylene-glycol (PEG) solution. A higher colon cleansing adequacy rate for a hyperosmolar 1L-PEG plus ascorbate prep has been recently reported. AIMS: We aimed to determine whether 1L-PEG outperforms 4L-PEG among inpatients. METHODS: post-hoc analysis of a large Italian multicenter prospective observational study among inpatients (QIPS study). We performed a propensity score matching between 1L-PEG and 4L-PEG group. The primary outcome was the rate of adequate colon cleansing as assessed by unblinded endoscopists through Boston scale. Secondary outcome was the safety profile. RESULTS: Among 1,004 patients undergoing colonoscopy, 724 (72%) were prescribed 4L-PEG and 280 (28%) 1L-PEG. The overall adequate colon cleansing rate was 69.2% (n = 695). We matched 274 pairs of patients with similar distribution of confounders. The rate of patients with adequate colon cleansing was higher in 1L-PEG than in 4L-PEG group (84.3% vs. 77.4%, p = 0.039). No different shift in serum concentration of electrolytes (namely Na+, K+, Ca2+), creatinine and hematocrit were observed for both preparations. CONCLUSION: We found a higher rate of adequate colon cleansing for colonoscopy with the 1L-PEG bowel prep vs. 4L-PEG, with apparent similar safety profile, among inpatients. A confirmatory randomized trial is needed. (ClinicalTrials.gov no: NCT04310332).

How to identify patients who are less likely to have metachronous neoplasms after a colon cancer: a predictive model.

BACKGROUND: Patients with prior colon cancer have increased risk of metachronous colorectal neoplasms; therefore, endoscopic surveillance is indicated. Current recommendations are not risk-stratified. We investigated predictive factors for colorectal neoplasms to build a model to spare colonoscopies for low-risk patients. METHODS: This was a multicenter, retrospective study including patients who underwent surgery for colon cancer in 2001 - 2008 (derivation cohort) and 2009 - 2013 (validation cohort). A predictive model for neoplasm occurrence at second surveillance colonoscopy was developed and validated. RESULTS: 421 and 203 patients were included in derivation and validation cohort, respectively. At second surveillance colonoscopy, 112 (26.6 %) and 55 (27.1 %) patients had metachronous neoplasms in derivation and validation groups; three cancers were detected in the latter. History of left-sided colon cancer (OR 1.64, 95 %CI 1.02 - 2.64), ≥ 1 advanced adenoma at index colonoscopy (OR 1.90, 95 %CI 1.05 - 3.43), and ≥ 1 adenoma at first surveillance colonoscopy (OR 2.06, 95 %CI 1.29 - 3.27) were independently predictive of metachronous colorectal neoplasms at second surveillance colonoscopy. For patients without such risk factors, diagnostic accuracy parameters were: 89.3 % (95 %CI 82.0 %-94.3 %) and 78.2 % (95 %CI 65.0 %-88.2 %) sensitivity, and 28.5 % (95 %CI 23.5 %-33.9 %) and 33.8 % (95 %CI 26.2 %-42.0 %) specificity in derivation and validation group, respectively. No cancer would be missed. CONCLUSIONS: Patients with prior left-sided colon cancer or ≥ 1 advanced adenoma at index colonoscopy or ≥ 1 adenoma at first surveillance colonoscopy had a significantly higher risk of neoplasms at second surveillance colonoscopy; patients without such factors had much lower risk and could safely skip the second surveillance colonoscopy. A prospective, multicenter validation study is needed.