Bowel Lesions Research Articles

Computed tomography enterography radiomics and machine learning for identification of Crohn’s disease

BackgroundCrohn’s disease is a severe chronic and relapsing inflammatory bowel disease. Although contrast-enhanced computed tomography enterography is commonly used to evaluate crohn’s disease, its imaging findings are often nonspecific and can overlap with other bowel diseases. Recent studies have explored the application of radiomics-based machine learning algorithms to aid in the diagnosis of medical images. This study aims to develop a non-invasive method for detecting bowel lesions associated with Crohn’s disease using CT enterography radiomics and machine learning algorithms.MethodsA total of 139 patients with pathologically confirmed Crohn’s disease were retrospectively enrolled in this study. Radiomics features were extracted from both arterial- and venous-phase CT enterography images, representing both bowel lesions with Crohn’s disease and segments of normal bowel. A machine learning classification system was constructed by combining six selected radiomics features with eight classification algorithms. The models were trained using leave-one-out cross-validation and evaluated for accuracy.ResultsThe classification model demonstrated robust performance and high accuracy, with an area under the curve of 0.938 and 0.961 for the arterial- and venous-phase images, respectively. The model achieved an accuracy of 0.938 for arterial-phase images and 0.961 for venous-phase images.ConclusionsThis study successfully identified a radiomics machine learning method that effectively differentiates Crohn’s disease bowel lesions from normal bowel segments. Further studies with larger sample sizes and external cohorts are needed to validate these findings.

Establishing an AI model and application for automated capsule endoscopy recognition based on convolutional neural networks (with video).

Although capsule endoscopy (CE) is a crucial tool for diagnosing small bowel diseases, the need to process a vast number of images imposes a significant workload on physicians, leading to a high risk of missed diagnoses. This study aims to develop an artificial intelligence (AI) model and application based on convolutional neural networks that can automatically recognize various lesions in small bowel capsule endoscopy. Three small bowel capsule endoscopy datasets were used for AI model training, validation, and testing, encompassing 12 categories of images. The model's performance was evaluated using metrics such as AUC, sensitivity, specificity, precision, accuracy, and F1 score to select the best model. A human-machine comparison experiment was conducted using the best model and endoscopists with varying levels of experience. Model interpretability was analyzed using Grad-CAM and SHAP techniques. Finally, a clinical application was developed based on the best model using PyQt5 technology. A total of 34,303 images were included in this study. The best model, MobileNetv3-large, achieved a weighted average sensitivity of 87.17%, specificity of 98.77%, and an AUC of 0.9897 across all categories. The application developed based on this model performed exceptionally well in comparison with endoscopists, achieving an accuracy of 87.17% and a processing speed of 75.04 frames per second, surpassing endoscopists of varying experience levels. The AI model and application developed based on convolutional neural networks can quickly and accurately identify 12 types of small bowel lesions. With its high sensitivity, this system can effectively assist physicians in interpreting small bowel capsule endoscopy images.Future studies will validate the AI system for video evaluations and real-world clinical integration.

Prediction of Significant Lesions on Capsule Endoscopy in Patients with Suspected Small Bowel Bleeding: External Validation of SSB Capsule Dx Score

Introduction: Deciding which patients with suspected small bowel bleeding (SSB) would benefit most from small bowel capsule endoscopy (SBCE) is challenging. Our aim was to perform an external validation of the recently developed SSB Capsule Diagnostic (Dx) score that includes 3 variables (hospital admission with overt bleeding, hemoglobin <6.4 g/dL and age <54 years) and has been shown to be potentially useful in limiting the use of SBCE in SSB low-risk patients. Methods: Retrospectively included all adult patients submitted to SBCE for SSB between November 2007 and December 2019. Patients’ demographic, clinical and laboratorial data at the time of SBCE were recorded. Small bowel lesions were classified according to Saurin classification. The SSB Capsule Dx score was calculated, and its calibration and discrimination ability were assessed. Results: We assessed 473 SBCEs for SSB. Patients’ mean age was 61.2 ± 17.9 years and 65.8% were female. P2 lesions were present in 36.2% of SBCEs. There was a significant association between the score and P2 lesions (p < 0.001). Mean score was −0.21 ± 0.87 having a fair accuracy toward the outcome (C-statistic 0.700; 95% confidence interval, 0.652–0.749; p < 0.001). A cutoff value of 0 was found to have a high sensitivity (86.0%) and negative predictive value (84.9%) for the diagnosis of P2 lesions at SBCE. Conclusion: Patients with a SSB Capsule Dx score <0 are unlikely to have a significant lesion on SBCE, thus its routine use in the clinical practice may be useful in the identification of low-risk SSB patients.

Pixel to Pathology: A Prospective Cross-Sectional Study on the Role of Multidetector Computed Tomography in the Evaluation of Malignant Large Bowel Lesions With Histopathological Correlation.

Colorectal lesions can present as a mass or as focal or diffuse thickening of the colon wall and may also be associated with abnormalities in the perienteric region. Multidetector computed tomography (MDCT) enables simultaneous imaging of both extracolonic structures and the gut wall. It is instrumental in assessing tumor extent, detecting pericolic dissemination, including lymph node involvement, and identifying metastases. This study aims to evaluate the role of MDCT in diagnosing large bowel lesions in patients referred from the surgery department with suspected large bowel pathology and to correlate MDCT findings with histopathological results. In the Department of Radiodiagnosis at Mahatma Gandhi Medical College and Research Institute (MGMCRI), Pondicherry, this study was carried out from February 2023 to January 2024 over a one-year span. CT imaging was carried out utilizing a GE Optima 128 Slice MDCT scanner for instances of clinically suspected large bowel pathology. IV contrast, rectal, oral (positive), and plain CT were the imaging protocols that were used. Thirty of the 45 patients were female, and 15 were male; most of the patients were in the 46-60 age range. Twenty-three of the 24 instances with a CT diagnosis of malignancy had a histological confirmation. On CT, all 21 benign lesions were correctly detected. Benign lesions showed homogenous attenuation, mild symmetric wall thickening, and diffuse bowel involvement. Heterogeneous attenuation, pronounced asymmetric wall thickening, and localized bowel involvement were observed in malignant lesions. For the diagnosis and differentiation of benign from malignant colon and rectum lesions,MDCThas shown itself to be an excellent technique. MDCT provides additional information about related pericolic abnormalities, lymph node presence, neighboring organ infiltration, and distant metastases in addition to identifying the lesions.

EP142 - "Unravelling the Enigma: Navigating the Diagnostic Dilemma of Abdominal Toxoplasmosis"

Abstract Introduction Abdominal toxoplasmosis, an infrequent yet medically significant manifestation of Toxoplasma infection, primarily caused by the protozoan parasite Toxoplasma gondii. Although toxoplasmosis can affect various organs, its occurrence in the abdominal region introduces unique diagnostic complexities, presenting symptoms such as abdominal pain, gastrointestinal disruptions, and systemic complications. Case Report In a notable case, a middle-aged patient presented with perplexing gastrointestinal symptoms, including nausea, vomiting, abdominal pain, and irregular bowel habits. With a history of anaemia and thrombocytopenia necessitating transfusions and yielding inconclusive bone marrow biopsies, the patient underwent OGD and colonoscopy with biopsies during initial management, revealing a hiatus hernia and diverticulosis that did not explain the symptoms. Despite being discharged pending biopsy results, subsequent normal findings failed to clarify. Upon readmission with persistent gastrointestinal complaints, a CTAP revealed two suspicious neoplastic focal areas of small bowel wall thickening and enlarged mesenteric lymph nodes. The MDT discussions ensued to rule out malignancy, leading to a follow-up PET scan suggesting nodal disease, raising concerns about lymphoma, small bowel adenocarcinoma, or carcinoid tumour and recommended an ultrasound-guided lymph node biopsy, 5 HIAA if asymptomatic, and diagnostic laparoscopy if symptomatic or obstructed. Subsequent core biopsy of a neck lymph node unveiled appearances consistent with toxoplasmosis, prompting treatment with anti-parasitic drugs. The patient awaits capsule endoscopy, and a follow-up CT scan displayed a reduction in bowel lesions and lymph nodes. Conclusion This case highlights the swift and precise identification of toxoplasmosis, resulting in an optimal patient outcome. The clinical manifestations of toxoplasmosis are diverse and lack distinct characteristics, making it challenging to definitively diagnose, particularly in immunocompetent individuals.

Robotic Pauli repair of parastomal hernia.

Robotic retro-muscular mesh repair of parastomal hernia (PH) is possible with transversus abdominis release. The aim of this study is to report mid-term results with this method. Patients who underwent surgery for PH were enrolled in a prospective study. Patients with concomitant midline hernia also underwent Rives repair. Polypropylene meshes with an absorbable buffer mesh were used in the retro-muscular pocket-except PTFE was chosen for inflammatory bowel disease patients. Follow-up was by telephone and physical examination and CT if suspicion of complication. Of the 56 included ostomy patients, 44 had colostomies, 10 had ileostomies, and two had urostomies. Thirty-nine percent had a concomitant hernia, and 25% had recurrent PH. The mean operating time without accessory repair was 156min (SD 37), and with accessory repairs 220min (SD 62). One bowel lesion and one vascular injury to the stoma occurred, prompting intraoperative stoma revision without postoperative morbidity. Postoperative complications ensued in 12 patients (23%). One patient had stoma necrosis after 3weeks and the stoma was relocated. One patient had a mesh infection, 3 patients temporary ileus, one patient hypovolemic renal failure, and two patients experienced a flare-up of their Crohn's Disease. One parastomal abscess was drained percutaneously. The median postoperative stay was 3days (1-21; mean 3.7days), and the readmission rate was 8.9%. The recurrence rate at mean and median 24months follow-up is 5.4%, two in ileostomy patients and one colostomy patient with unaltered bulging where CT shows a relatively large prolapse and recorded as a failure. Robotic Pauli repair has shown promising results. However, repair fails of PH at small bowel ostomies with short mesentery. ePauli/TAR is now the preliminary choice for eligible patients at our site. Comparative studies with the intraperitoneal mesh standard are needed.

Novel ultrasound capsule endoscopy for gastrointestinal scanning: An in vivo animal study.

EUS is an important modality for diagnosis and assessment of gastrointestinal (GI) subepithelial lesions. However, EUS is invasive and operator-dependent and requires sedation in most cases. The newly developed ultrasound capsule endoscopy (USCE) system, with both white-light and ultrasound imaging modalities, is a minimally invasive method for superficial and submucosal imaging of the esophagus. This animal study aimed to evaluate the feasibility and efficacy of the USCE system for upper GI tract and small bowel scanning. Three Bama miniature pigs were selected to scan their esophagus, stomach, small bowel, and simulated submucosal lesions. USCE was performed first, followed by EUS. The feasibility of USCE was measured by obtaining ultrasound images of normal GI walls and submucosal lesions under the guidance of optical viewing. The efficacy of USCE was evaluated by comparing tissue structures and lesion features shown on ultrasound images obtained with both instruments. Under the optical mode of USCE, the GI tract was well visualized, and all simulated lesions were located. Clear ultrasound images of normal GI tract and submucosal lesions were acquired. Ultrasound images of the esophagus, stomach, and small bowel were characterized by differentiated multilayer structures on USCE, which was consistent with the structures displayed on EUS. And the visualization of submucosal lesions, using both USCE and EUS, was characterized by a hypoechoic and well-demarcated mass in the layer of submucosa. This animal study indicated the feasibility and potential clinical efficacy of this USCE for simultaneous optical mucosal visualization and transmural ultrasound imaging of upper GI tract and small bowel, providing possibility of using this technology for a wider range of GI tract.

Predicting disease recurrence in patients with endometriosis: an observational study

BackgroundDespite surgical and pharmacological interventions, endometriosis can recur. Reliable information regarding risk of recurrence following a first diagnosis is scant. The aim of this study was to examine clinical and survey data in the setting of disease recurrence to identify predictors of risk of endometriosis recurrence.MethodsThis observational study reviewed data from 794 patients having surgery for pelvic pain or endometriosis. Patients were stratified into two analytic groups based on self-reported or surgically confirmed recurrent endometriosis. Statistical analyses included univariate, followed by multivariate logistic regression to identify risk factors of recurrence, with least absolute shrinkage and selection operator (Lasso) regularisation. Risk-calibrated Supersparse Linear Integer Models (RiskSLIM) and survival analyses (with Lasso) were undertaken to identify predictive features of recurrence.ResultsSeveral significant features were repeatedly identified in association with recurrence, including adhesions, high rASRM score, deep disease, bowel lesions, adenomyosis, emergency room attendance for pelvic pain, younger age at menarche, higher gravidity, high blood pressure and older age. In the surgically confirmed group, with a score of 5, the RiskSLIM method was able to predict the risk of recurrence (compared to a single diagnosis) at 95.3% and included adenomyosis and adhesions in the model. Survival analysis further highlighted bowel lesions, adhesions and adenomyosis.ConclusionsFollowing an initial diagnosis of endometriosis, clinical decision-making regarding disease management should take into consideration the presence of bowel lesions, adhesions and adenomyosis, which increase the risk of endometriosis recurrence.

Role of Bowel Hydrosonography in Evaluation of Inflammatory Bowel Diseases Compared to CT or MR Enterography

Abstract Background bowel diseases (IBD). CT entrography (CTE), bowel US, magnetic resonance enterography Crohn’s disease and ulcerative colitis are chronic, relapsing inflammatory (MRE) are decisive medical imaging modalities for diagnosing and locating bowel lesions with its extramural extent and complications. They assess the degree of activity, help clinicians to identify patients in need of surgery, and can be used for patient follow-up. Aim To evaluate the utility of small bowel hydrosonography (SBH) in assessment if patient s with proven IBD in comparison to the currently available CT/MR enterographic techniques. Methods The study was conducted on 41 patients with histologically proven IBD. All patients were subjected to clinical assessment, laboratory investigations, MRE/CTE then bowel ultrasound. Results This study was conducted on 41 patients (19.5%) with ulcerative colitis and (80.5 %) with Crohn's disease; Terminal illeum is the most affected detected segment (22.1%) by bowel ultrasound and(22.5%) by CTE/MRE, ultrasound also showed statistically significance results in detection evidence of complicated Crohn’s disease such as fistula and abscess. Conclusion In comparison to MRE/CTE, bowel hydrosonography is a useful, non- invasive, and feasible bedside imaging tool for the detection of inflammation, detection of complications, and follow-up of IBD patients.

Effects of bit chewing on postoperative gastric reflux volume in horses with small bowel lesions: A preliminary clinical controlled trial

Effects of bit chewing on postoperative gastric reflux volume in horses with small bowel lesions: A preliminary clinical controlled trial

A Novel Method to Localize Patency Capsule by Ileocolonoscopy Facilitates Endoscopic Assessment of the Small and Large Intestine in Patients with Crohn’s Disease

Introduction: Patients with Crohn’s disease (CD) require an assessment of small bowel lesions, while difficulties exist in performing small intestinal examinations, especially in small-sized medical offices. Small bowel capsule endoscopy (SBCE) is handy and can be performed in most clinical settings. The only drawback of SBCE is a requirement of patency testing prior to the exam because it sometimes requires CT scanning to localize the ingested patency capsule (PC), which may be a substantial burden for the patient. We have developed a novel PC detection system named PICS (patency capsule, ileocolonoscopy, and small bowel capsule endoscopy) method by which we can avoid CT scanning. In the PICS method, ileocolonoscopy (ICS) is performed after 30–33 h of PC ingestion and the PC can be localized by ICS in patients who have not excreted the PC, and the entire intestine can be examined in combination with subsequent SBCE without additional bowel preparation. The aim of this study was to assess the usefulness and safety of the PICS method for CD patients. Methods: CD patients who underwent PICS method from April 2021 to March 2023 were reviewed for clinical data, outcome of PICS method including the rates of PC detection by ICS, the number of patients underwent SBCE, and adverse events. Lewis score was used to assess SBCE results. Results: The PICS method was performed in 54 patients. The median age of patients was 28.5 years old, and 64.8% of them were ileocolic type. The median disease duration was 10.5 months and 24.1% had history of small bowel resection. Five cases (9.3%) confirmed gastrointestinal patency by ICS, and none of the cases required CT scanning. One patient who could not be confirmed patency by ICS, and the other patient who excreted PC but was found ileal stenosis by ICS did not undergo SBCE. Remaining 52 patients received SBCE, and the median Lewis score of them was 0 (IQR 0, 450). There were no adverse events including small bowel obstruction by PC and SBCE retention in this series. Conclusion: The PICS method is not only feasible and safe but also convenient to assess disease extent in patients with CD. By localizing PC with ICS, additional CT scanning could be unnecessary for SBCE, which benefits both physicians and CD patients.

Long-Term Monitoring and Clinical Implications of Small Bowel Capsule Endoscopy in Patients with Crohn’s Disease with Small Bowel Lesions: A Retrospective Analysis

Introduction: Crohn’s disease (CD) induces persistent inflammation throughout the gastrointestinal (GI) tract, potentially resulting in complications such as intestinal stenosis and fistulas, particularly in the small bowel. Small bowel capsule endoscopy (SBCE) is recommended for monitoring CD, especially when GI tract patency is maintained. This study aimed to retrospectively assess patients with CD who underwent SBCE to determine the timing of clinical changes and address the current lack of evidence regarding GI tract patency loss during CD treatment. Methods: Of the 166 consecutive patients who underwent SBCE at our institution, 120 were followed up and included in this study. Forty-six patients were excluded due to colitis type or immediate treatment changes post-SBCE. This study focused on the primary and secondary endpoints, including the cumulative stricture-free rate of the GI tract, emergency hospitalization post-SBCE, and post-SBCE treatment strategies, at the discretion of the attending physicians. Results: Demographic data revealed that the mean age of the study population was 43 years and that there was a male predominance (75%). The median disease duration was 12 years and the mean Crohn’s Disease Activity Index was 98. During a 1,486-day observation period, 37% of patients experienced treatment changes. A Lewis score of >264 and perianal lesions were identified as independent risk factors for additional treatment needs. Emergency hospitalization occurred in 6% of patients and GI patency failure in 11%. Female sex and Lewis score>264 were associated with higher risks. GI patency rate declined 2 years after SBCE. Conclusions: For patients who experienced no treatment changes based on SBCE results, it is recommended to undergo SBCE monitoring at intervals of no longer than 2 years.

A Study Of Small Bowel Lesions Using Computed Tomography Enterography

A Study Of Small Bowel Lesions Using Computed Tomography Enterography

Double balloon enteroscopy versus single balloon enteroscopy: A comparative study.

Device assisted enteroscopy (DAE) like the double balloon enteroscopy (DBE) and single balloon enteroscopy (SBE) are postulated to ease small bowel examination and performance of therapy. However, studies comparing the effectiveness of these 2 modalities have yielded varying results. The aim of this study is to compare the efficacy and safety of SBE and DBE. We retrospectively reviewed records of patients who underwent DBE (n = 82) or SBE (n = 45) for small bowel exam in our unit from January 2014 to January 2022. Our primary outcomes were to compare the technical success and diagnostic success rates between DBE and SBE. Our secondary outcomes were to compare the therapeutic success, and complication rates. The main indications were suspected GI bleeding (DBE 41.5% vs SBE 48.9%), iron deficiency anemia (DBE 9.8% vs SBE 4.4%) and small bowel lesions (DBE 28.0% vs SBE 44.4%) detected either from prior capsule endoscopy or radiological imaging. Majority of the enteroscopy exam was by antegrade approach (DBE 67.1% vs SBE 77.8%). We found no significant difference in the technical success (DBE 95.1% vs SBE 97.8%, P = .46), diagnostic success (DBE 69.5% vs SBE 77.8%, P = .36) and the therapeutic success rate (DBE 63.2% vs SBE 54.3%, P = .09) between the groups. Complications occurred in 1 case from each group (mucosal tear). None of the complications were major. In patients who underwent enteroscopy, the diagnostic and therapeutic performance of SBE is similar to DBE. Both procedures were safe with low complication rates.

Efficacy of serum leucine-rich alpha-2 glycoprotein in predicting findings of Crohn's disease small bowel lesion in capsule endoscopy.

Small bowel capsule endoscopy (SBCE) is an evaluation method for small bowel (SB) lesions in Crohn's disease (CD). However, the relationship between SBCE findings and the serological biomarker leucine-rich alpha-2 glycoprotein (LRG) remains unclear. We aimed to establish appropriate cutoff values of LRG to predict the presence of SB lesions in CD through SBCE. Patients with CD with SB lesions who had undergone SBCE and LRG measurements 1 month before and after the SBCE were included. The LRG values for ulcers ≥0.5 cm and other inflammatory lesions noted in SBCE were determined using the Youden Index, and the sensitivity and specificity were calculated. Additionally, the correlation between the SBCE scores (CD Activity in Capsule Endoscopy) and LRG values was evaluated. Forty patients without active colorectal lesions were included in the study. When the cutoff value of LRG for SB ulcers ≥ 0.5 cm was set at 14 μg/mL, the sensitivity was 92.3%, specificity was 81.5%, positive predictive value (PPV) was 70.6%, and negative predictive value (NPV) was 95.7%. In contrast, an LRG cutoff value of 12 μg/mL without inflammatory findings had a sensitivity of 91.7%, specificity of 82.1%, PPV of 68.8%, and NPV of 95.8%. CD Activity in Capsule Endoscopy correlated well with LRG values (Spearman's rank correlation coefficient ρ = 0.681, P< 0.001). An LRG cutoff value of 14 μg/mL may be useful in predicting the presence of SB ulcers ≥ 0.5 cm, and an LRG cutoff value of 12 μg/mL may be useful in predicting the absence of SB inflammatory findings.

Sennoside A+B Is as Effective as Polyethylene Glycol in Preparation for Small Intestine Capsule Endoscopy.

We aimed to compare the effectiveness of the polyethylene glycol (PEG) and sennoside A+B regimens after clear fluid diet and fasting in bowel preperation of capsule endoscopy. In this retrospective single-center study, patients who were consecutively examined with small bowel capsule endoscopy (SBCE) between May 2010 and March 2023 were evaluated. Patients who underwent PEG 4 L and sennoside A+B calcium 250 mL for small bowel preparation were assigned. The quality of the small bowel cleaning and the diagnostic yield in detecting of small bowel lesions were compared. Two hundred forty-two patients who underwent SBCE for various indications (PEG 74.4%, sennoside A+B 25.6%) were included in the study. The mean proximal small bowel cleaning scores was 1.97 ± 0.77 for PEG and 1.98 ± 0.04 (P = .83) for sennoside A+B; the mid small bowel cleaning scores was 1.76 ± 0.84 for PEG and 1.59 ± 0.05 (P = .108) for sennoside A+B; the mean distal small bowel cleaning scores was 1.27 ± 0.08 for PEG and 1.3 ± 0.54 (P = .805) for sennoside A+B; and the total small bowel cleaning scores was 1.66 ± 0.06 and 1.62 ± 0.04 (P = .622) for PEG and sennoside A+B, respectively. There were no significant differences regarding small bowel cleaning scores both segmentally and totally. At the same time, the diagnostic value of SBCE was similar in both groups. The effectiveness of sennoside A+B in SBCE preparation is similar to that of PEG and can be used in intestinal cleansing.

Endoscopic Diagnosis of Small Bowel Tumor.

Recent technological advances, including capsule endoscopy (CE) and balloon-assisted endoscopy (BAE), have revealed that small intestinal disease is more common than previously thought. CE has advantages, including a high diagnostic yield, discomfort-free, outpatient basis, and physiological images. BAE enabled endoscopic diagnosis and treatment in the deep small bowel. Computed tomography (CT) enterography with negative oral contrast can evaluate masses, wall thickening, and narrowing of the small intestine. In addition, enhanced CT can detect abnormalities outside the gastrointestinal tract that endoscopy cannot evaluate. Each modality has its advantages and disadvantages, and a good combination of multiple modalities leads to an accurate diagnosis. As a first-line modality, three-phase enhanced CT is preferred. If CT shows a mass, stenosis, or wall thickening, a BAE should be selected. If there are no abnormal findings on CT and no obstructive symptoms, CE should be selected. If there are significant findings in the CE, determine the indication for BAE and its insertion route based on these findings. Early diagnosis of small intestinal tumors is essential for favorable outcomes. For early diagnosis, the possibility of small bowel lesions should be considered in patients with unexplained symptoms and signs after examination of the upper and lower gastrointestinal tract.

Deep enteroscopy - what's new and what works best?

We aim to review the types of device-assisted deep enteroscopy endoscopes, tips for a successful procedure as well as areas for improvement. Deep enteroscopy allows for diagnostic and therapeutic intervention of the small bowel and can be used as an adjunct to video capsule endoscopy to improve the yield and management of small bowel lesions. Our top tips for deep enteroscopy success include reviewing patient history and prior imaging, utilizing CO 2 insufflation or water exchange, verifying quality measures and emphasizing ergonomics. With these, endoscopists can optimize patient outcomes while minimizing occupational risks. New aspects of deep enteroscopy equipment focus on high-resolution imaging, a larger working channel, and enhanced scope angulation. Proposed improvements include developing innovative technology to optimize the color and clarity of the high-resolution imaging, minimizing the number of staff required for the procedure, and decreasing ergonomic strain.

Automated detection of small bowel lesions based on capsule endoscopy using deep learning algorithm

BackgroundIn order to overcome the challenges of lesion detection in capsule endoscopy (CE), we improved the YOLOv5-based deep learning algorithm and established the CE-YOLOv5 algorithm to identify small bowel lesions captured by CE. MethodsA total of 124,678 typical abnormal images from 1,452 patients were enrolled to train the CE-YOLOv5 model. Then 298 patients with suspected small bowel lesions detected by CE were prospectively enrolled in the testing phase of the study. Small bowel images and videos from the above 298 patients were interpreted by the experts, non-experts and CE-YOLOv5, respectively. ResultsThe sensitivity of CE-YOLOv5 in diagnosing vascular lesions, ulcerated/erosive lesions, protruding lesions, parasite, diverticulum, active bleeding and villous lesions based on CE videos was 91.9 %, 92.2 %, 91.4 %, 93.1 %, 93.3 %, 95.1 %, and 100 % respectively. Furthermore, CE-YOLOv5 achieved specificity and accuracy of more than 90 % for all lesions. Compared with experts, the CE-YOLOv5 showed comparable overall sensitivity, specificity and accuracy (all P > 0.05). Compared with non-experts, the CE-YOLOv5 showed significantly higher overall sensitivity (P < 0.0001) and overall accuracy (P < 0.0001), and a moderately higher overall specificity (P = 0.0351). Furthermore, the time for AI-reading (5.62 ± 2.81 min) was significantly shorter than that for the other two groups (both P < 0.0001). ConclusionsCE-YOLOv5 diagnosed small bowel lesions in CE videos with high sensitivity, specificity and accuracy, providing a reliable approach for automated lesion detection in real-world clinical practice.

Reading of small bowel capsule endoscopy after frame reduction using an artificial intelligence algorithm

ObjectivesPoorly visualized images that appear during small bowel capsule endoscopy (SBCE) can confuse the interpretation of small bowel lesions and increase the physician’s workload. Using a validated artificial intelligence (AI) algorithm that can evaluate the mucosal visualization, we aimed to assess whether SBCE reading after the removal of poorly visualized images could affect the diagnosis of SBCE.MethodsA study was conducted to analyze 90 SBCE cases in which a small bowel examination was completed. Two experienced endoscopists alternately performed two types of readings. They used the AI algorithm to remove poorly visualized images for the frame reduction reading (AI user group) and conducted whole frame reading without AI (AI non-user group) for the same patient. A poorly visualized image was defined as an image with < 50% mucosal visualization. The study outcomes were diagnostic concordance and reading time between the two groups. The SBCE diagnosis was classified as Crohn’s disease, bleeding, polyp, angiodysplasia, and nonspecific finding.ResultsThe final SBCE diagnoses between the two groups showed statistically significant diagnostic concordance (k = 0.954, p < 0.001). The mean number of lesion images was 3008.5 ± 9964.9 in the AI non-user group and 1401.7 ± 4811.3 in the AI user group. There were no cases in which lesions were completely removed. Compared with the AI non-user group (120.9 min), the reading time was reduced by 35.6% in the AI user group (77.9 min).ConclusionsSBCE reading after reducing poorly visualized frames using the AI algorithm did not have a negative effect on the final diagnosis. SBCE reading method integrated with frame reduction and mucosal visualization evaluation will help improve AI-assisted SBCE interpretation.