Predictors and characteristics of angioectasias in patients with obscure gastrointestinal bleeding identified by video capsule endoscopy

Clinical usefulness and acceptability of small-bowel capsule endoscopy with panoramic imaging compared with axial imaging in Japanese patients

Objectives

We aimed to evaluate the usefulness and acceptability of CapsoCam Plus (CapsoCam) in Japanese patients.

Methods

This retrospective single-center study enrolled 930 patients with suspected small-bowel bleeding (SSBB) who underwent capsule endoscopy. Thirty-three patients using CapsoCam and PillCam SB3 (SB3) were matched using propensity score matching. The diagnostic yield and the acceptability of CapsoCam were evaluated.

Results

There was no SSBB case where capsule endoscopy was performed within 48 h of bleeding. CapsoCam had a significantly higher observation rate of the entire small bowel (97% vs. 73%, p = 0.006) and Vater's papilla (82% vs. 15%, p < 0.001) than SB3. The reading time of CapsoCam was significantly longer than that of SB3 (30 vs. 25 min, p < 0.001), and CapsoCam's time from the capsule endoscopy swallowing to read completion was longer than that of SB3 (37 vs. 12 h, p < 0.001). The two groups showed no difference in the capsule endoscopy findings according to the P classification. Notably, 85% of the patients using CapsoCam reported examination distress as "not at all" or "almost not," and 94% reported swallowing difficulty as "very easy" or "easy."

Conclusions

CapsoCam took time to read; however, it is a well-tolerated examination with a high observation rate of Vater's papilla and entire small-bowel mucosa. Detectability of bleeding sources was comparable in both modalities for cases of occult SSBB and overt SSBB more than 48 h after bleeding. CapsoCam is a useful modality for patients with SSBB.

The professional background of a referring physician predicts the diagnostic yield of small bowel capsule endoscopy in suspected small bowel bleeding

Background and study aims The diagnostic yield of small-bowel capsule endoscopy (SBCE) in suspected small bowel bleeding (SSBB) is highly variable. Different reimbursement systems and equipment costs also limit SBCE use in clinical practice. Thus, minimizing non-diagnostic procedures is advisable. This study aimed to assess the SBCE diagnostic yield and identify factors predicting diagnostic findings in a cohort of patients with SSBB. Patients and methods In this retrospective cohort study, we analyzed the medical records of patients who consecutively underwent SBCE for SSBB over 9 years. By logistic regression, we identified covariates predicting diagnostic findings at SBCE. Finally, we performed a post-hoc cost analysis based on previous gastroenterologist or endoscopist consultations versus direct SBCE ordering by other specialists. Results The final analysis included 584 patients. Most SBCEs were ordered by a gastroenterologist or endoscopist (74%). The number of SBCEs without any finding was significantly lower in the gastroenterologist/endoscopist group P <0.001). The SBCE diagnostic yield ordered by a gastroenterologist or endoscopist was significantly higher than that by other specialists (63% vs 52%, odds ratio [OR] 1.57; 95% confidence interval [CI] 1.07-2.26, P =0.019). At multivariate analysis, older age (OR 1.7, 95%CI 1.2-2.4, P =0.005), anemia (OR 4.9, 95%CI 1.9-12, P =0.001), small bowel transit time (OR 1, 95%CI 1-1.02, P =0.039), and referring physician (OR 1.8, 95%CI 1.1-2.7, P =0.003) independently predicted diagnostic findings. Implementing prior gastroenterologist or endoscopist referral vs direct SBCE ordering would reduce medical expenditures by 16%. Conclusions The professional background of referring physicians significantly improves the diagnostic yield of SBCE and contributes to controlling public health costs.

Automatic detection of small bowel lesions with different bleeding risks based on deep learning models

BACKGROUND

Deep learning provides an efficient automatic image recognition method for small bowel (SB) capsule endoscopy (CE) that can assist physicians in diagnosis. However, the existing deep learning models present some unresolved challenges.

AIM

To propose a novel and effective classification and detection model to automatically identify various SB lesions and their bleeding risks, and label the lesions accurately so as to enhance the diagnostic efficiency of physicians and the ability to identify high-risk bleeding groups.

METHODS

The proposed model represents a two-stage method that combined image classification with object detection. First, we utilized the improved ResNet-50 classification model to classify endoscopic images into SB lesion images, normal SB mucosa images, and invalid images. Then, the improved YOLO-V5 detection model was utilized to detect the type of lesion and its risk of bleeding, and the location of the lesion was marked. We constructed training and testing sets and compared model-assisted reading with physician reading.

RESULTS

The accuracy of the model constructed in this study reached 98.96%, which was higher than the accuracy of other systems using only a single module. The sensitivity, specificity, and accuracy of the model-assisted reading detection of all images were 99.17%, 99.92%, and 99.86%, which were significantly higher than those of the endoscopists' diagnoses. The image processing time of the model was 48 ms/image, and the image processing time of the physicians was 0.40 ± 0.24 s/image (P < 0.001).

CONCLUSION

The deep learning model of image classification combined with object detection exhibits a satisfactory diagnostic effect on a variety of SB lesions and their bleeding risks in CE images, which enhances the diagnostic efficiency of physicians and improves the ability of physicians to identify high-risk bleeding groups.

Predictive Model for Positive Video Capsule Endoscopy in Iron Deficiency Anemia

BACKGROUND AND AIMS

Bleeding from the gastrointestinal tract can contribute to the development of iron deficiency anemia (IDA) among individuals without another obvious source of bleeding. In order to identify patients most likely to benefit from examination of the small bowel, our aim was to create a risk score for positive video capsule endoscopy (VCE) in IDA utilizing a multicenter collection of studies.

METHODS

We performed a retrospective multicenter study utilizing VCE studies performed for an indication of IDA between 1/1/2005 and 7/31/2018. VCE findings were graded based on the P0-P2 grading system. The primary outcome of interest was a positive (P2) VCE. Data were analyzed with Student's t test for continuous variables and the Fisher's exact test for categorical variables. Logistic regression was used to identify independent associations with positive VCE.

RESULTS

In total, 765 VCE procedures were included with 355 (46.5%) male subjects and a median age of 63.2 (SD 15.3) years. One hundred ninety studies (24.8%) were positive (P2) for small bowel bleeding. Four variables associated with positive VCE which were incorporated into a point scoring system: (+) 1 for age ≥ 66 years, active smoking and cardiac arrythmia and (-) 1 for preceding hemoglobin level ≥ 8.5. The risk probabilities for positive VCE-assigned scores - 1, 0, 1, and 2 + were 12.3% (95% CI 7.3-17.3%), 20% (14.9-25.1%), 34.8% (28.6-41%), and 39% (30-47.8%).

CONCLUSION

In order to improve the diagnostic yield of capsule examinations, risk factors should be applied to clinical decision-making. We created a risk score for positive VCE in IDA, including the risk factors of age, smoking, history of cardiac arrythmia, and preceding hemoglobin level.

Artificial intelligence and capsule endoscopy: automatic detection of vascular lesions using a convolutional neural network

Background

Capsule endoscopy (CE) is the first line for evaluation of patients with obscure gastrointestinal bleeding. A wide range of small intestinal vascular lesions with different hemorrhagic potential are frequently found in these patients. Nevertheless, reading CE exams is time-consuming and prone to errors. Convolutional neural networks (CNN) are artificial intelligence tools with high performance levels in image analysis. This study aimed to develop a CNN-based model for identification and differentiation of vascular lesions with distinct hemorrhagic potential in CE images.

Methods

The development of the CNN was based on a database of CE images. This database included images of normal small intestinal mucosa, red spots, and angiectasia/varices. The hemorrhagic risk was assessed by Saurin's classification. For CNN development, 11,588 images (9525 normal mucosa, 1026 red spots, and 1037 angiectasia/varices) were ultimately extracted. Two image datasets were created for CNN training and testing.

Results

The network was 91.8% sensitive and 95.9% specific for detection of vascular lesions, providing accurate predictions in 94.4% of cases. In particular, the CNN had a sensitivity and specificity of 97.1% and 95.3%, respectively, for detection of red spots. Detection of angiectasia/varices occurred with a sensitivity of 94.1% and a specificity of 95.1%. The CNN had a frame reading rate of 145 frames/sec.

Conclusions

The developed algorithm is the first CNN-based model to accurately detect and distinguish enteric vascular lesions with different hemorrhagic risk. CNN-assisted CE reading may improve the diagnosis of these lesions and overall CE efficiency.

Deep learning and capsule endoscopy: automatic identification and differentiation of small bowel lesions with distinct haemorrhagic potential using a convolutional neural network

OBJECTIVE

Capsule endoscopy (CE) is pivotal for evaluation of small bowel disease. Obscure gastrointestinal bleeding most often originates from the small bowel. CE frequently identifies a wide range of lesions with different bleeding potentials in these patients. However, reading CE examinations is a time-consuming task. Convolutional neural networks (CNNs) are highly efficient artificial intelligence tools for image analysis. This study aims to develop a CNN-based model for identification and differentiation of multiple small bowel lesions with distinct haemorrhagic potential using CE images.

DESIGN

We developed, trained, and validated a denary CNN based on CE images. Each frame was labelled according to the type of lesion (lymphangiectasia, xanthomas, ulcers, erosions, vascular lesions, protruding lesions, and blood). The haemorrhagic potential was assessed by Saurin's classification. The entire dataset was divided into training and validation sets. The performance of the CNN was measured by the area under the receiving operating characteristic curve, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV).

RESULTS

A total of 53 555 CE images were included. The model had an overall accuracy of 99%, a sensitivity of 88%, a specificity of 99%, a PPV of 87%, and an NPV of 99% for detection of multiple small bowel abnormalities and respective classification of bleeding potential.

CONCLUSION

We developed and tested a CNN-based model for automatic detection of multiple types of small bowel lesions and classification of the respective bleeding potential. This system may improve the diagnostic yield of CE for these lesions and overall CE efficiency.

A comprehensive systematic review and meta-analysis of risk factors for rebleeding following device-assisted enteroscopy therapy of small-bowel vascular lesions

INTRODUCTION

the aim of this study was to determine the risk factors for rebleeding following device-assisted enteroscopy therapy of small bowel vascular lesions.

METHODS

this is a systematic review and meta-analysis. A literature search was performed from January 2003 to October 2019. All studies reporting on at least one risk factor for bleeding recurrence after endoscopic therapy of small bowel vascular lesions were included. A meta-analysis of those risk factors reported in at least three studies was performed to assess their association with rebleeding. The OR and 95 % CI were used for binary outcome data. Heterogeneity analysis was performed using the Tau and I2 index. If I2 > 20 %, potential sources of heterogeneity were identified by sensitivity analyses and a random-effect model was used.

RESULTS

the search identified a total of 572 articles and 35 full-text records were assessed for eligibility after screening. Finally, eight studies that included 548 patients were selected. The overall median rebleeding rate was 38.5 % (range: 10.9-53.3 %) with a median follow-up of 24.5 months. Female sex (OR: 1.96, 95 % CI: 1.14-3.37, p = 0.01, I2 = 0 %), Osler-Weber syndrome (OR: 4.35, 95 % CI: 1.22-15.45, p = 0.02, I2 = 0 %) and cardiac disease (OR: 1.89, 95 % CI: 1.12-2.97, p = 0.005, I2: 0 %) were associated with rebleeding. According to the sensitivity analysis, overt bleeding (OR: 2.13, 95 % CI: 1.22-3.70, p = 0.007, I2 = 0 %), multiple lesions (OR: 4.57, 95 % CI: 2.04-10.22, p < 0.001, I2 = 0 %) and liver cirrhosis (OR: 2.61, 95 % CI: 1.11-6.13, p = 0.03, I2 = 0 %) were also predictors for rebleeding.

CONCLUSIONS

patient characteristics and comorbidities should be considered for follow-up patient management after effective device-assisted endoscopic therapy, as they can predict rebleeding.

Scoring systems in clinical small-bowel capsule endoscopy: all you need to know!

Capsule endoscopy (CE) emerged out of the pressing clinical need to image the small bowel (SB) in cases of midgut bleeding and provide an overall comfortable and reliable gastrointestinal (GI) diagnosis 1. Since its wider adoption in clinical practice, significant progress has been made in several areas including software development, hardware features and clinical indications, while innovative applications of CE never cease to appear 2 3. Currently, several manufacturers provide endoscopic capsules with more or less similar technological features 4. Although there is engaging and continuous academic and industry-fueled R&D, promising furtherment of CE technology 4 5, the current status of clinical CE remains that of by and large an imaging modality. Clinical relevance of CE images is cornerstone in the decision-making process for medical management. In one of the larger to date SB CE studies, 4,206 abnormal images were detected in 3,280 patients 6. Thus, CE leads to the identification of a large amount of potential pathology, some of which are pertinent (or relevant) while some (probably the majority) are not.

Soon artificial intelligence (AI) is likely to carry out several roles currently performed by humans; in fact, we are witnessing only the first stages of a transition in the clinical adoption of AI-based solutions in several aspects of gastroenterology including CE 7. Until then though, human-based decision-making profoundly impacts patient care and – although not suggested in the updated European Society of Gastrointestinal Endoscopy (ESGE) European curriculum 8 9 – it should be an integral part of CE training. Frequently, interpretation of CE images by experts or at least experienced readers differs. In a tandem CE reading study, expert review of discordant cases revealed a 50 % (13/25 discordant results) error rate by experienced readers, corresponding (in 5/13 cases) to ‘over-classification’ of an irrelevant abnormality 10. Another comparative study showed an ‘over-classification’ of such irrelevant abnormalities in ~10 % of CE readings 11. One thing which has been for a while on the table – in relation to optimizing and/or standardizing CE reporting and subsequent decision-making – is the need for reproducible scoring systems and for a reliable common language among clinicians responsible for further patient’s management.

Over the years, several of these scoring systems were developed while others appear in the wake of software and hardware improvements aiming to replace and/or complement their predecessors. This review presents a comprehensive account of the currently available classification/scoring systems in clinical CE spanning from predicting the bleeding potential of identified SB lesions (with emphasis on vascular lesions), and the individual rebleeding risk; scoring systems for the prediction of SB lesions in patients with obscure gastrointestinal bleeding (OGlB), having the potential to improve patient selection and rationalize the use of enteroscopy, with better allocation of resources, optimized diagnostic workflow and tailored treatment. This review also includes scores for reporting the inflammatory burden, the cleansing level that underscores confidence in CE reporting and the mass or bulge question in CE. Essentially, the aim is to become a main text for reference when scoring is required and facilitate the inclusion of -through readiness of access- one of the other in the final report.

Role of Video Capsule Endoscopy as a Prelude to Deep Enteroscopy

Video capsule endoscopy and device-assisted enteroscopy are complementary technologies. Capsule endoscopy is a highly acceptable technology with high diagnostic yield that can guide a subsequent enteroscopy approach. This article aims to focus on the role of video capsule endoscopy as a prelude to deep enteroscopy with a focus on the strengths and limitations of either approach.

Small Bowel Angioectasias Rebleeding and the Identification of Higher Risk Patients

BACKGROUND

Small bowel capsule endoscopy (SBCE) is the gold standard for suspected small bowel bleeding (SBB). Angioectasias are the most common vascular anomalies in the gastrointestinal tract and have been reported as the source of SBB in up to 80% of patients. Considering their frequency, their usual intermittent bleeding nature, and their risk of rebleeding, the aim of this study was to identify some features and possible predictors of rebleeding in the presence of these lesions.

METHODS

This is a retrospective study, which included consecutive SBCE with angioectasias between April 2008 and December 2017 with a minimum follow-up of 12 months. Rebleeding was defined as a drop of hemoglobin ≥ 2 g/dl and/or in the presence of hematochezia or melenas with negative esophagogastroduodenoscopy and ileocolonoscopy. Data were collected from medical records, and angioectasias were classified by number, location, size, and type. Univariate and multivariable statistical analysis was performed to identify possible predictors of rebleeding.

RESULTS

From a total of 630 patients submitted to SBCE for suspected SBB, 129 with angioectasias were included; 59.7% were female, with a median age of 72 (19-91) years old and a mean follow-up of 44.0 ± 31.9 months. In 32.6% (n = 42) of the patients, at least one episode of rebleeding was documented. The presence of heart failure (OR 3.41; IC95% 1.18-9.89; p = 0.024), the size of the angioectasias (OR 5.41; IC95% 2.15-13.6; p < 0.001), and smoking status (OR 3.15; IC95% 1.07-9.27; p = 0.038) were independent predictor factors of rebleeding.

CONCLUSION

Heart failure, smoking status, and angioectasias with a size superior to 5 mm are independent predictor factors of rebleeding in a population with angioectasias diagnosed by SBCE.

Causes and Clinical Characteristics of Small Bowel Bleeding in Northern Vietnam

AIM

Causes, clinical features, and diagnostic approaches for small bowel (SB) bleeding were analyzed to derive recommendations in dealing with this clinical condition.

METHODS

We included 54 patients undergoing surgical treatment for SB bleeding, from January 2009 to December 2019. Detailed clinical data, diagnosis procedures, and causes of bleeding were collected.

RESULTS

Among 54 cases with SB bleeding, the most common causes were tumors (64.8%), followed by angiopathy (14.8%), ulcers (9.3%), diverticula (5.6%), tuberculosis (3.7%), and enteritis (1.9%). Most tumors (32/35 cases, 91.4%) and vascular lesions (8/8 cases, 100%) were located in the jejunum. The incidence of tumors was higher in the older (30/41 cases, 73.1%) than that in patients younger than 40 years of age (5/13 cases, 38.5%, P < 0.01). Common initial findings were melena (68.5%) and hematochezia (31.5%). The overall diagnostic yield of computed tomographic enterography (CTE) was 57.4% (31/54 cases), with the figures for tumors, vascular lesions, and inflammatory lesions being 71.4% (25/35 cases), 62.5% (5/8 cases), and 12.5% (1/8 cases), respectively. Double-balloon enteroscopy (DBE) definitively identified SB bleeding sources in 16/22 (72.7%) patients.

CONCLUSION

Tumors, angiopathy, ulcers, and diverticula were the most common causes of SB bleeding in Northern Vietnamese population. CTE has a high detection rate for tumors in patients with SB bleeding. CTE as a triage tool may identify patients before double-balloon enteroscopy because of the high prevalence of SB tumors.

Mapping the distribution of small bowel angioectasias

Background: Angioectasias are a prominent cause of small bowel (SB) bleeding frequently identified during capsule endoscopy (CE). Subsequent management depends upon grade/severity and location. There is increasing evidence that the location of SB angioectasias is not random. We aimed to map the distribution of SB angioectasias, and assess whether this impacted clinical outcomes. Materials and methods: Retrospective study examining CEs performed over a 10-year period at a tertiary referral centre. Information regarding number, location, and Saurin classification (P0-2) of SB angioectasias was collected. Clinically significant angioectasias (P1/P2) and active SB bleeding were analysed further. Outcomes of patients with P2 angioectasias or active SB bleeding were recorded. Results: 164 SBCE examinations reported angioectasias. 554 P1-2 angioectasias and active bleeds were seen, 435 (78.52%) within the first tertile of SB transit time (SBTT). 277 (50%) angioectasias were identified within the first 10% of SBTT. 40/75 (53.3%) patients with >1 P2 angioectasia and/or active bleed were referred for intervention. Of initial interventions, 24 patients underwent upper GI endoscopy; 13 underwent double balloon enteroscopy (DBE). 9/37(24.3%) had no identifiable angioectasias on endoscopy. Of those receiving ablative therapy, 20/28 (71.4%) re-presented with iron-deficiency anaemia or bleeding. In this group, average angioectasia position was 15.6% of SBTT, compared with 7.9% in those who did not re-represent (p = 0.344). Patients who re-presented had an average 1.6 additional P1 angioectasias, compared with 7.6 amongst those who did not return (p = 0.017). Conclusions: Clinically significant angioectasias are overwhelmingly located within the proximal SB. The majority are within reach of conventional endoscopy. However, AEs are often multiple and many patients re-present following intervention.

The Spectrum of Small Intestinal Lesions in Patients with Unexplained Iron Deficiency Anemia Detected by Video Capsule Endoscopy

Background and objectives: Video-capsule endoscopy (VCE) has shown a large range (38⁻83%) of diagnostic yield in unexplained iron deficiency anemia (IDA) and obscure-occult bleeding. Therefore, we retrospectively investigated the VCE-detected spectrum and the prevalence of small bowel injuries and associated risk factors in inpatients with both of the above reported conditions. Methods: We selected inpatients with IDA (hemoglobin <12 g/dL in women, <13 g/dL in men) and obscure-occult bleeding. We excluded VCE indications other than IDA. Complete medical histories and laboratory tests were collected. All subjects underwent PillCam SB2/SB3. The VCE feature Lewis score was calculated when appropriate. We used the t-test and Fisher's exact test for continuous and categorical variables, respectively, in univariate analysis. For multivariate analysis, we used binomial logistic regression. Results: We retrieved 109 patients (female:male ratio of 53:56; age 63.4 ± 18.9 years). Eighty patients (73.4%) showed ≥1 small bowel lesions. The Lewis score was calculated in 41 patients: 13 (31.7%) showed a mild (<135) and 28 (68.3%) a moderate-severe (135⁻790 and >790, respectively) score. In univariate analysis, the small bowel transit time (6.2 ± 2.9 versus 5.2 ± 2.1 h; p = 0.049) and non-steroidal anti-inflammatory drug use for at least two weeks (17.5% versus 0%; p = 0.01) were significantly higher in subjects with injuries. These associations were not confirmed at multivariate analysis. The severity of a lesion directly correlated with proton pump inhibitor (PPI) use and duration (not confirmed in multivariate analysis). VCE can reveal the source of obscure-occult bleeding in a high percentage of unexplained IDAs. A wide spectrum of endoscopic pictures may be found. Known as well as supposed risk factors for small bowel lesions may be detected.