Efficacy of new playback functions at reducing small-bowel wireless capsule endoscopy reading times

Diagnostic accuracy of artificial intelligence-aided capsule endoscopy (TOP100) in overt small bowel bleeding

BACKGROUND

Capsule endoscopy (CE) is the first-choice exploration in case of overt small bowel bleeding (SBB). An early CE is known to increase diagnostic yield, but long reading times may delay therapeutics. The study evaluates the diagnostic performance of the artificial intelligence tool TOP100 in patients with overt SBB undergoing early CE with Pillcam SB3.

METHODS

Patients who underwent early CE (up to 14 days from the bleeding episode) for suspected overt SBB were included. One experienced endoscopist prospectively performed standard reading (SR) and a second blind experienced endoscopist performed a TOP100-based reading (TR). The primary endpoint was TR diagnostic accuracy for lesions with high bleeding potential (P2).

RESULTS

A total of 111 patients were analyzed. The most common clinical presentation was melena (64%). CE showed angiodysplasias in 40.5% of patients (45/111). In per-patient analysis, TR showed a sensitivity of 90.48% (95% CI 82.09-95.80), specificity of 100% (95% CI 87.23-100) with a PPV of 100% (95% CI 94.01-100), NPV of 77.14% (95% CI 63.58-86.71) and diagnostic accuracy of 92.79 (86.29-96.84). At multivariate analysis, adequate intestinal cleansing was the only independent predictor of concordance between TR and SR (OR 2.909, p = 0.019). The median reading time for SR and TR was 23 min (18.0-26.8) and 1.9 min (range 1.7-2.1), respectively (p < 0.001).

CONCLUSIONS

TOP100 provides a fast-reading mode for early CE in case of overt small bowel bleeding. It identifies most patients with active bleeding and angiodysplasias, aiding in the prioritization of therapeutic procedures. However, its accuracy in detecting ulcers, varices and P1 lesions seems insufficient.

Computer vision-based solutions to overcome the limitations of wireless capsule endoscopy

Endoscopic investigation plays a critical role in the diagnosis of gastrointestinal (GI) diseases. Since 2001, Wireless Capsule Endoscopy (WCE) has been available for small bowel exploration and is in continuous development. Over the last decade, WCE has achieved impressive improvements in areas such as miniaturisation, image quality and battery life. As a result, WCE is currently a very useful alternative to wired enteroscopy in the investigation of various small bowel abnormalities and has the potential to become the leading screening technique for the entire gastrointestinal tract. However, commercial solutions still have several limitations, namely incomplete examination and limited diagnostic capacity. These deficiencies are related to technical issues, such as image quality, motion estimation and power consumption management. Computational methods, based on image processing and analysis, can help to overcome these challenges and reduce both the time required by reviewers and human interpretation errors. Research groups have proposed a series of methods including algorithms for locating the capsule or lesion, assessing intestinal motility and improving image quality.In this work, we provide a critical review of computational vision-based methods for WCE image analysis aimed at overcoming the technological challenges of capsules. This article also reviews several representative public datasets used to evaluate the performance of WCE techniques and methods. Finally, some promising solutions of computational methods based on the analysis of multiple-camera endoscopic images are presented.

Quality indicators for capsule endoscopy and deep enteroscopy

BACKGROUND AND AIMS

Capsule endoscopy (CE) and deep enteroscopy (DE) can be useful for diagnosing and treating suspected small-bowel disease. Guidelines and detailed recommendations exist for the use of CE/DE, but comprehensive quality indicators are lacking. The goal of this task force was to develop quality indicators for appropriate use of CE/DE by using a modified RAND/UCLA Appropriateness Method.

METHODS

An expert panel of 7 gastroenterologists with diverse practice experience was assembled to identify quality indicators. A literature review was conducted to develop a list of proposed quality indicators applicable to preprocedure, intraprocedure, and postprocedure periods. The panelists reviewed the literature; identified and modified proposed quality indicators; rated them on the basis of scientific evidence, validity, and necessity; and determined proposed performance targets. Agreement and consensus with the proposed indicators were verified using the RAND/UCLA Appropriateness Method.

RESULTS

The voting procedure to prioritize metrics emphasized selecting measures to improve quality and overall patient care. Panelists rated indicators on the perceived appropriateness and necessity for clinical practice. After voting and discussion, 2 quality indicators ranked as inappropriate or uncertain were excluded. Each quality indicator was categorized by measure type, performance target, and summary of evidence. The task force identified 13 quality indicators for CE and DE.

CONCLUSIONS

Comprehensive quality indicators have not existed for CE or DE. The task force identified quality indicators that can be incorporated into clinical practice. The panel also addressed existing knowledge gaps and posed research questions to better inform future research and quality guidelines for these procedures.

Quality Indicators for Capsule Endoscopy and Deep Enteroscopy

INTRODUCTION

Capsule endoscopy (CE) and deep enteroscopy (DE) can be useful for diagnosing and treating suspected small-bowel disease. Guidelines and detailed recommendations exist for the use of CE/DE, but comprehensive quality indicators are lacking. The goal of this task force was to develop quality indicators for appropriate use of CE/DE by using a modified RAND/UCLA Appropriateness Method.

METHODS

An expert panel of 7 gastroenterologists with diverse practice experience was assembled to identify quality indicators. A literature review was conducted to develop a list of proposed quality indicators applicable to preprocedure, intraprocedure, and postprocedure periods. The panelists reviewed the literature; identified and modified proposed quality indicators; rated them on the basis of scientific evidence, validity, and necessity; and determined proposed performance targets. Agreement and consensus with the proposed indicators were verified using the RAND/UCLA Appropriateness Method.

RESULTS

The voting procedure to prioritize metrics emphasized selecting measures to improve quality and overall patient care. Panelists rated indicators on the perceived appropriateness and necessity for clinical practice. After voting and discussion, 2 quality indicators ranked as inappropriate or uncertain were excluded. Each quality indicator was categorized by measure type, performance target, and summary of evidence. The task force identified 13 quality indicators for CE and DE.

DISCUSSION

Comprehensive quality indicators have not existed for CE or DE. The task force identified quality indicators that can be incorporated into clinical practice. The panel also addressed existing knowledge gaps and posed research questions to better inform future research and quality guidelines for these procedures.

Development of a Deep-Learning Algorithm for Small Bowel-Lesion Detection and a Study of the Improvement in the False-Positive Rate

Deep learning has recently been gaining attention as a promising technology to improve the identification of lesions, and deep-learning algorithms for lesion detection have been actively developed in small-bowel capsule endoscopy (SBCE). We developed a detection algorithm for abnormal findings by deep learning (convolutional neural network) the SBCE imaging data of 30 cases with abnormal findings. To enable the detection of a wide variety of abnormal findings, the training data were balanced to include all major findings identified in SBCE (bleeding, angiodysplasia, ulceration, and neoplastic lesions). To reduce the false-positive rate, "findings that may be responsible for hemorrhage" and "findings that may require therapeutic intervention" were extracted from the images of abnormal findings and added to the training dataset. For the performance evaluation, the sensitivity and the specificity were calculated using 271 detectable findings in 35 cases. The sensitivity was calculated using 68,494 images of non-abnormal findings. The sensitivity and specificity were 93.4% and 97.8%, respectively. The average number of images detected by the algorithm as having abnormal findings was 7514. We developed an image-reading support system using deep learning for SBCE and obtained a good detection performance.

Video capsule endoscopy: pushing the boundaries with software technology

Video capsule endoscopy (VCE) has transformed imaging of the small bowel as it is a non-invasive and well tolerated modality with excellent diagnostic capabilities. The way we read VCE has not changed much since its introduction nearly two decades ago. Reading is still very time intensive and prone to reader error. This review outlines the evidence regarding software enhancements which aim to address these challenges. These include the suspected blood indicator (SBI), automated fast viewing modes including QuickView, lesion characterization tools such Fuji Intelligent Color Enhancement, and three-dimensional (3D) representation tools. We also outline the exciting new evidence of artificial intelligence (AI) and deep learning (DL), which promises to revolutionize capsule reading. DL algorithms have been developed for identifying organs of origin, intestinal motility events, active bleeding, coeliac disease, polyp detection, hookworms and angioectasias, all with impressively high sensitivity and accuracy. More recently, an algorithm has been created to detect multiple abnormalities with a sensitivity of 99.9% and reading time of only 5.9 minutes. These algorithms will need to be validated robustly. However, it will not be long before we see this in clinical practice, aiding the clinician in rapid and accurate diagnosis.

Use of rapid reading software to reduce capsule endoscopy reading times while maintaining accuracy

BACKGROUNDS AND AIMS

A typical capsule endoscopy (CE) case generates tens of thousands of images, with abnormalities often confined to a just few frames. Omni Mode is a novel EndoCapsule software algorithm (Olympus, Tokyo, Japan) that proposes to intelligently remove duplicate images while maintaining accuracy in lesion detection.

METHODS

This prospective multicenter study took place across 9 European centers. Consecutive, unselected CE cases were read conventionally in normal mode, with every captured frame reviewed. Cases were subsequently anonymized and randomly allocated to another center where they were read using Omni Mode. Detected lesions and reading times were recorded, with findings compared between both viewing modes. The clinical significance of lesions was described according to the P classification (P0, P1, and P2). Where a discrepancy in lesion detection in either mode was found, expert blinded review at a consensus meeting was undertaken.

RESULTS

The patient population undergoing CE had a mean age of 49.5 years (range, 18-91), with the investigation of anemia or GI bleeding accounting for 71.8% of cases. The average small-bowel transit time was 4 hours, 26 minutes. The mean reading time in normal mode was 42.5 minutes. The use of Omni Mode was significantly faster (P < .0001), with an average time saving of 24.6 minutes (95% confidence interval, 22.8-26.9). The 2127 lesions were identified and classified according to the P classification as P0 (1234), P1 (656), and P2 (237). Lesions were identified using both reading modes in 40% (n = 936), and 1186 lesions were identified by either normal or Omni Mode alone. Normal mode interpretation was associated with 647 lesions being missed, giving an accuracy of .70. Omni Mode interpretation led to 539 lesions being missed, with an accuracy of .75. There was no significant difference in clinical conclusions made between either reading mode.

CONCLUSIONS

This study shows that CE reading times can be reduced by an average of 40%, without any reduction in clinical accuracy.

Capsule endoscopy for small-intestinal disorders: Current status

Small-bowel capsule endoscopy (SBCE) is used widely because of its non-invasive and patient-friendly nature. SBCE can visualize entire small-intestinal mucosa and facilitate detection of small-intestinal abnormalities. In this review article, we focus on the current status of SBCE. Several platforms for SBCE are available worldwide. Third-generation SBCE (PillCam® SB3) has a high-resolution camera equipped with an adaptive frame rate system. Several software modes have been developed to reduce the reading time for capsule endoscopy and to minimize the possibility of missing lesions. The main complication of SBCE is capsule retention. Thus, the main contraindication for SBCE is known or suspected gastrointestinal obstruction unless intestinal patency is proven. Possible indications for SBCE are obscure gastrointestinal bleeding, Crohn's disease, small-intestinal polyps and tumors, and celiac disease. Colon capsule endoscopy (CCE) can observe inflamed colonic mucosa non-invasively, and allows for the continuous and non-invasive observation of the entire intestinal tract (pan-endoscopy). Recently, application of CCE as pan-enteric endoscopy for inflammatory bowel diseases (including Crohn's disease) has been reported. In the near future, reading for CE will be assisted by artificial intelligence, and reading CE videos for long periods will not be required.

Optimising the performance and interpretation of small bowel capsule endoscopy

Small bowel capsule endoscopy has become a commonly used tool in the investigation of gastrointestinal symptoms and is now widely available in clinical practice. In contrast to conventional endoscopy, there is a lack of clear consensus on when competency is achieved or the way in which capsule endoscopy should be performed in order to maintain quality and clinical accuracy. Here we explore the evidence on the key factors that influence the quality of small bowel capsule endoscopy services.

Video Capsule Endoscopy: Technology, Reading, and Troubleshooting

Video capsule endoscopy (VCE) has completed the endoscopic visualization of the entire luminal gastrointestinal tract. VCE can be performed in inpatients and outpatients, requires appropriate bowel preparation before the study, and can be administered via oral swallowing or endoscopic device placement into the small bowel based on outlined patient-dependent factors. Current commercially available VCE systems were reviewed and compared for individual features and attributes. This article focuses on preparation for VCE, currently available VCE technology, how to read a VCE study, and risks and contraindications to VCE.

Capsule endoscopy of the small bowel

Capsule endoscopy (CE) is a first line small bowel investigative modality which provides more sensitive mucosal imaging than comparators. It is a non-invasive, non-irradiating tool well tolerated by patients. The risk of retention of the capsule can be minimised by ensuring luminal patency using the Agile patency device. Research continues into how to minimise missed pathology and variability in the identification of pathology or interpretation of images. The consensus is that bowel preparation using laxatives improves visibility and diagnostic yield. Research includes the development of image recognition software, both to eliminate sequentially identical images to improve viewing speed and to select or enhance images likely to represent pathology. However, careful reading by experienced capsule endoscopists remains the benchmark. This should be performed at a speed comfortable to the viewer, probably at a maximum of 15 frames per second. Some prior experience of endoscopy appears to be helpful for novice capsule endoscopists and formal training on a hands-on training course seems to improve pathology recognition, for novices and for those with CE experience.

Evaluation of performance of the Omni mode for detecting video capsule endoscopy images: A multicenter randomized controlled trial

BACKGROUND AND STUDY AIMS

Olympus recently developed a new algorithm called Omni mode that discards redundant video capsule endoscopy (VCE) images. The current study aimed to demonstrate the non-inferiority of the Omni mode in terms of true positives (TPs) and the superiority of the Omni mode with regard to reading time against a control (ordinary ES-10 system).

PATIENTS AND METHODS

This multicenter prospective study included 40 patients with various small bowel diseases. VCE images were evaluated by 7 readers and 3 judging committee members. Two randomly allocated readers assessed the VCE images obtained using the 2 modalities for each patient. The order of the modalities was switched between the 2 readers and the interval between readings by the same reader was 2 weeks. The judging committee predefined clinically relevant lesions as major lesions and irrelevant lesions as minor lesions. The number of TPs for major and minor lesions and the reading times were compared between the modalities. The predefined non-inferiority margin for the TP ratio of the Omni mode compared with the control was 0.9.

RESULTS

The estimated TP ratios and 95 % confidence intervals for total, major, and minor lesions were 0.87 (0.80 - 0.95), 0.93 (0.83 - 1.04), and 0.83 (0.74 - 0.94), respectively. Although non-inferiority was not demonstrated, the rate of detection of major lesions was not significantly different between the modalities. The reading time was significantly lower when using the Omni mode than when using the control.

CONCLUSIONS

The Omni mode may be only appropriate for the assessment of major lesions.

Challenges and Future of Wireless Capsule Endoscopy

In 2015, capsule endoscopy was introduced as the main investigation method for small bowel mucosal diseases, and its role in colonic diseases has been gradually revealed. Future challenges for capsule endoscopy, besides improvements of image quality and visualization of each part of the small bowel and colonic mucosa, include the development of gastric capsules, the capacity to perform histological examination of the mucosa, and maybe in the future, some capsule endoscopy-driven therapeutics. The aim of this review was to evaluate the clinical demands and feasibility of achieving the aforementioned objectives.

Quantitative measurements in capsule endoscopy

This review summarizes several approaches for quantitative measurement in capsule endoscopy. Video capsule endoscopy (VCE) typically provides wireless imaging of small bowel. Currently, a variety of quantitative measurements are implemented in commercially available hardware/software. The majority is proprietary and hence undisclosed algorithms. Measurement of amount of luminal contamination allows calculating scores from whole VCE studies. Other scores express the severity of small bowel lesions in Crohn׳s disease or the degree of villous atrophy in celiac disease. Image processing with numerous algorithms of textural and color feature extraction is further in the research focuses for automated image analysis. These tools aim to select single images with relevant lesions as blood, ulcers, polyps and tumors or to omit images showing only luminal contamination. Analysis of motility pattern, size measurement and determination of capsule localization are additional topics. Non-visual wireless capsules transmitting data acquired with specific sensors from the gastrointestinal (GI) tract are available for clinical routine. This includes pH measurement in the esophagus for the diagnosis of acid gastro-esophageal reflux. A wireless motility capsule provides GI motility analysis on the basis of pH, pressure, and temperature measurement. Electromagnetically tracking of another motility capsule allows visualization of motility. However, measurement of substances by GI capsules is of great interest but still at an early stage of development.

The evaluation of the OMOM capsule endoscopy with similar pictures elimination mode

BACKGROUND AND OBJECTIVE

The reading of capsule endoscopy (CE) is time-consuming. Thus, several time-saving features have been added to the CE software by different CE companies. However, the similar pictures elimination mode in the OMOM capsule endoscopy software has not been evaluated yet. Our aim is to evaluate the sensitivity and specificity of the similar pictures elimination mode.

METHODS

We retrospectively studied 200 consecutive capsule endoscopy procedures in Nanfang Hospital between April 2012 and July 2012, among which 52 cases were excluded for poor bowel preparation, failure to reach the colon or data missing. All images of the small-bowel were re-evaluated with four different reading modes (conventional mode and three levels of similar pictures elimination mode) by four experienced endoscopists. Then, the reading time, the number of detected lesions, κ value, sensitivity and specificity were compared between 4 methods.

RESULTS

The mean reading time with levels I to III, respectively, was significantly shorter than with conventional mode (P<0.001, conventional mode 33.25min, level I 24.90min, level II 20.54min, level III 14.96min). The sensitivity between conventional mode and level I were similar (93.8% vs. 87.7%, P=0.073), while the sensitivity of levels II and III was significantly lower than conventional mode (77.8% vs. 93.8%, 70% vs. 93.8%, respectively). The specificity between all 4 modes were similar (100% for conventional mode, 98.5% for levels I to III).

CONCLUSIONS

Level I reading mode was as efficient as the conventional mode and less time-consuming. Its sensitivity and specificity were similar compared with the conventional mode.

Continuing challenges in the diagnosis and management of obscure gastrointestinal bleeding

The diagnosis and management of obscure gastrointestinal bleeding (OGIB) have changed dramatically since the introduction of video capsule endoscopy (VCE) followed by deep enteroscopy and other imaging technologies in the last decade. Significant advances have been made, yet there remains room for improvement in our diagnostic yield and treatment capabilities for recurrent OGIB. In this review, we will summarize the latest technologies for the diagnosis of OGIB, limitations of VCE, technological enhancement in VCE, and different management options for OGIB.

Small-bowel capsule endoscopy: a ten-point contemporary review

The introduction of capsule endoscopy (CE) in clinical practice increased the interest for the study of the small-bowel. Consequently, in about 10 years, an impressive quantity of literature on indications, diagnostic yield (DY), safety profile and technical evolution of CE has been published as well as several reviews. At present time, there are 5 small-bowel capsule enteroscopy (SBCE) models in the worldwide market. Head-to-head trials have showed in the great majority of studies comparable results in terms of DY, image quality and completion rate. CE meta-analyses formed the basis of national/international guidelines; these guidelines place CE in a prime position for the diagnostic work-up of patients with obscure gastrointestinal bleeding, known and/or suspected Crohn's disease and possible small-bowel neoplasia. A 2-L polyethylene glycol-based purge, administered the day before the procedure, is the most widely practiced preparation regimen. Whether this regimen can be further improved (i.e., by further decreasing its volume, changing the timing of administration, coupling it with prokinetics and/or other factors) or if it can really affect the DY, is still under discussion. Faecal calprotectin has been used in SBCE studies in two settings: in patients taking non-steroidal anti-inflammatory drugs, to evaluate the type and extent of mucosal damage and, more importantly from a clinical point of view, in patients with known or suspected Crohn's disease for assessment of inflammation activity. Although there is still a lot of debate around the exact reasons of SBCE poor performance in various small-bowel segments, it is worth to remember that the capsule progress is non-steerable, hence more rapid in the proximal than in lower segments of the small-bowel. Capsule aspiration, a relatively unexpected complication, has been reported with increasing frequency. This is probably related with the increase in the mean age of patients undergoing CE. CE video review is a time-consuming procedure. Therefore, several attempts have been made to develop technical software features, in order to make CE video analysis easier and shorter (without jeopardizing its accuracy). Suspected Blood Indicator, QuickView and Fujinon Intelligent Chromo Endoscopy are some of the software tools that have been checked in various clinical studies to date.

Evaluation of four time-saving methods of reading capsule endoscopy videos

OBJECTIVE

Review of wireless capsule endoscopy recordings is time consuming. The aim of this study was to evaluate four time-saving methods offered with Rapid Software.

METHODS

A total of 100 wireless capsule endoscopy videos with abnormal findings were evaluated using five different ways of viewing: (a) manual mode at a speed of 10 frames per second (fps), (b) manual mode at a speed of 20 fps, (c) manual mode with a simultaneous display of two images at a speed of 20 fps, (d) automatic mode at a speed of 10 fps, and (e) quickview mode at a speed of 3 fps. Then, we calculated the concordance of abnormal findings between each one of the four time-saving methods using method A.

RESULTS

The mean reading time with time-saving methods was significantly shorter than with method A (method A: 59.8 min, method B: 30 min, method C: 30.2 min, method D: 32.2 min, method E: 16.3 min). The agreement in finding abnormal lesions between method A and the four evaluated methods was excellent and almost perfect (κ>0.8), except for quickview in recognizing polyps. Diagnostic miss rate was 1% for method D, 2% for B and C, and 12% for E. No tumors and cases of celiac or Crohn's disease were lost by all four methods.

CONCLUSION

We conclude that manual mode/20 fps, the simultaneous projection of two images/20 fps, and automatic mode/10 fps have minimal diagnostic miss rates and can safely replace slower modes in clinical practice. The quickview mode is a safe diagnostic tool only when larger or diffuse lesions are suspected, such as Crohn's or celiac disease.