Colonoscopy Videos Research Articles

Identification of cecum time-location in a colonoscopy video by deep learning analysis of colonoscope movement.

BackgroundCecal intubation time is an important component for quality colonoscopy. Cecum is the turning point that determines the insertion and withdrawal phase of the colonoscope. For this reason, obtaining information related with location of the cecum in the endoscopic procedure is very useful. Also, it is necessary to detect the direction of colonoscope’s movement and time-location of the cecum.MethodsIn order to analysis the direction of scope’s movement, the Horn–Schunck algorithm was used to compute the pixel’s motion change between consecutive frames. Horn–Schunk-algorithm applied images were trained and tested through convolutional neural network deep learning methods, and classified to the insertion, withdrawal and stop movements. Based on the scope’s movement, the graph was drawn with a value of +1 for insertion, −1 for withdrawal, and 0 for stop. We regarded the turning point as a cecum candidate point when the total graph area sum in a certain section recorded the lowest.ResultsA total of 328,927 frame images were obtained from 112 patients. The overall accuracy, drawn from 5-fold cross-validation, was 95.6%. When the value of “t” was 30 s, accuracy of cecum discovery was 96.7%. In order to increase visibility, the movement of the scope was added to summary report of colonoscopy video. Insertion, withdrawal, and stop movements were mapped to each color and expressed with various scale. As the scale increased, the distinction between the insertion phase and the withdrawal phase became clearer.ConclusionInformation obtained in this study can be utilized as metadata for proficiency assessment. Since insertion and withdrawal are technically different movements, data of scope’s movement and phase can be quantified and utilized to express pattern unique to the colonoscopist and to assess proficiency. Also, we hope that the findings of this study can contribute to the informatics field of medical records so that medical charts can be transmitted graphically and effectively in the field of colonoscopy.

A locally based feature descriptor for abnormalities detection

Wireless capsule endoscopy (WCE) is a novel imaging technique that can view the entire small bowel in human body. Therefore, it has been gradually adopted compared with traditional endoscopies for gastrointestinal diseases. However, the task of reviewing the vast amount of images produced by a WCE test is exhaustive for the physicians. This paper presents a new feature extraction scheme for pathological inflammation and ulcer regions discrimination in WCE images. In addition, the novel approach is adopted for polyp recognition in colonoscopy videos. A novel idea based on extracting certain local features from the image is proposed. Then, the occurrence histogram of these features is used as descriptor of the image. The new feature descriptor scheme is grayscale rotation invariant and computationally simple as the operator can be realized with a few operations in a small neighborhood. The proposed operator does not discard the contrast information. Besides, we propose to test the quality of the model using logarithmic loss metric and show how calibration can be useful in reducing the aforementioned measure. Extensive classification experiments have been applied on different datasets, which prove that the occurrence histogram of the extracted features is powerful. The proposed method achieved 99.1%, 99.7% and 99.2% in terms of the precision in the first, second and third datasets, respectively, and surpassed some known local descriptors on a texture dataset.

Localization of Small Neoplastic Lesions in Colonoscopy by Estimating Edge, Texture and Motion Saliency.

Early screening in Colorectal Cancer consists in finding and removing small precancerous masses or neoplastic lesions developed from the mucosa, usually lesions smaller than 10 mm. Localization of small neoplastic lesions is a very challenging task since colon exploration is highly dependent on the expert training and colon preparation. Several strategies have attempted to locate neoplasias, but usually they are huge lesions that a trained gastroenterologist could hardly miss. This work presents a saliency-based strategy to localize polypoid and non-polypoid neoplastic lesions smaller than 10 mm in colonoscopy videos by combining spatio-temporal descriptors. For doing so, a per-frame-multi-scale representation is computed and edge, texture and motion features are extracted. Each of these features is used to construct a primary saliency map which are then combined to obtain a coarse saliency map. Finally, the neoplasia is localized as the bounding box of a circular region, approximated by the Hough transform, with the largest salience. The proposed approach was evaluated in 8 short colonoscopy videos obtaining an average of Annotated Area Covered of 0.75 and a precision of 0.82.

Automated Detection of Non-Informative Frames for Colonoscopy Through a Combination of Deep Learning and Feature Extraction.

Colonoscopy is a standard medical examination used to inspect the mucosal surface and detect abnormalities of the colon. Objective assessment and scoring of disease features in the colon are important in conditions such as colorectal cancer and inflammatory bowel disease. However, subjectivity in human disease assessment and measurement is hampered by interobserver variation and several biases. A computer-aided system for colonoscopy video analysis could facilitate diagnosis and disease severity measurement, which would aid in treatment selection and clinical outcome prediction. However, a large number of images captured during colonoscopy are non-informative, making detecting and removing those frames an important first step in performing automated analysis. In this paper, we present a combination of deep learning and conventional feature extraction to distinguish non-informative from informative images in patients with ulcerative colitis. Our result shows that the combination of bottleneck features in the RGB color space and hand-crafted features in the HSV color space can boost the classification performance. Our proposed method was validated using 5-fold cross-validation and achieved an average AUC of 0.939 and an average F1 score of 0.775.

Performance of a Deep Learning Model vs Human Reviewers in Grading Endoscopic Disease Severity of Patients With Ulcerative Colitis

Assessing endoscopic disease severity in ulcerative colitis (UC) is a key element in determining therapeutic response, but its use in clinical practice is limited by the requirement for experienced human reviewers. To determine whether deep learning models can grade the endoscopic severity of UC as well as experienced human reviewers. In this diagnostic study, retrospective grading of endoscopic images using the 4-level Mayo subscore was performed by 2 independent reviewers with score discrepancies adjudicated by a third reviewer. Using 16 514 images from 3082 patients with UC who underwent colonoscopy at a single tertiary care referral center in the United States between January 1, 2007, and December 31, 2017, a 159-layer convolutional neural network (CNN) was constructed as a deep learning model to train and categorize images into 2 clinically relevant groups: remission (Mayo subscore 0 or 1) and moderate to severe disease (Mayo subscore, 2 or 3). Ninety percent of the cohort was used to build the model and 10% was used to test it; the process was repeated 10 times. A set of 30 full-motion colonoscopy videos, unseen by the model, was then used for external validation to mimic real-world application. Model performance was assessed using area under the receiver operating curve (AUROC), sensitivity and specificity, positive predictive value (PPV), and negative predictive value (NPV). Kappa statistics (κ) were used to measure agreement of the CNN relative to adjudicated human reference cores. The authors included 16 514 images from 3082 unique patients (median [IQR] age, 41.3 [26.1-61.8] years, 1678 [54.4%] female), with 3980 images (24.1%) classified as moderate-to-severe disease by the adjudicated reference score. The CNN was excellent for distinguishing endoscopic remission from moderate-to-severe disease with an AUROC of 0.966 (95% CI, 0.967-0.972); a PPV of 0.87 (95% CI, 0.85-0.88) with a sensitivity of 83.0% (95% CI, 80.8%-85.4%) and specificty of 96.0% (95% CI, 95.1%-97.1%); and NPV of 0.94 (95% CI, 0.93-0.95). Weighted κ agreement between the CNN and the adjudicated reference score was also good for identifying exact Mayo subscores (κ = 0.84; 95% CI, 0.83-0.86) and was similar to the agreement between experienced reviewers (κ = 0.86; 95% CI, 0.85-0.87). Applying the CNN to entire colonoscopy videos had similar accuracy for identifying moderate to severe disease (AUROC, 0.97; 95% CI, 0.963-0.969). This study found that deep learning model performance was similar to experienced human reviewers in grading endoscopic severity of UC. Given its scalability, this approach could improve the use of colonoscopy for UC in both research and routine practice.

Construction of artificial intelligence assisted diagnosis model for colonoscopy

Objective To establish an artificial intelligence deep learning model using clinical colonoscopy images and video to assist the diagnosis by colonoscopy. Methods More than 600 000 colonoscopy images were collected in endoscopic center of the Second Affiliated Hospital of Zhejiang University School of Medicine from 2014 to 2018, and endoscopic experts recorded a large number of high-quality operation video of colonoscopy as analysis data. After repeated discussion by six experts, the classified intestinal sites and pathological features were determined, and fuzzy and confusable images were deleted. The final selection result was approximately 1 out of 4. And then the features of images were marked using an independently developed software. The deep learning algorithm was developed using TensorFlow platform of Google. Results After repeated comparison and analysis of the results of machine training and judgment results combined with pathology from endoscopic experts, the sensitivity of the model for some diseases (such as colon polyps) was 99% under laboratory conditions. In the clinical colonoscopy test, the sensitivity, specificity, and overall accuracy of this model for diagnosis of colon polyps were 98.30% (4 187/4 259), 88.10% (17 620/20 000), and 92.92% [2×98.30%×88.10%/(98.30%+ 88.10%)], respectively. The sensitivity and specificity for ulcerative colitis were 78.32% (2 671/3 410), and 67.06% (13 412/20 000), respectively. The diagnosis time spent on a single image was 0.5±0.03 s, and it was the real time for application, including system recognition, text prompt in video image, background record and storage. Conclusion The artificial intelligence assisted diagnosis model developed by our team can identify colonic polyps, colorectal cancer, colorectal eminence, colonic diverticulum, ulcerative colitis, etc. The auxiliary diagnosis model of colon disease can guide beginners to carry out colonoscopy, and can improve lesion detection rate, reduce misdiagnosis rate, and improve the overall operating efficiency of endoscopic center, which is conducive to the quality control of colonoscopy. Key words: Colonoscopy; Artificial intelligence; Quality control; Assisted diagnosis

Improving Automatic Polyp Detection Using CNN by Exploiting Temporal Dependency in Colonoscopy Video.

Automatic polyp detection has been shown to be difficult due to various polyp-like structures in the colon and high interclass variations in polyp size, color, shape, and texture. An efficient method should not only have a high correct detection rate (high sensitivity) but also a low false detection rate (high precision and specificity). The state-of-the-art detection methods include convolutional neural networks (CNN). However, CNNs have shown to be vulnerable to small perturbations and noise; they sometimes miss the same polyp appearing in neighboring frames and produce a high number of false positives. We aim to tackle this problem and improve the overall performance of the CNN-based object detectors for polyp detection in colonoscopy videos. Our method consists of two stages: a region of interest (RoI) proposal by CNN-based object detector networks and a false positive (FP) reduction unit. The FP reduction unit exploits the temporal dependencies among image frames in video by integrating the bidirectional temporal information obtained by RoIs in a set of consecutive frames. This information is used to make the final decision. The experimental results show that the bidirectional temporal information has been helpful in estimating polyp positions and accurately predict the FPs. This provides an overall performance improvement in terms of sensitivity, precision, and specificity compared to conventional false positive learning method, and thus achieves the state-of-the-art results on the CVC-ClinicVideoDB video data set.

P076 USING DEEP LEARNING FOR AUTOMATED GRADING OF ENDOSCOPIC DISEASE SEVERITY IN ULCERATIVE COLITIS

Objectives: Endoscopic severity of ulcerative colitis predicts clinical outcomes, making it critical for assessing efficacy of new therapies. However, its use in routine practice is limited by availability of experienced and trained human reviewers. We investigated the feasibility of using deep learning algorithms to independently and accurately grade severity of ulcerative colitis from still images and full motion video colonoscopies. Setting: Patients with ulcerative colitis undergoing colonoscopy at a single tertiary care referral center in the United States between 2007-2017. Design: Retrospective grading of ulcerative colitis severity was performed on colonoscopic images using the standardized Mayo score by at least 2 human reviewers as the gold standard for diagnosis. A 159-layer convolutional neural network was employed as a deep learning algorithm for feature extraction and used to train and categorize images into two clinically-relevant groups: remission (defined as Mayo 0-1) and moderate-to-severe disease (defined as Mayo 2-3). Seventy-percent of the cohort was used to train the algorithm and 30% of the cohort was used to test it. An independent set of graded full motion video colonoscopies (N=30) unseen by the algorithm was then used for external validation. Main Outcome Measure: Performance of the deep learning automated grading system was assessed using Area Under the Receive Operating Curve (AuROC), positive (PPV) and negative predictive values (NPV),and kappa statistics for agreement relative to the gold standard of human reviewers. 14,830 endoscopic still images over 3,038 unique individuals with ulcerative colitis underwent grading. The accuracy of the convolutional neural network approach for distinguishing endoscopic remission from moderate-severe disease was excellent at 93.7% (95% CI 96.4 to 97.9) with an AuROC of 0.979, yielding a PPV of 88.41% and NPV of 93.75%. Misclassification of exact Mayo endoscopic severity score was more common for intermediate endoscopic severities (Mayo grade 1 and 2), with an overall weighted Kappa agreement of k=0.855 (95% CI 0.847 to 0.863) between deep learning prediction and human reference score. Applying score prediction methods to unseen, full-motion colonoscopy videos demonstrated similar accuracy at identifying endoscopic remission (AuROC 0.981) and agreement of scores with human reference scores (k=0.821, 95% CI 0.803 to 0.839, Table 1). A deep learning algorithm accurately grades severity of ulcerative colitis as compared with experienced human reviewers. Automated analysis of colonoscopy videos could substantially improve access to standardized assessments of ulcerative colitis in routine practice and research.

Polyp detection with computer-aided diagnosis in white light colonoscopy: comparison of three different methods.

Background and study aims Detection of polyps during colonoscopy is essential for screening colorectal cancer and computer-aided-diagnosis (CAD) could be helpful for this objective. The goal of this study was to assess the efficacy of CAD in detection of polyps in video colonoscopy by using three methods we have proposed and applied for diagnosis of polyps in wireless capsule colonoscopy. Patients and methods Forty-two patients were included in the study, each one bearing one polyp. A dataset was generated with a total of 1680 polyp instances and 1360 frames of normal mucosa. We used three methods, that are all binary classifiers, labelling a frame as either containing a polyp or not. Two of the methods (Methods 1 and 2) are threshold-based and address the problem of polyp detection (i. e. separation between normal mucosa frames and polyp frames) and the problem of polyp localization (i. e. the ability to locate the polyp in a frame). The third method (Method 3) belongs to the class of machine learning methods and only addresses the polyp detection problem. The mathematical techniques underlying these three methods rely on appropriate fusion of information about the shape, color and texture content of the objects presented in the medical images. Results Regarding polyp localization, the best method is Method 1 with a sensitivity of 71.8 %. Comparing the performance of the three methods in the detection of polyps, independently of the precision in the location of the lesions, Method 3 stands out, achieving a sensitivity of 99.7 %, an accuracy of 91.1 %, and a specificity of 84.9 %. Conclusion CAD, using the three studied methods, showed good accuracy in the detection of polyps with white light colonoscopy.

Colonic Polyp Detection in Endoscopic Videos With Single Shot Detection Based Deep Convolutional Neural Network.

A major rise in the prevalence and influence of colorectal cancer (CRC) leads to substantially increasing healthcare costs and even death. It is widely accepted that early detection and removal of colonic polyps can prevent CRC. Detection of colonic polyps in colonoscopy videos is problematic because of complex environment of colon and various shapes of polyps. Currently, researchers indicate feasibility of Convolutional Neural Network (CNN)-based detection of polyps but better feature extractors are needed to improve detection performance. In this paper, we investigated the potential of the single shot detector (SSD) framework for detecting polyps in colonoscopy videos. SSD is a one-stage method, which uses a feed-forward CNN to produce a collection of fixed-size bounding boxes for each object from different feature maps. Three different feature extractors, including ResNet50, VGG16, and InceptionV3 were assessed. Multi-scale feature maps integrated into SSD were designed for ResNet50 and InceptionV3, respectively. We validated this method on the 2015 MICCAI polyp detection challenge datasets, compared it with teams attended the challenge, YOLOV3 and two-stage method, Faster-RCNN. Our results demonstrated that the proposed method surpassed all the teams in MICCAI challenge and YOLOV3 and was comparable with two-stage method. Especially in detection speed aspect, our proposed method outperformed all the methods, met real-time application requirement. Meanwhile, we also indicated that among all the feature extractors, InceptionV3 obtained the best result of precision and recall. In conclusion, SSD- based method achieved excellent detection performance in polyp detection and can potentially improve diagnostic accuracy and efficiency.

The frequency and clinical and endoscopic features of mixed NSAIDs-induced gastrointestinal injuries

Non-steroidal anti-inflammatory drugs (NSAIDs) can cause undesirable reactions in all parts of the gastrointestinal tract (GIT). However, the frequency of mixed injuries of various GIT parts due to the use of these drugs has not been investigated.Objective:to estimate the frequency of mixed NSAID-induced injuries of the upper GIT, small and large intestine.Patients and methods.A total of 112 patients (62.5% were women) (mean age, 56.2±14.6 years) with rheumatic diseases who had regularly taken NSAIDs were examined. All the patients underwent esophagogastroduodenoscopy and video colonoscopy. Video capsule endoscopy was performed in 35 patients with signs of NSAID-induced gastropathy.Results and discussion.The signs of NSAID-induced gastropathy (gastric and duodenal erosions and/or ulcers) were found in 43.8% of patients; those of NSAID-induced enteropathy (small bowel hemorrhages, erosions, and ulcers) were present in 68.6%; and those of NSAIDinduced colopathy (colonic hemorrhages, erosions, and ulcers) were in 14.3%. The concurrence of NSAID-induced gastro- and colonopathy was present in 28.6% of the patients (odds ratio 12.2; 95% confidence interval, 2.619–56.84); that of NSAID-induced gastro-, entero-, and colopathy was in 10 (20.4% of all the patients with NSAID-induced gastropathy). There was a significant association of the risk of mixed pathology in all GIT parts with the diagnosis of spondylarthritis, the presence of abdominal pain, the signs of dysbiosis and bacterial overgrowth syndrome, as well as with the carriage of CYP2C19 gene polymorphism (the CYP2C19*17*1/*17 allele).Conclusion.Mixed injury of various GIT parts due to the use of NSAIDs is a frequent and serious pathology that requires comprehensive diagnostic tests and combined use of preventive therapies with different mechanisms of action.

A unique case of ileal Dieulafoy's lesion diagnosed and treatedbycap-assisted water immersion colonoscopy.

Ileal Dieulafoy’s lesions are rare causes of GI bleeding and are identified primarily through capsule endoscopy, single-balloon enteroscopy, or double-balloon enteroscopy. We describe a unique case of ileal Dieulafoy’s lesion identified and treated through cap-assisted water-immersion colonoscopy (Video 1, available online at www.VideoGIE.org). An 83-year-old woman with a history of transient ischemic attack on an aspirin regimen presented with a 5-day history of melena and anemia. She had a history of GI bleeding 1 year previously from presumed peptic ulcer and right-sided colonic arteriovenous malformations. Repeated EGD showed retained clips on previous ulcer sites. Colonoscopy showed a 1-cm nonbleeding ascending arteriovenous malformation with more proximal intermittent extrusion of fresh blood from the ileocecal valve. Advancement of a pediatric colonoscope (Evis Exera III PCF-PH190L/l; Olympus Medical Systems, Tokyo, Japan) with distal attachment cap (Reveal; US Endoscopy, Mentor, Ohio, USA) into the ileum revealed a trail of small scattered clots. Further advancement of the colonoscope 30 cm into the ileum (measured by withdrawal from the lesion to the ileocecal valve) demonstrated a 4-mm nonadherent clot. Irrigation of the clot uncovered a 1-mm actively pulsating bleeding vessel. Hemostasis was achieved with 3 hemoclips, and a tattoo was placed (Fig. 1). After discharge, no recurrent bleeding was observed. We highlight the use of cap-assisted water-immersion colonoscopy as a means to advance the colonoscope deeper into the small bowel compared with traditional gas distention colonoscopy. The cap extends the length of the colonoscope 3 to 4 mm, which allows the proximal adjacent intestinal folds to be pushed down ahead of the scope. In some areas where inevitable looping results and advancing the scope even 1 cm is difficult, the cap may assist in pushing down the proximal adjacent fold to help anchor the colonoscope forward and beyond where an otherwise introduced loop has difficulty advancing. Therapeutically, the cap stabilizes lesions into the field of view, preventing inadvertent deflection of the tip. The water-immersion method minimizes excessive looping and lengthening of the colon. The shortened colon length potentially allows more proximal advancement of the colonoscope. Deeper intubation of the ileum allows lesions like this one to be reached without the need for capsule or balloon enteroscopy. Cap-assisted water-immersion colonoscopy may allow more efficient and cost-effective means of identifying and treating ileal lesions thought to be otherwise unreachable by gas distention colonoscopy. All authors disclosed no financial relationships relevant to this publication. eyJraWQiOiI4ZjUxYWNhY2IzYjhiNjNlNzFlYmIzYWFmYTU5NmZmYyIsImFsZyI6IlJTMjU2In0.eyJzdWIiOiI0YmI5NTA1YmQ2ZDcxMzczNmFkNmJiNmU3ZTk5ZGNjZCIsImtpZCI6IjhmNTFhY2FjYjNiOGI2M2U3MWViYjNhYWZhNTk2ZmZjIiwiZXhwIjoxNjc4NzQ4OTgwfQ.kUT-cUQDKFui2G-CNZc1qE2Xnpx_8VggXnt8HInPZ1TvsHQOAxBgrnGlQOrXb21rfpZsaBelzmgtc9QugNCH8jIEl0bwzSpqBj-hOyGqwahx7zMmrxqBkbF_XWU0Y-I9h-iyt_OxK_0vxGuPhe7-r4geZhxa-MZR6JSvu4t63wYcweKtN5YC6RadudcV3pjhJjo6B8XLf_lvRkrLMikeprWX06TON7qKuIP7MIFc0oboQmEKmh7you_wYi4YhKJFkAokaXdGUCjWlgAMx_a5xaBb-smXynCQVKSFYbXdhTNq41T5wbpshJJi0VyosoYFQFJGBCqpOVVbzao3upp8eg Download .mp4 (219.31 MB) Help with .mp4 files Video 1. Ileal intubation and endoscopic therapy with cap-assisted water-immersion colonoscopy.

Colonoscopy Videos on YouTube: Are They a Good Source of Patient Education? Presidential Poster Award

Colonoscopy Videos on YouTube: Are They a Good Source of Patient Education? Presidential Poster Award

Development and validation of a deep-learning algorithm for the detection of polyps during colonoscopy.

The detection and removal of precancerous polyps via colonoscopy is the gold standard for the prevention of colon cancer. However, the detection rate of adenomatous polyps can vary significantly among endoscopists. Here, we show that a machine-learning algorithm can detect polyps in clinical colonoscopies, in real time and with high sensitivity and specificity. We developed the deep-learning algorithm by using data from 1,290 patients, and validated it on newly collected 27,113 colonoscopy images from 1,138 patients with at least one detected polyp (per-image-sensitivity, 94.38%; per-image-specificity, 95.92%; area under the receiver operating characteristic curve, 0.984), on a public database of 612 polyp-containing images (per-image-sensitivity, 88.24%), on 138 colonoscopy videos with histologically confirmed polyps (per-image-sensitivity of 91.64%; per-polyp-sensitivity, 100%), and on 54 unaltered full-range colonoscopy videos without polyps (per-image-specificity, 95.40%). By using a multi-threaded processing system, the algorithm can process at least 25 frames per second with a latency of 76.80 ± 5.60 ms in real-time video analysis. The software may aid endoscopists while performing colonoscopies, and help assess differences in polyp and adenoma detection performance among endoscopists.

Blue-light imaging compared with high-definition white light for real-time histology prediction of colorectal polyps less than 1 centimeter: a prospective randomized study

Blue-light imaging (BLI) is a new chromoendoscopy technique, potentially useful for differentiating neoplastic from nonneoplastic lesions. The present study was aimed at comparing BLI with high-definition white light (HDWL) in the real-time histology prediction of colon polyps<10mm. Consecutive outpatients undergoing colonoscopy with the ELUXEO 7000 endoscopy platform and 760 series video colonoscopes (Fujifilm Co, Tokyo, Japan) who had at least 1 polyp<10mm were randomized to BLI or HDWL for polyp characterization. The accuracy of high-confidence real-time histology prediction (adenoma vs not adenoma) by either BLI or HDWL for polyps<10mm (primary end-point) and diminutive (≤5mm) polyps was calculated, along with sensitivity, specificity, and positive and negative predictive values, with histopathology as the reference standard. A total of 483 polyps were detected in 245 randomized patients (125 and 120 in the BLI and HDWL arms, respectively). A total of 358 were diminutive, and 283 were adenomas. Overall, 222 (85.7%) and 193 (86.1%) polyps were characterized with high confidence by BLI and HDWL, respectively (P= .887), with an overall accuracy of 92% and 84%, respectively (P= .011). The accuracy was significantly higher by BLIthan HDWL, also for diminutive polyps (92% vs 83%; P= .008). When BLI was used, the negative predictive value for diminutive rectosigmoid polyps was 88%, and the post-polypectomy surveillance interval wascorrectly attributed in 85.7% and 93.7% of patients, respectively, according to U.S. and European guidelines. BLI was superior to HDWL for the real-time prediction of histology in polyps<10mm. A BLI-dedicated classification might further improve the endoscopist performance. (Clinical trial registration number: NCT03274115.).

GTCreator: a flexible annotation tool for image-based datasets.

Methodology evaluation for decision support systems for health is a time-consuming task. To assess performance of polyp detection methods in colonoscopy videos, clinicians have to deal with the annotation of thousands of images. Current existing tools could be improved in terms of flexibility and ease of use. We introduce GTCreator, a flexible annotation tool for providing image and text annotations to image-based datasets. It keeps the main basic functionalities of other similar tools while extending other capabilities such as allowing multiple annotators to work simultaneously on the same task or enhanced dataset browsing and easy annotation transfer aiming to speed up annotation processes in large datasets. The comparison with other similar tools shows that GTCreator allows to obtain fast and precise annotation of image datasets, being the only one which offers full annotation editing and browsing capabilites. Our proposed annotation tool has been proven to be efficient for large image dataset annotation, as well as showing potential of use in other stages of method evaluation such as experimental setup or results analysis.

Educational Colonoscopy Video Enhances Bowel Preparation Quality and Comprehension in an Inner City Population.

Quality of bowel preparation and patient knowledge remains a major barrier for completing colorectal cancer screening. Few studies have tested unique ways to impact patient understanding centering on interactive computer programs, pictures, and brochures. Two studies explored instructional videos but focused on patient compliance and anxiety as endpoints. Furthermore, excessive video length and content may limit their impact on a broad patient population. No study so far has studied a video's impact on preparation quality and patient understanding of the colonoscopy procedure. We conducted a single blinded prospective study of inner city patients presenting for a first time screening colonoscopy. During their initial visit patients were randomized to watch an instructional colonoscopy video or a video discussing gastroesophageal reflux disease (GERD). All patients watched a 6 minutes long video with the same spokesperson, completed a demographic questionnaire (Supplemental Digital Content 1, http://links.lww.com/JCG/A352) and were enrolled only if screened within 30 days of their visit. On the day of the colonoscopy, patients completed a 14 question quiz of their knowledge. Blinded endoscopist graded patient preparations based on the Ottawa scale. All authors had access to the study data and reviewed and approved the final manuscript. Among the 104 subjects enrolled in the study, 56 were in the colonoscopy video group, 48 were in GERD video group, and 12 were excluded. Overall, 48% were male and 52% female; 90% of patients had less than a high school education, 76% were African American, and 67% used a 4 L split-dose preparation. There were no differences between either video group with regard to any of the above categories. Comparisons between the 2 groups revealed that the colonoscopy video group had significantly better Ottawa bowel preparation score (4.77 vs. 6.85; P=0.01) than the GERD video group. The colonoscopy video group also had less-inadequate repeat bowel preparations versus the GERD video group (9% vs. 23%; P<0.01). The overall score on the knowledge questionnaire (Supplemental Digital Content 1, http://links.lww.com/JCG/A352) was significantly higher in the colonoscopy video group as compared with the GERD video group (12.77 vs. 11.08; P<0.001. In all patients the overall quiz score positively correlated with preparation quality (odds ratio, 2.31; confidence interval, 1.35-3.94; P<0.001). Our unique population represented an overwhelmingly under-educated (85% had a high school education or less) and minority group (76% African American). They are one of the most at risk of having multiple barriers such as comprehension and reading difficulties resulting in poor preparation examinations and no shows to procedures. Our instructional video proved to be high yield in this population. The patients assigned to watch the colonoscopy video showed a significant increase in "excellent" grade adequate bowel preparation quality by >23% and a significant decrease in "inadequate" bowel preparations by almost 50%. Our study proves that an educational video can improve both comprehension with regard to all aspects of colonoscopy. ClinicalTrials.gov number, NCT02906969.

Classification of Informative Frames in Colonoscopy Videos Using Convolutional Neural Networks with Binarized Weights.

Colorectal cancer is one of the common cancers in the United States. Polyps are one of the major causes of colonic cancer, and early detection of polyps will increase the chance of cancer treatments. In this paper, we propose a novel classification of informative frames based on a convolutional neural network with binarized weights. The proposed CNN is trained with colonoscopy frames along with the labels of the frames as input data. We also used binarized weights and kernels to reduce the size of CNN and make it suitable for implementation in medical hardware. We evaluate our proposed method using Asu Mayo Test clinic database, which contains colonoscopy videos of different patients. Our proposed method reaches a dice score of 71.20% and accuracy of more than 90% using the mentioned dataset.

Polyp Segmentation in Colonoscopy Images Using Fully Convolutional Network.

Colorectal cancer is one of the highest causes of cancer-related death, especially in men. Polyps are one of the main causes of colorectal cancer, and early diagnosis of polyps by colonoscopy could result in successful treatment. Diagnosis of polyps in colonoscopy videos is a challenging task due to variations in the size and shape of polyps. In this paper, we proposed a polyp segmentation method based on the convolutional neural network. Two strategies enhance the performance of the method. First, we perform a novel image patch selection method in the training phase of the network. Second, in the test phase, we perform effective post-processing on the probability map that is produced by the network. Evaluation of the proposed method using the CVC-ColonDB database shows that our proposed method achieves more accurate results in comparison with previous colonoscopy video-segmentation methods.

Motion artifact detection in colonoscopy images

Abstract Computer-aided detection is an integral part of medical image evaluation process because examination of each image takes a long time and generally experts’ do not have enough time for the elimination of images with motion artifact (blurred images). Computer-aided detection is required for both increasing accuracy rate and saving experts’ time. Large intestine does not have straight structure thus camera of the colonoscopy should be moved continuously to examine inside of the large intestine and this movement causes motion artifact on colonoscopy images. In this study, images were selected from open-source colonoscopy videos and obtained at Kayseri Training and Research Hospital. Totally 100 images were analyzed half of which were clear. Firstly, a modified version of histogram equalization was applied in the pre-processing step to all images in our dataset, and then, used Laplacian, wavelet transform (WT), and discrete cosine transform-based (DCT) approaches to extract features for the discrimination of images with no artifact (clear) and images with motion artifact. The Laplacian-based feature extraction method was used for the first time in the literature on colonoscopy images. The comparison between Laplacian-based features and previously used methods such as WT and DCT has been performed. In the classification phase of our study, support vector machines (SVM), linear discriminant analysis (LDA), and k nearest neighbors (k-NN) were used as the classifiers. The results showed that Laplacian-based features were more successful in the detection of images with motion artifact when compared to popular methods used in the literature. As a result, a combination of features extracted using already existing approaches (WT and DCT) and the Laplacian-based methods reached 85% accuracy levels with SVM classification approach