Establishment and validation of an artificial intelligence-based model for real-time detection and classification of colorectal adenoma.

Colorectal cancer (CRC) prevention requires early detection and removal of adenomas. We aimed to develop a computational model for real-time detection and classification of colorectal adenoma. Computationally constrained background based on real-time detection, we propose an improved adaptive lightweight ensemble model for real-time detection and classification of adenomas and other polyps. Firstly, we devised an adaptive lightweight network modification and effective training strategy to diminish the computational requirements for real-time detection. Secondly, by integrating the adaptive lightweight YOLOv4 with the single shot multibox detector network, we established the adaptive small object detection ensemble (ASODE) model, which enhances the precision of detecting target polyps without significantly increasing the model's memory footprint. We conducted simulated training using clinical colonoscopy images and videos to validate the method's performance, extracting features from 1148 polyps and employing a confidence threshold of 0.5 to filter out low-confidence sample predictions. Finally, compared to state-of-the-art models, our ASODE model demonstrated superior performance. In the test set, the sensitivity of images and videos reached 87.96% and 92.31%, respectively. Additionally, the ASODE model achieved an accuracy of 92.70% for adenoma detection with a false positive rate of 8.18%. Training results indicate the effectiveness of our method in classifying small polyps. Our model exhibits remarkable performance in real-time detection of colorectal adenomas, serving as a reliable tool for assisting endoscopists.

Evaluation of an Artificial Intelligence-Augmented Digital System for Histologic Classification of Colorectal Polyps.

Importance: Colorectal polyps are common, and their histopathologic classification is used in the planning of follow-up surveillance. Substantial variation has been observed in pathologists' classification of colorectal polyps, and improved assessment by pathologists may be associated with reduced subsequent underuse and overuse of colonoscopy. Objective: To compare standard microscopic assessment with an artificial intelligence (AI)-augmented digital system that annotates regions of interest within digitized polyp tissue and predicts polyp type using a deep learning model to assist pathologists in colorectal polyp classification. Design, Setting, and Participants: In this diagnostic study conducted at a tertiary academic medical center and a community hospital in New Hampshire, 100 slides with colorectal polyp samples were read by 15 pathologists using a microscope and an AI-augmented digital system, with a washout period of at least 12 weeks between use of each modality. The study was conducted from February 10 to July 10, 2020. Main Outcomes and Measures: Accuracy and time of evaluation were used to compare pathologists' performance when a microscope was used with their performance when the AI-augmented digital system was used. Outcomes were compared using paired t tests and mixed-effects models. Results: In assessments of 100 slides with colorectal polyp specimens, use of the AI-augmented digital system significantly improved pathologists' classification accuracy compared with microscopic assessment from 73.9% (95% CI, 71.7%-76.2%) to 80.8% (95% CI, 78.8%-82.8%) (P < .001). The overall difference in the evaluation time per slide between the digital system (mean, 21.7 seconds; 95% CI, 20.8-22.7 seconds) and microscopic examination (mean, 13.0 seconds; 95% CI, 12.4-13.5 seconds) was -8.8 seconds (95% CI, -9.8 to -7.7 seconds), but this difference decreased as pathologists became more familiar and experienced with the digital system; the difference between the time of evaluation on the last set of 20 slides for all pathologists when using the microscope and the digital system was 4.8 seconds (95% CI, 3.0-6.5 seconds). Conclusions and Relevance: In this diagnostic study, an AI-augmented digital system significantly improved the accuracy of pathologic interpretation of colorectal polyps compared with microscopic assessment. If applied broadly to clinical practice, this tool may be associated with decreases in subsequent overuse and underuse of colonoscopy and thus with improved patient outcomes and reduced health care costs.

Serrated polyps and polyposis of the colon: a brief review for surgeon endoscopists.

Serrated polyps (SPs) were once considered benign, clinically unimportant lesions. However, it is now recognized that through the serrated neoplasia pathway (SNP), SPs play a role in the development of 15%-30% of cases of colorectal cancers (CRC). Furthermore, a high proportion of postcolonoscopy CRCs are believed to arise from SNP. Serrated polyps are classified into hyperplastic polyps, sessile serrated lesions, sessile serrated lesions with dysplasia, traditionally serrated adenomas, and unclassified serrated adenoma, each with a distinct morphological and molecular profile. Despite improved understanding, SPs remain a clinical challenge owing to evolving terminology, frequent pathologic misclassification, endoscopic underdetection, and high rates of incomplete removal. Surgeon endoscopists and surgeons who perform colorectal procedures will undoubtedly come across patients with SPs, and this paper summarizes some of the clinical challenges they will encounter. We also discuss the diagnosis and management of patients with serrated polyposis syndrome (SPS).

Colonoscopy polyp detection and classification: Dataset creation and comparative evaluations.

Colorectal cancer (CRC) is one of the most common types of cancer with a high mortality rate. Colonoscopy is the preferred procedure for CRC screening and has proven to be effective in reducing CRC mortality. Thus, a reliable computer-aided polyp detection and classification system can significantly increase the effectiveness of colonoscopy. In this paper, we create an endoscopic dataset collected from various sources and annotate the ground truth of polyp location and classification results with the help of experienced gastroenterologists. The dataset can serve as a benchmark platform to train and evaluate the machine learning models for polyp classification. We have also compared the performance of eight state-of-the-art deep learning-based object detection models. The results demonstrate that deep CNN models are promising in CRC screening. This work can serve as a baseline for future research in polyp detection and classification.

Multiclassification of Endoscopic Colonoscopy Images Based on Deep Transfer Learning.

With the continuous improvement of human living standards, dietary habits are constantly changing, which brings various bowel problems. Among them, the morbidity and mortality rates of colorectal cancer have maintained a significant upward trend. In recent years, the application of deep learning in the medical field has become increasingly spread aboard and deep. In a colonoscopy, Artificial Intelligence based on deep learning is mainly used to assist in the detection of colorectal polyps and the classification of colorectal lesions. But when it comes to classification, it can lead to confusion between polyps and other diseases. In order to accurately diagnose various diseases in the intestines and improve the classification accuracy of polyps, this work proposes a multiclassification method for medical colonoscopy images based on deep learning, which mainly classifies the four conditions of polyps, inflammation, tumor, and normal. In view of the relatively small number of data sets, the network firstly trained by transfer learning on ImageNet was used as the pretraining model, and the prior knowledge learned from the source domain learning task was applied to the classification task about intestinal illnesses. Then, we fine-tune the model to make it more suitable for the task of intestinal classification by our data sets. Finally, the model is applied to the multiclassification of medical colonoscopy images. Experimental results show that the method in this work can significantly improve the recognition rate of polyps while ensuring the classification accuracy of other categories, so as to assist the doctor in the diagnosis of surgical resection.

External validation of blue light imaging (BLI) criteria for the optical characterization of colorectal polyps by endoscopy experts.

BACKGROUND AND AIM: Recently, the BLI Adenoma Serrated International Classification (BASIC) system was developed by European experts to differentiate colorectal polyps. Our aim was to validate the BASIC classification system among the US-based endoscopy experts. METHODS: Participants utilized a web-based interactive learning system where the group was asked to characterize polyps using the BASIC criteria: polyp surface (presence of mucus, regular/irregular and [pseudo]depressed), pit appearance (featureless, round/non-round with/without dark spots; homogeneous/heterogeneous distribution with/without focal loss), and vessels (present/absent, lacy, peri-cryptal, irregular). The final testing consisted of reviewing BLI images/videos to determine whether the criteria accurately predicted the histology results. Confidence in adenoma identification (rated "1" to "5") and agreement in polyp (adenoma vs non-adenoma) identification and characterization per BASIC criteria were derived. Strength of interobserver agreement with kappa (k) value was reported for adenoma identification. RESULTS: Ten endoscopy experts from the United States identified conventional adenoma (vs non-adenoma) with 94.4% accuracy, 95.0% sensitivity, 93.8% specificity, 93.8% positive predictive value, and 94.9% negative predictive value using BASIC criteria. Overall strength of interobserver agreement was high: kappa 0.89 (0.82-0.96). Agreement for the individual criteria was as follows: surface mucus (93.8%), regularity (65.6%), type of pit (40.6%), pit visibility (66.9%), pit distribution (57%), vessel visibility (73%), and being lacy (46%) and peri-cryptal (61%). The confidence in diagnosis was rated at high ≥4 in 67% of the cases. CONCLUSIONS: A group of US-based endoscopy experts have validated a simple and easily reproducible BLI classification system to characterize colorectal polyps with >90% accuracy and a high level of interobserver agreement.

Computational learning of features for automated colonic polyp classification.

Shape, texture, and color are critical features for assessing the degree of dysplasia in colonic polyps. A comprehensive analysis of these features is presented in this paper. Shape features are extracted using generic Fourier descriptor. The nonsubsampled contourlet transform is used as texture and color feature descriptor, with different combinations of filters. Analysis of variance (ANOVA) is applied to measure statistical significance of the contribution of different descriptors between two colonic polyps: non-neoplastic and neoplastic. Final descriptors selected after ANOVA are optimized using the fuzzy entropy-based feature ranking algorithm. Finally, classification is performed using Least Square Support Vector Machine and Multi-layer Perceptron with five-fold cross-validation to avoid overfitting. Evaluation of our analytical approach using two datasets suggested that the feature descriptors could efficiently designate a colonic polyp, which subsequently can help the early detection of colorectal carcinoma. Based on the comparison with four deep learning models, we demonstrate that the proposed approach out-performs the existing feature-based methods of colonic polyp identification.

New polyp image classification technique using transfer learning of network-in-network structure in endoscopic images.

While colorectal cancer is known to occur in the gastrointestinal tract. It is the third most common form of cancer of 27 major types of cancer in South Korea and worldwide. Colorectal polyps are known to increase the potential of developing colorectal cancer. Detected polyps need to be resected to reduce the risk of developing cancer. This research improved the performance of polyp classification through the fine-tuning of Network-in-Network (NIN) after applying a pre-trained model of the ImageNet database. Random shuffling is performed 20 times on 1000 colonoscopy images. Each set of data are divided into 800 images of training data and 200 images of test data. An accuracy evaluation is performed on 200 images of test data in 20 experiments. Three compared methods were constructed from AlexNet by transferring the weights trained by three different state-of-the-art databases. A normal AlexNet based method without transfer learning was also compared. The accuracy of the proposed method was higher in statistical significance than the accuracy of four other state-of-the-art methods, and showed an 18.9% improvement over the normal AlexNet based method. The area under the curve was approximately 0.930 ± 0.020, and the recall rate was 0.929 ± 0.029. An automatic algorithm can assist endoscopists in identifying polyps that are adenomatous by considering a high recall rate and accuracy. This system can enable the timely resection of polyps at an early stage.

Hyperplastic polyp or sessile serrated lesion? The contribution of serial sections to reclassification.

BACKGROUND: The histological discrimination of hyperplastic polyps from sessile serrated lesions can be difficult. Sessile serrated lesions and hyperplastic polyps are types of serrated polyps which confer different malignancy risks, and surveillance intervals, and are sometimes difficult to discriminate. Our aim was to reclassify previously diagnosed hyperplastic polyps as sessile serrated lesions or confirmed hyperplastic polyps, using additional serial sections. METHODS: Clinicopathological data for all colorectal hyperplastic polyps diagnosed in 2016 and 2017 was collected. The slides were reviewed and classified as hyperplastic polyps, sessile serrated lesion, or other, using current World Health Organization criteria. Eight additional serial sections were performed for the confirmed hyperplastic polyp group and reviewed. RESULTS: Of an initial 147 hyperplastic polyps from 93 patients, 9 (6.1%) were classified as sessile serrated lesions, 103 as hyperplastic polyps, and 35 as other. Of the 103 confirmed hyperplastic polyps, 7 (6.8%) were proximal, and 8 (7.8%) had a largest fragment size of ≥5 mm and < 10 mm. After 8 additional serial sections, 11 (10.7%) were reclassified as sessile serrated lesions. They were all less than 5 mm and represented 14.3% of proximal polyps and 10.4% of distal polyps. An average of 3.6 serial sections were required for a change in diagnosis. CONCLUSION: Histopathological distinction between hyperplastic polyps and sessile serrated lesions remains a challenge. This study has uncovered a potential role for the use of additional serial sections in the morphological reappraisal of small hyperplastic polyps, especially when proximally located.

A comparative study on polyp classification using convolutional neural networks.

Colorectal cancer is the third most common cancer diagnosed in both men and women in the United States. Most colorectal cancers start as a growth on the inner lining of the colon or rectum, called 'polyp'. Not all polyps are cancerous, but some can develop into cancer. Early detection and recognition of the type of polyps is critical to prevent cancer and change outcomes. However, visual classification of polyps is challenging due to varying illumination conditions of endoscopy, variant texture, appearance, and overlapping morphology between polyps. More importantly, evaluation of polyp patterns by gastroenterologists is subjective leading to a poor agreement among observers. Deep convolutional neural networks have proven very successful in object classification across various object categories. In this work, we compare the performance of the state-of-the-art general object classification models for polyp classification. We trained a total of six CNN models end-to-end using a dataset of 157 video sequences composed of two types of polyps: hyperplastic and adenomatous. Our results demonstrate that the state-of-the-art CNN models can successfully classify polyps with an accuracy comparable or better than reported among gastroenterologists. The results of this study can guide future research in polyp classification.

Feasibility of using narrow band imaging international colorectal endoscopic classification for diagnosing colorectal neoplasia in China: A multicenter pilot observational study.

OBJECTIVE: We aimed to investigate whether Chinese endoscopists without narrow-band imaging (NBI) experiences could achieve high accuracy in the real-time diagnosis of colorectal polyps using NBI International Colorectal Endoscopic (NICE) classification after web-based training. METHODS: Altogether 15 endoscopists from five centers with no NBI experiences followed a short, web-based training program on the NICE classification and took web-based test. Their performances were compared with 15 matched experienced endoscopists with no NBI experience who received no NBI training. These 15 trained endoscopists then made real-time diagnoses of colorectal neoplasia. A logistic regression was used to assess potential predictors of diagnostic performance. RESULTS: Compared with those who received no training, trained endoscopists achieved comparable overall accuracy (85.3% vs 83.1%, P = 0.408) and accuracy at a high-confidence level (87.0% vs 86.0%, P = 0.670), but had a higher confidence rate (86.1% vs 83.7%, P = 0.004) for the diagnosis of neoplasia. Real-time diagnostic accuracy, sensitivity and specificity were 94.3% (95% confidence interval [CI] 91.5%-96.2%), 96.2% (95% CI 93.4%-97.9%) and 85.3% (95% CI 74.8%-92.1%) at high-confidence level. The high-confidence level was the strongest predictor of real-time diagnostic accuracy (odds ratio 12.66, P < 0.001). CONCLUSIONS: Web-based training can improve the confidence level of endoscopists in accurately diagnosing colorectal polyps using the NICE classification. Chinese endoscopists can achieve high accuracy in diagnosing colorectal neoplasia at a high confidence level (ClinicalTrials ID: NCT02033980).

Computer-assisted assessment of colonic polyp histopathology using probe-based confocal laser endomicroscopy.

INTRODUCTION: Probe-based confocal laser endomicroscopy (pCLE) is a promising modality for classifying polyp histology in vivo, but decision making in real-time is hampered by high-magnification targeting and by the learning curve for image interpretation. The aim of this study is to test the feasibility of a system combining the use of a low-magnification, wider field-of-view pCLE probe and a computer-assisted diagnosis (CAD) algorithm that automatically classifies colonic polyps. METHODS: This feasibility study utilized images of polyps from 26 patients who underwent colonoscopy with pCLE. The pCLE images were reviewed offline by two expert and five junior endoscopists blinded to index histopathology. A subset of images was used to train classification software based on the consensus of two GI histopathologists. Images were processed to extract image features as inputs to a linear support vector machine classifier. We compared the CAD algorithm's prediction accuracy against the classification accuracy of the endoscopists. RESULTS: We utilized 96 neoplastic and 93 non-neoplastic confocal images from 27 neoplastic and 20 non-neoplastic polyps. The CAD algorithm had sensitivity of 95%, specificity of 94%, and accuracy of 94%. The expert endoscopists had sensitivities of 98% and 95%, specificities of 98% and 96%, and accuracies of 98% and 96%, while the junior endoscopists had, on average, a sensitivity of 60%, specificity of 85%, and accuracy of 73%. CONCLUSION: The CAD algorithm showed comparable performance to offline review by expert endoscopists and improved performance when compared to junior endoscopists and may be useful for assisting clinical decision making in real time.

Diagnostic yield of the Japan NBI Expert Team (JNET) classification for endoscopic diagnosis of superficial colorectal neoplasms in a large-scale clinical practice database.

Background: Magnifying Narrow Band Imaging (NBI) during colonoscopy is a reliable method for differential and depth diagnoses of colorectal lesions. This study examined the diagnostic yield of magnifying NBI based on the Japan NBI Expert Team (JNET) classification in a clinical setting using a large-scale clinical practice database. Types 1, 2A, 2B and 3 correspond to the histopathological classifications of hyperplastic polyp/sessile-serrated polyp, low-grade intramucosal neoplasia, high-grade intramucosal neoplasia/shallow submucosal invasive cancer, and deep submucosal invasive cancer, respectively. Methods: The prospective records of colonoscopy reports and pathological data of 1558 consecutive superficial colorectal lesions removed by colonoscopy were retrospectively analysed. After excluding 156 lesions, the documented JNET classifications of the remaining 1402 colorectal lesions were analysed. Diagnostic yield was analysed and also compared between expert endoscopists and nonexpert endoscopists. Results: The sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and accuracy were respectively 75%, 96%, 74%, 96% and 93% for type 1; 91%, 70%, 92%, 67% and 87% for type 2A; 42%, 95%, 26%, 98% and 93% for type 2B; and 35%, 100%, 93%, 98% and 98% for type 3. Nonexpert and expert endoscopists alike had specificity, NPV and accuracy >90% for types 1, 2B and 3, and a sensitivity and PPV >90% for type 2A. Type 2B had a low sensitivity of 42% because it included various histological features. Conclusions: The JNET classification proved useful in a clinical setting both for expert and nonexpert endoscopists, as was expected from the original JNET definition, but type 2B requires further investigation using pit pattern diagnosis.

Terminology, Molecular Features, Epidemiology, and Management of Serrated Colorectal Neoplasia.

In addition to the adenoma to carcinoma sequence, colorectal carcinogenesis can occur via the serrated pathway. Studies have focused on clarification of categories and molecular features of serrated polyps, as well as endoscopic detection and risk assessment. Guidelines from the World Health Organization propose assigning serrated polyps to categories of hyperplastic polyps, traditional serrated adenomas, and sessile serrated lesions (SSLs). Traditional serrated adenomas and SSLs are precursors to colorectal cancer. The serrated pathway is characterized by mutations in RAS and RAF, disruptions to the Wnt signaling pathway, and widespread methylation of CpG islands. Epidemiology studies of serrated polyps have been hampered by inconsistencies in terminology and reporting, but the prevalence of serrated class polyps is 20%-40% in average-risk individuals; most serrated polyps detected are hyperplastic. SSLs, the most common premalignant serrated subtype, and are found in up to 15% of average-risk patients by high-detecting endoscopists. Variations in rate of endoscopic detection of serrated polyps indicate the need for careful examination, with adequate bowel preparation and sufficient withdrawal times. Risk factors for SSLs include white race, family history of colorectal cancer, smoking, and alcohol intake. Patients with serrated polyps, particularly SSLs and traditional serrated adenomas, have an increased risk of synchronous and metachronous advanced neoplasia. Surveillance guidelines vary among countries, but SSLs and proximal hyperplastic polyps require special attention in assignment of surveillance interval-especially in light of concerns regarding incomplete detection and resection.

Serrated lesions of the colon A window on a more clear classification.

Serrated polyps evaluation represents a challenge for pathologists for lacking of univocal criteria that leads to different inter -individual interpretation. The aim of our review is to offer an alternative simpler histologic and endoscopic approach to these lesions for a more correct relationship between endoscopists and pathologists.

Clinical usefulness of magnifying colonoscopy for the diagnosis of ulcerative colitis-associated neoplasia.

BACKGROUND AND AIM: The aim of this investigation was to evaluate the efficacy of Japanese magnifying colonoscopic classifications for ulcerative colitis-associated neoplasia (UCAN). METHODS: We reviewed the colonoscopy records from 2011 to 2018 at our institutions and identified cases of endoscopically or surgically resected UCAN observed by magnifying narrow-band imaging (NBI) endoscopy and magnifying chromoendoscopy. Association between magnifying endoscopic classification and histopathological findings was investigated retrospectively. Japan NBI expert team (JNET) classification and pit pattern classification were applied. RESULTS: There were 17 patients who had a diagnosis of UCAN. Tumors of types 2A, 2B and 3 by JNET classification correlated with the histopathological findings of low-grade dysplasia (LGD)/high-grade dysplasia (HGD), HGD, and massively submucosal invasive (mSM) carcinoma, respectively. Tumors of types III/IV, VI low irregularity, and VI high irregularity/VN by pit pattern classification were correlated with the histopathological findings of LGD/HGD, HGD, and mSM carcinoma, respectively. CONCLUSIONS: Japan NBI expert team classification and pit pattern classification may be predictive of the histological diagnosis and invasion depth of UCAN. This needs to be investigated prospectively in a large cohort or in a randomized clinical trial.

Objective validity of the Japan Narrow-Band Imaging Expert Team classification system for the differential diagnosis of colorectal polyps.

BACKGROUND AND AIM: The Japan Narrow-Band Imaging (NBI) Expert Team (JNET) classification is a recently proposed NBI magnifying endoscopy-based classification system for colorectal tumors. Although the usefulness of this system has been reported by JNET experts, its objective validity remains unclear. We tested its validity and usefulness for the diagnosis of colorectal polyps by including colonoscopy experts and non-experts as test participants. METHODS: Forty NBI images of polyps of various JNET types were shown to 22 doctors (11 experts and 11 non-gastrointestinal [GI] trainees) who had not examined the patients. The doctors diagnosed the polyps based solely on the surface and vessel patterns in the magnified images and the JNET classification system. Concordance rates of their diagnoses with the pathological findings of the polyps were determined, and the results for experts and non-GI trainees were compared. RESULTS: Both for colonoscopy experts and non-GI trainees, the JNET classification system was particularly useful for classifying polyps as benign or malignant. Although the accuracy rates for classifying polyps into each JNET type varied among colonoscopy experts, those who were familiar with the JNET classification system were able to diagnose polyps with approximately 90% accuracy. Common mistakes were attributable to misunderstandings of the wording in the JNET classification chart and lack of proper training. CONCLUSION: The JNET classification system is a practical approach for the diagnosis of colorectal polyps. Training is required even for experienced colonoscopists to adopt the system properly. Common pitfalls must be shared among colonoscopists to improve the accuracy of the diagnosis.