Should We Resect and Discard Low Risk Diminutive Colon Polyps

Ecogastroenterology: cultivating sustainable clinical excellence in an environmentally conscious landscape

Gastrointestinal practices, especially endoscopy, have a substantial environmental impact, marked by notable greenhouse gas emissions and waste generation. As the world struggles with climate change, there emerges a pressing need to re-evaluate and reform the environmental footprint within gastrointestinal medicine. The challenge lies in finding a harmonious balance between ensuring clinical effectiveness and upholding environmental responsibility. This task involves recognising that the most significant reduction in the carbon footprint of endoscopy is achieved by avoiding unnecessary procedures; addressing the use of single-use endoscopes and accessories; and extending beyond the procedural suites to include clinics, virtual care, and conferences, among other aspects of gastrointestinal practice. The emerging digital realm in health care is crucial, given the potential environmental advantages of virtual gastroenterological care. Through an in-depth analysis, this review presents a path towards sustainable gastrointestinal practices, emphasising integrated strategies that prioritise both patient care and environmental stewardship.

Artificial intelligence for characterization of colorectal polyps: Prospective multicenter study

Background and study aims Optical diagnosis poses challenges to implementation of "resect and discard" strategies. This study aimed to assess the feasibility and performance of a new commercially available system for colorectal polyps. Patients and methods Nine expert endoscopists in three centers performed colonoscopies using artificial intelligence-equipped colonoscopes (CAD EYE, Fujifilm). Histology and predictions were compared, with hyperplastic polyps and sessile serrated lesions grouped for analysis. Results Overall, 253 polyps in 119 patients were documented (n=152 adenomas, n=78 hyperplastic polyps, n=23 sessile serrated lesions). CAD EYE detected polyps before endoscopists in 81 of 253 cases (32%). The mean polyp size was 5.5 mm (SD 0.6 mm). Polyp morphology was Paris Ip (4 %), Is (28 %), IIa (60 %), and IIb (8 %). CAD EYE achieved a sensitivity of 80%, specificity of 83%, positive predictive value (PPV) of 96%, and negative predictive value (NPV) of 72%. Expert endoscopists had a sensitivity of 88%, specificity of 83%, PPV of 96%, and NPV of 72%. Diagnostic accuracy was similar between CAD EYE (81%) and endoscopists (86%). However, sensitivity was greater with endoscopists as compared with CAD EYE ( P <0.05). CAD EYE classified sessile serrated lesions as hyperplasia in 22 of 23 cases, and endoscopists correctly classified 16 of 23 cases. Conclusions The CAD EYE system shows promise for detecting and characterizing colorectal polyps. Larger studies are needed, however, to confirm these findings.

The Utility of Narrow-Band Imaging International Colorectal Endoscopic Classification in Predicting the Histologies of Diminutive Colorectal Polyps Using I-Scan Optical Enhancement: A Prospective Study

(1) Background: This study aimed to evaluate the accuracy of predicting the histology of diminutive colonic polyps (DCPs) (≤5 mm) using i-scan optical enhancement (OE) based on the narrow-band imaging international colorectal endoscopic (NICE) classification. The study compared the diagnostic accuracy between experts who were already familiar with the NICE classification and trainees who were not, both before and after receiving brief training on the NICE classification. (2) Method: This prospective, single-center clinical trial was conducted at the Dong-A University Hospital from March 2020 to August 2020 and involved two groups of participants. The first group comprised two experienced endoscopists who were proficient in using i-scan OE and had received formal training in optical diagnosis and dye-less chromoendoscopy (DLC) techniques. The second group consisted of three endoscopists in the process of training in internal medicine at the Dong-A University Hospital. Each endoscopist examined the polyps and evaluated them using the NICE classification through i-scan OE. The results were not among the participants. Trained endoscopists were divided into pre- and post-training groups. (3) Results: During the study, a total of 259 DCPs were assessed using i-scan OE by the two expert endoscopists. They made real-time histological predictions according to the NICE classification criteria. For the trainee group, before training, the area under the receiver operating characteristic curves (AUROCs) for predicting histopathological results using i-scan OE were 0.791, 0.775, and 0.818. However, after receiving training, the AUROCs improved to 0.935, 0.949, and 0.963, which were not significantly different from the results achieved by the expert endoscopists. (4) Conclusions: This study highlights the potential of i-scan OE, along with the NICE classification, in predicting the histopathological results of DCPs during colonoscopy. In addition, this study suggests that even an endoscopist without experience in DLC can effectively use i-scan OE to improve diagnostic performance with only brief training.

Real-World Validation of a Computer-Aided Diagnosis System for Prediction of Polyp Histology in Colonoscopy: A Prospective Multicenter Study

INTRODUCTION

Computer-aided diagnosis (CADx) of polyp histology could support endoscopists in clinical decision-making. However, this has not been validated in a real-world setting.

METHODS

We performed a prospective, multicenter study comparing CADx and endoscopist predictions of polyp histology in real-time colonoscopy. Optical diagnosis based on visual inspection of polyps was made by experienced endoscopists. After this, the automated output from the CADx support tool was recorded. All imaged polyps were resected for histological assessment. Primary outcome was difference in diagnostic performance between CADx and endoscopist prediction of polyp histology. Subgroup analysis was performed for polyp size, bowel preparation, difficulty of location of the polyps, and endoscopist experience.

RESULTS

A total of 661 eligible polyps were resected in 320 patients aged ≥40 years between March 2021 and July 2022. CADx had an overall accuracy of 71.6% (95% confidence interval [CI] 68.0-75.0), compared with 75.2% (95% CI 71.7-78.4) for endoscopists ( P = 0.023). The sensitivity of CADx for neoplastic polyps was 61.8% (95% CI 56.9-66.5), compared with 70.3% (95% CI 65.7-74.7) for endoscopists ( P < 0.001). The interobserver agreement between CADx and endoscopist predictions of polyp histology was moderate (83.1% agreement, κ 0.661). When there was concordance between CADx and endoscopist predictions, the accuracy increased to 78.1%.

DISCUSSION

The overall diagnostic accuracy and sensitivity for neoplastic polyps was higher in experienced endoscopists compared with CADx predictions, with moderate interobserver agreement. Concordance in predictions increased this diagnostic accuracy. Further research is required to improve the performance of CADx and to establish its role in clinical practice.

Rationalising the use of specimen pots following colorectal polypectomy: a small step towards greener endoscopy

Aims

In this study, we aim to determine whether combining multiple small colorectal polyps within a single specimen pot can reduce carbon footprint, without an associated deleterious clinical impact.

Methods

This was a retrospective observational study of colorectal polyps resected during 2019, within the Imperial College Healthcare Trust. The numbers of pots for polypectomy specimens were calculated and corresponding histology results were extracted. We modelled the potential reduction in carbon footprint if all less than 10 mm polyps were sent together and the number of advanced lesions we would not be able to locate if we adopted this strategy. Carbon footprint was estimated based on previous study using a life-cycle assessment, at 0.28 kgCO2e per pot.

Results

A total of 11 781 lower gastrointestinal endoscopies were performed. There were 5125 polyps removed and 4192 pots used, equating to a carbon footprint of 1174 kgCO2e. There were 4563 (89%) polyps measuring 0-10 mm. 6 (0.1%) of these polyps were cancers, while 12 (0.2%) demonstrated high-grade dysplasia. If we combined all small polyps in a single pot, total pot usage could be reduced by one-third (n=2779).

Conclusion

A change in practice by placing small polyps collectively in one pot would have resulted in reduction in carbon footprint equivalent to 396 kgCO2e (emissions from 982 miles driven by an average passenger car). The reduction in carbon footprint from judicious use of specimen pots would be amplified with a change in practice on a national level.

Impact of the Volume and Distribution of Training Datasets in the Development of Deep-Learning Models for the Diagnosis of Colorectal Polyps in Endoscopy Images

Background: Establishment of an artificial intelligence model in gastrointestinal endoscopy has no standardized dataset. The optimal volume or class distribution of training datasets has not been evaluated. An artificial intelligence model was previously created by the authors to classify endoscopic images of colorectal polyps into four categories, including advanced colorectal cancer, early cancers/high-grade dysplasia, tubular adenoma, and nonneoplasm. The aim of this study was to evaluate the impact of the volume and distribution of training dataset classes in the development of deep-learning models for colorectal polyp histopathology prediction from endoscopic images. Methods: The same 3828 endoscopic images that were used to create earlier models were used. An additional 6838 images were used to find the optimal volume and class distribution for a deep-learning model. Various amounts of data volume and class distributions were tried to establish deep-learning models. The training of deep-learning models uniformly used no-code platform Neuro-T. Accuracy was the primary outcome on four-class prediction. Results: The highest internal-test classification accuracy in the original dataset, doubled dataset, and tripled dataset was commonly shown by doubling the proportion of data for fewer categories (2:2:1:1 for advanced colorectal cancer: early cancers/high-grade dysplasia: tubular adenoma: non-neoplasm). Doubling the proportion of data for fewer categories in the original dataset showed the highest accuracy (86.4%, 95% confidence interval: 85.0–97.8%) compared to that of the doubled or tripled dataset. The total required number of images in this performance was only 2418 images. Gradient-weighted class activation mapping confirmed that the part that the deep-learning model pays attention to coincides with the part that the endoscopist pays attention to. Conclusion: As a result of a data-volume-dependent performance plateau in the classification model of colonoscopy, a dataset that has been doubled or tripled is not always beneficial to training. Deep-learning models would be more accurate if the proportion of fewer category lesions was increased.

No-Code Platform-Based Deep-Learning Models for Prediction of Colorectal Polyp Histology from White-Light Endoscopy Images: Development and Performance Verification

Background: The authors previously developed deep-learning models for the prediction of colorectal polyp histology (advanced colorectal cancer, early cancer/high-grade dysplasia, tubular adenoma with or without low-grade dysplasia, or non-neoplasm) from endoscopic images. While the model achieved 67.3% internal-test accuracy and 79.2% external-test accuracy, model development was labour-intensive and required specialised programming expertise. Moreover, the 240-image external-test dataset included only three advanced and eight early cancers, so it was difficult to generalise model performance. These limitations may be mitigated by deep-learning models developed using no-code platforms. Objective: To establish no-code platform-based deep-learning models for the prediction of colorectal polyp histology from white-light endoscopy images and compare their diagnostic performance with traditional models. Methods: The same 3828 endoscopic images used to establish previous models were used to establish new models based on no-code platforms Neuro-T, VLAD, and Create ML-Image Classifier. A prospective multicentre validation study was then conducted using 3818 novel images. The primary outcome was the accuracy of four-category prediction. Results: The model established using Neuro-T achieved the highest internal-test accuracy (75.3%, 95% confidence interval: 71.0–79.6%) and external-test accuracy (80.2%, 76.9–83.5%) but required the longest training time. In contrast, the model established using Create ML-Image Classifier required only 3 min for training and still achieved 72.7% (70.8–74.6%) external-test accuracy. Attention map analysis revealed that the imaging features used by the no-code deep-learning models were similar to those used by endoscopists during visual inspection. Conclusion: No-code deep-learning tools allow for the rapid development of models with high accuracy for predicting colorectal polyp histology.

Non-optical polyp-based resect and discard strategy: A prospective clinical study

BACKGROUND

Post-polypectomy surveillance intervals are currently determined based on pathology results.

AIM

To evaluate a polyp-based resect and discard model that assigns surveillance intervals based solely on polyp number and size.

METHODS

Patients undergoing elective colonoscopies at the Montreal University Medical Center were enrolled prospectively. The polyp-based strategy was used to assign the next surveillance interval using polyp size and number. Surveillance intervals were also assigned using optical diagnosis for small polyps (< 10 mm). The primary outcome was surveillance interval agreement between the polyp-based model, optical diagnosis, and the pathology-based reference standard using the 2020 United States Multi-Society Task Force guidelines. Secondary outcomes included the proportion of reduction in required histopathology evaluations and proportion of immediate post-colonoscopy recommendations provided to patients.

RESULTS

Of 944 patients (mean age 62.6 years, 49.3% male, 933 polyps) were enrolled. The surveillance interval agreement for the polyp-based strategy was 98.0% [95% confidence interval (CI): 0.97-0.99] compared with pathology-based assignment. Optical diagnosis-based intervals achieved 95.8% (95%CI: 0.94-0.97) agreement with pathology. When using the polyp-based strategy and optical diagnosis, the need for pathology assessment was reduced by 87.8% and 70.6%, respectively. The polyp-based strategy provided 93.7% of patients with immediate surveillance interval recommendations vs 76.1% for optical diagnosis.

CONCLUSION

The polyp-based strategy achieved almost perfect surveillance interval agreement compared with pathology-based assignments, significantly reduced the number of required pathology evaluations, and provided most patients with immediate surveillance interval recommendations.

Artificial intelligence-assisted colonoscopy: a narrative review of current data and clinical applications

Colonoscopy is the reference standard procedure for the prevention and diagnosis of colorectal cancer, which is a leading cause of cancer-related deaths in Singapore. Artificial intelligence systems are automated, objective and reproducible. Artificial intelligence-assisted colonoscopy has recently been introduced into clinical practice as a clinical decision support tool. This review article provides a summary of the current published data and discusses ongoing research and current clinical applications of artificial intelligence-assisted colonoscopy.

Optical diagnosis of colorectal polyps using novel blue light imaging classification among trainee endoscopists

BACKGROUND

Blue light imaging (BLI) has been shown to improve the characterization of colorectal polyps among the endoscopy experts. We aimed to determine if this technology could be taught to endoscopy trainees while maintaining high accuracy and interobserver agreement.

METHODS

Twenty-one gastroenterology trainees (fellows) from two academic institutions participated in this prospective study. Each trainee completed a web-based learning comprising four modules: pre-test, didactic videos explaining the BLI Adenoma Serrated International Classification (BASIC), interactive examples, and post-test assessment. The pre- and post-test modules consisted of reviewing video images of colon polyps in high definition white light imaging and BLI and then applying the BASIC classification to determine if the polyps were likely to be adenomatous. Confidence in adenoma identification (rated '1' to '5'), accuracy in polyp (adenoma vs. non-adenoma) identification, and agreement in characterization per BASIC criteria were derived.

RESULTS

Trainee accuracy in the adenoma diagnosis improved from 74.7% (pre-test) to 85.4% (post-test) (P < 0.01). There was a trend towards higher accuracy in polyp characterization with subsequent years of training (1st year fellows 77.4%, 2nd year 88.5%, and final year 94.0%) with consistent improvements after the e-learning across years of trainees. Overall, trainees were able to identify adenoma with a high sensitivity of 86.9%, specificity 83.9%, positive predictive value of 84.4%, and negative predictive value of 86.5%. However, their interobserver agreement in adenoma diagnosis was moderate (k = 0.52).

CONCLUSION

The novel BLI classification can be easily taught to gastroenterology trainees using an online module and accuracy improves with years of training reaching >90% for colorectal polyp characterization.

Too Good to Be True? Evaluation of Colonoscopy Sensitivity Assumptions Used in Policy Models

BACKGROUND

Models can help guide colorectal cancer (CRC) screening policy. While models are carefully calibrated and validated, there is less scrutiny of assumptions about test performance.

METHODS

We examined the validity of the CRC-SPIN model and colonoscopy sensitivity assumptions. Standard sensitivity assumptions, consistent with published decision analyses, assume sensitivity equal to 0.75 for diminutive adenomas (<6mm), 0.85 for small adenomas (6-10mm), 0.95 for large adenomas ( {greater than or equal to} 10mm), and 0.95 for preclinical cancer. We also selected adenoma sensitivity that resulted in more accurate predictions. Targets were drawn from the Wheat Bran Fiber study. We examined how well the model predicted outcomes measured over a three-year follow-up period, including: the number of adenomas detected, the size of the largest adenoma detected, and incident CRC.

RESULTS

Using standard sensitivity assumptions, the model predicted adenoma prevalence that was too low (42.5% versus 48.9% observed, with 95% confidence interval 45.3%-50.7%) and detection of too few large adenomas (5.1% versus 14.% observed, with 95% confidence interval 11.8%-17.4%). Predictions were close to targets when we set sensitivities to 0.20 for diminutive adenomas, 0.60 for small adenomas, 0.80 for 10-20mm adenomas, and 0.98 for adenomas 20mm and larger.

CONCLUSIONS

Colonoscopy may be less accurate than currently assumed, especially for diminutive adenomas. Alternatively, the CRC-SPIN model may not accurately simulate onset and progression of adenomas in higher-risk populations.

IMPACT

Misspecification of either colonoscopy sensitivity or disease progression in high-risk populations may impact the predicted effectiveness of CRC screening. When possible, decision analyses used to inform policy should address these uncertainties.

Automatic image and text-based description for colorectal polyps using BASIC classification

Colorectal polyps (CRP) are precursor lesions of colorectal cancer (CRC). Correct identification of CRPs during in-vivo colonoscopy is supported by the endoscopist's expertise and medical classification models. A recent developed classification model is the Blue light imaging Adenoma Serrated International Classification (BASIC) which describes the differences between non-neoplastic and neoplastic lesions acquired with blue light imaging (BLI). Computer-aided detection (CADe) and diagnosis (CADx) systems are efficient at visually assisting with medical decisions but fall short at translating decisions into relevant clinical information. The communication between machine and medical expert is of crucial importance to improve diagnosis of CRP during in-vivo procedures. In this work, the combination of a polyp image classification model and a language model is proposed to develop a CADx system that automatically generates text comparable to the human language employed by endoscopists. The developed system generates equivalent sentences as the human-reference and describes CRP images acquired with white light (WL), blue light imaging (BLI) and linked color imaging (LCI). An image feature encoder and a BERT module are employed to build the AI model and an external test set is used to evaluate the results and compute the linguistic metrics. The experimental results show the construction of complete sentences with an established metric scores of BLEU-1 = 0.67, ROUGE-L = 0.83 and METEOR = 0.50. The developed CADx system for automatic CRP image captioning facilitates future advances towards automatic reporting and may help reduce time-consuming histology assessment.

Editors' Choice of Noteworthy Clinical Endoscopy Publications in the First Decade

This is a special review to celebrate the 10^th anniversary of Clinical Endoscopy. Each deputy editor has selected articles from one’s subspecialty that are significant in terms of the number of downloads, citations, and clinical importance. The articles included original articles, review articles, systematic reviews, and meta-analyses.

Computer-Aided Diagnosis of Diminutive Colorectal Polyps in Endoscopic Images: Systematic Review and Meta-analysis of Diagnostic Test Accuracy

Background

Most colorectal polyps are diminutive and benign, especially those in the rectosigmoid colon, and the resection of these polyps is not cost-effective. Advancements in image-enhanced endoscopy have improved the optical prediction of colorectal polyp histology. However, subjective interpretability and inter- and intraobserver variability prohibits widespread implementation. The number of studies on computer-aided diagnosis (CAD) is increasing; however, their small sample sizes limit statistical significance.

Objective

This review aims to evaluate the diagnostic test accuracy of CAD models in predicting the histology of diminutive colorectal polyps by using endoscopic images.

Methods

Core databases were searched for studies that were based on endoscopic imaging, used CAD models for the histologic diagnosis of diminutive colorectal polyps, and presented data on diagnostic performance. A systematic review and diagnostic test accuracy meta-analysis were performed.

Results

Overall, 13 studies were included. The pooled area under the curve, sensitivity, specificity, and diagnostic odds ratio of CAD models for the diagnosis of diminutive colorectal polyps (adenomatous or neoplastic vs nonadenomatous or nonneoplastic) were 0.96 (95% CI 0.93-0.97), 0.93 (95% CI 0.91-0.95), 0.87 (95% CI 0.76-0.93), and 87 (95% CI 38-201), respectively. The meta-regression analysis showed no heterogeneity, and no publication bias was detected. Subgroup analyses showed robust results. The negative predictive value of CAD models for the diagnosis of adenomatous polyps in the rectosigmoid colon was 0.96 (95% CI 0.95-0.97), and this value exceeded the threshold of the diagnosis and leave strategy.

Conclusions

CAD models show potential for the optical histological diagnosis of diminutive colorectal polyps via the use of endoscopic images.

Trial Registration

PROSPERO CRD42021232189; https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=232189

RNA-sequencing identification and validation of genes differentially expressed in high-risk adenoma, advanced colorectal cancer, and normal controls

Distinct gene expression patterns that occur during the adenoma-carcinoma sequence need to be determined to analyze the underlying mechanism in each step of colorectal cancer progression. Elucidation of biomarkers for colorectal polyps that harbor malignancy potential is important for prevention of colorectal cancer. Here, we use RNA sequencing to determine gene expression profile in patients with high-risk adenoma treated with endoscopic submucosal dissection by comparing with gene expression in patients with advanced colorectal cancer and normal controls. We collected 70 samples, which consisted of 27 colorectal polyps, 24 cancer tissues, and 19 normal colorectal mucosa. RNA sequencing was performed on an Illumina platform to select differentially expressed genes (DEGs) between colorectal polyps and cancer, polyps and controls, and cancer and normal controls. The Kyoto Gene and Genome Encyclopedia (KEGG) and gene ontology (GO) analysis, gene-concept network, GSEA, and a decision tree were used to evaluate the DEGs. We selected the most highly expressed genes in high-risk polyps and validated their expression using real-time PCR and immunohistochemistry. Compared to patients with colorectal cancer, 82 upregulated and 24 downregulated genes were detected in high-risk adenoma. In comparison with normal controls, 33 upregulated and 79 downregulated genes were found in high-risk adenoma. In total, six genes were retrieved as the highest and second highest expressed in advanced polyps and cancers among the three groups. Among the six genes, ANAX3 and CD44 expression in real-time PCR for validation was in good accordance with RNA sequencing. We identified differential expression of mRNAs among high-risk adenoma, advanced colorectal cancer, and normal controls, including that of CD44 and ANXA3, suggesting that this cluster of genes as a marker of high-risk colorectal adenoma.

Colon capsule endoscopy in colorectal cancer screening: a randomised controlled trial

Introduction

The use of capsule endoscopy has become an approved method in small bowel diagnostics, but the same level of integration is not seen in large bowel diagnostics. We will use colon capsule endoscopy (CCE) as a filter test in colorectal cancer (CRC) screening between the faecal immunochemical test (FIT) and colonoscopy. We aim to investigate the clinical performance, population acceptability, and economic implications of the procedure in a large-scale clinical trial.

Methods and analysis

We will randomly allocate 124 214 Danish citizens eligible for participation in the national CRC screening programme within the Region of Southern Denmark to either an intervention group or a control group. Prior to submitting a FIT, citizens randomised to the intervention group will be informed about their opportunity to undergo CCE, instead of colonoscopy, if the FIT is positive. Suspected cancers; >3 adenomas <10 mm in size, 1 adenoma >10 mm in size or >4 adenomas regardless of size, detected during CCE will generate an invitation to colonoscopy as per regular screening guidelines, whereas citizens with suspected low risk polyps will re-enter the biennial screening programme. Citizens with no CCE findings will be excluded from screening for 8 years. In the control group, citizens will follow standard screening procedures.

Ethics and dissemination

All participants must consent prior to capsule ingestion. All collected data will be handled and stored in accordance with current data protection legislation. Approvals from the regional ethics committee (ref. S-20190100) and the Danish data protection agency have been obtained (ref. 19/29858).

Trial registration details

The study has been registered with ClinicalTrials.gov under: NCT04049357.

Clinical validation of the SIMPLE classification for optical diagnosis of colorectal polyps

Background and study aims  A novel endoscopic optical diagnosis classification system (SIMPLE) has recently been developed. This study aimed to evaluate the SIMPLE classification in a clinical cohort. Patients and methods  All diminutive and small colorectal polyps found in a cohort of individuals undergoing screening, diagnostic, or surveillance colonoscopies underwent optical diagnosis using image-enhanced endoscopy (IEE) and the SIMPLE classification. The primary outcome was the agreement of surveillance intervals determined by optical diagnosis compared with pathology-based results for diminutive polyps. Secondary outcomes included the negative predictive value (NPV) for rectosigmoid adenomas, the percentage of pathology exams avoided, and the percentage of immediate surveillance interval recommendations. Analysis of optical diagnosis for polyps ≤ 10 mm was also performed. Results  399 patients (median age 62.6 years; 55.6 % female) were enrolled. For patients with at least one polyp ≤ 5 mm undergoing optical diagnosis, agreement with pathology-based surveillance intervals was 93.5 % (95 % confidence interval [CI] 91.4-95.6). The NPV for rectosigmoid adenomas was 86.7 % (95 %CI 77.5-93.2). When using optical diagnosis, pathology analysis could be avoided in 61.5 % (95 %CI 56.9-66.2) of diminutive polyps, and post-colonoscopy surveillance intervals could be given immediately to 70.9 % (95 %CI 66.5-75.4) of patients. For patients with at least one ≤ 10 mm polyp, agreement with pathology-based surveillance intervals was 92.7 % (95 %CI 89.7-95.1). NPV for rectosigmoid adenomas ≤ 10 mm was 85.1 % (95 %CI CI 76.3-91.6). Conclusions  IEE with the SIMPLE classification achieved the quality benchmark for the resect and discard strategy; however, the NPV for rectosigmoid polyps requires improvement.

Colonoscopy and artificial intelligence: Bridging the gap or a gap needing to be bridged?

Research in artificial intelligence (AI) in gastroenterology has increased over the last decade. Colonoscopy represents the most widely published field with regards to its use in gastroenterology. Most studies to date center on polyp detection and characterization, as well as real-time evaluation of adequacy of mucosal exposure for inspection. This review article discusses how advances in AI has bridged certain gaps in colonoscopy. In addition, the gaps formed with the development of AI that currently prevent its routine use in colonoscopy will be explored.

A CNN CADx System for Multimodal Classification of Colorectal Polyps Combining WL, BLI, and LCI Modalities

Colorectal polyps are critical indicators of colorectal cancer (CRC). Blue Laser Imaging and Linked Color Imaging are two modalities that allow improved visualization of the colon. In conjunction with the Blue Laser Imaging (BLI) Adenoma Serrated International Classification (BASIC) classification, endoscopists are capable of distinguishing benign and pre-malignant polyps. Despite these advancements, this classification still prevails a high misclassification rate for pre-malignant colorectal polyps. This work proposes a computer aided diagnosis (CADx) system that exploits the additional information contained in two novel imaging modalities, enabling more informative decision-making during colonoscopy. We train and benchmark six commonly used CNN architectures and compare the results with 19 endoscopists that employed the standard clinical classification model (BASIC). The proposed CADx system for classifying colorectal polyps achieves an area under the curve (AUC) of 0.97. Furthermore, we incorporate visual explanatory information together with a probability score, jointly computed from White Light, Blue Laser Imaging, and Linked Color Imaging. Our CADx system for automatic polyp malignancy classification facilitates future advances towards patient safety and may reduce time-consuming and costly histology assessment.