Clinical consequences of computer-aided colorectal polyp detection.

BACKGROUND AND AIM: Randomised trials show improved polyp detection with computer-aided detection (CADe), mostly of small lesions. However, operator and selection bias may affect CADe's true benefit. Clinical outcomes of increased detection have not yet been fully elucidated. METHODS: In this multicentre trial, CADe combining convolutional and recurrent neural networks was used for polyp detection. Blinded endoscopists were monitored in real time by a second observer with CADe access. CADe detections prompted reinspection. Adenoma detection rates (ADR) and polyp detection rates were measured prestudy and poststudy. Histological assessments were done by independent histopathologists. The primary outcome compared polyp detection between endoscopists and CADe. RESULTS: In 946 patients (51.9% male, mean age 64), a total of 2141 polyps were identified, including 989 adenomas. CADe was not superior to human polyp detection (sensitivity 94.6% vs 96.0%) but outperformed them when restricted to adenomas. Unblinding led to an additional yield of 86 true positive polyp detections (1.1% ADR increase per patient; 73.8% were <5 mm). CADe also increased non-neoplastic polyp detection by an absolute value of 4.9% of the cases (1.8% increase of entire polyp load). Procedure time increased with 6.6±6.5 min (+42.6%). In 22/946 patients, the additional detection of adenomas changed surveillance intervals (2.3%), mostly by increasing the number of small adenomas beyond the cut-off. CONCLUSION: Even if CADe appears to be slightly more sensitive than human endoscopists, the additional gain in ADR was minimal and follow-up intervals rarely changed. Additional inspection of non-neoplastic lesions was increased, adding to the inspection and/or polypectomy workload.

Expected value of artificial intelligence in gastrointestinal endoscopy: European Society of Gastrointestinal Endoscopy (ESGE) Position Statement.

This ESGE Position Statement defines the expected value of artificial intelligence (AI) for the diagnosis and management of gastrointestinal neoplasia within the framework of the performance measures already defined by ESGE. This is based on the clinical relevance of the expected task and the preliminary evidence regarding artificial intelligence in artificial or clinical settings. MAIN RECOMMENDATIONS:: (1) For acceptance of AI in assessment of completeness of upper GI endoscopy, the adequate level of mucosal inspection with AI should be comparable to that assessed by experienced endoscopists. (2) For acceptance of AI in assessment of completeness of upper GI endoscopy, automated recognition and photodocumentation of relevant anatomical landmarks should be obtained in ≥90% of the procedures. (3) For acceptance of AI in the detection of Barrett's high grade intraepithelial neoplasia or cancer, the AI-assisted detection rate for suspicious lesions for targeted biopsies should be comparable to that of experienced endoscopists with or without advanced imaging techniques. (4) For acceptance of AI in the management of Barrett's neoplasia, AI-assisted selection of lesions amenable to endoscopic resection should be comparable to that of experienced endoscopists. (5) For acceptance of AI in the diagnosis of gastric precancerous conditions, AI-assisted diagnosis of atrophy and intestinal metaplasia should be comparable to that provided by the established biopsy protocol, including the estimation of extent, and consequent allocation to the correct endoscopic surveillance interval. (6) For acceptance of artificial intelligence for automated lesion detection in small-bowel capsule endoscopy (SBCE), the performance of AI-assisted reading should be comparable to that of experienced endoscopists for lesion detection, without increasing but possibly reducing the reading time of the operator. (7) For acceptance of AI in the detection of colorectal polyps, the AI-assisted adenoma detection rate should be comparable to that of experienced endoscopists. (8) For acceptance of AI optical diagnosis (computer-aided diagnosis [CADx]) of diminutive polyps (≤5 mm), AI-assisted characterization should match performance standards for implementing resect-and-discard and diagnose-and-leave strategies. (9) For acceptance of AI in the management of polyps ≥ 6 mm, AI-assisted characterization should be comparable to that of experienced endoscopists in selecting lesions amenable to endoscopic resection.

Establishing key research questions for the implementation of artificial intelligence in colonoscopy: a modified Delphi method.

BACKGROUND : Artificial intelligence (AI) research in colonoscopy is progressing rapidly but widespread clinical implementation is not yet a reality. We aimed to identify the top implementation research priorities. METHODS : An established modified Delphi approach for research priority setting was used. Fifteen international experts, including endoscopists and translational computer scientists/engineers, from nine countries participated in an online survey over 9 months. Questions related to AI implementation in colonoscopy were generated as a long-list in the first round, and then scored in two subsequent rounds to identify the top 10 research questions. RESULTS : The top 10 ranked questions were categorized into five themes. Theme 1: clinical trial design/end points (4 questions), related to optimum trial designs for polyp detection and characterization, determining the optimal end points for evaluation of AI, and demonstrating impact on interval cancer rates. Theme 2: technological developments (3 questions), including improving detection of more challenging and advanced lesions, reduction of false-positive rates, and minimizing latency. Theme 3: clinical adoption/integration (1 question), concerning the effective combination of detection and characterization into one workflow. Theme 4: data access/annotation (1 question), concerning more efficient or automated data annotation methods to reduce the burden on human experts. Theme 5: regulatory approval (1 question), related to making regulatory approval processes more efficient. CONCLUSIONS : This is the first reported international research priority setting exercise for AI in colonoscopy. The study findings should be used as a framework to guide future research with key stakeholders to accelerate the clinical implementation of AI in endoscopy.

Artificial intelligence and its impact on quality improvement in upper and lower gastrointestinal endoscopy.

Artificial intelligence (AI) and its application in medicine has grown large interest. Within gastrointestinal (GI) endoscopy, the field of colonoscopy and polyp detection is the most investigated, however, upper GI follows the lead. Since endoscopy is performed by humans, it is inherently an imperfect procedure. Computer-aided diagnosis may improve its quality by helping prevent missing lesions and supporting optical diagnosis for those detected. An entire evolution in AI systems has been established in the last decades, resulting in optimization of the diagnostic performance with lower variability and matching or even outperformance of expert endoscopists. This shows a great potential for future quality improvement of endoscopy, given the outstanding diagnostic features of AI. With this narrative review, we highlight the potential benefit of AI to improve overall quality in daily endoscopy and describe the most recent developments for characterization and diagnosis as well as the recent conditions for regulatory approval.

Automatic, computer-aided determination of endoscopic and histological inflammation in patients with mild to moderate ulcerative colitis based on red density.

BACKGROUND: The objective evaluation of endoscopic disease activity is key in ulcerative colitis (UC). A composite of endoscopic and histological factors is the goal in UC treatment. We aimed to develop an operator-independent computer-based tool to determine UC activity based on endoscopic images. METHODS: First, we built a computer algorithm using data from 29 consecutive patients with UC and 6 healthy controls (construction cohort). The algorithm (red density: RD) was based on the red channel of the red-green-blue pixel values and pattern recognition from endoscopic images. The algorithm was refined in sequential steps to optimise correlation with endoscopic and histological disease activity. In a second phase, the operating properties were tested in patients with UC flares requiring treatment escalation. To validate the algorithm, we tested the correlation between RD score and clinical, endoscopic and histological features in a validation cohort. RESULTS: We constructed the algorithm based on the integration of pixel colour data from the redness colour map along with vascular pattern detection. These data were linked with Robarts histological index (RHI) in a multiple regression analysis. In the construction cohort, RD correlated with RHI (r=0.74, p<0.0001), Mayo endoscopic subscores (r=0.76, p<0.0001) and UC Endoscopic Index of Severity scores (r=0.74, p<0.0001). The RD sensitivity to change had a standardised effect size of 1.16. In the validation set, RD correlated with RHI (r=0.65, p=0.00002). CONCLUSIONS: RD provides an objective computer-based score that accurately assesses disease activity in UC. In a validation study, RD correlated with endoscopic and histological disease activity.

Pitfalls in training and validation of deep learning systems.

The number of publications in endoscopic journals that present deep learning applications has risen tremendously over the past years. Deep learning has shown great promise for automated detection, diagnosis and quality improvement in endoscopy. However, the interdisciplinary nature of these works has undoubtedly made it more difficult to estimate their value and applicability. In this review, the pitfalls and common misconducts when training and validating deep learning systems are discussed and some practical guidelines are proposed that should be taken into account when acquiring data and handling it to ensure an unbiased system that will generalize for application in routine clinical practice. Finally, some considerations are presented to ensure correct validation and comparison of AI systems.

Optimization for Medical Image Segmentation: Theory and Practice When Evaluating With Dice Score or Jaccard Index.

In many medical imaging and classical computer vision tasks, the Dice score and Jaccard index are used to evaluate the segmentation performance. Despite the existence and great empirical success of metric-sensitive losses, i.e. relaxations of these metrics such as soft Dice, soft Jaccard and Lovász-Softmax, many researchers still use per-pixel losses, such as (weighted) cross-entropy to train CNNs for segmentation. Therefore, the target metric is in many cases not directly optimized. We investigate from a theoretical perspective, the relation within the group of metric-sensitive loss functions and question the existence of an optimal weighting scheme for weighted cross-entropy to optimize the Dice score and Jaccard index at test time. We find that the Dice score and Jaccard index approximate each other relatively and absolutely, but we find no such approximation for a weighted Hamming similarity. For the Tversky loss, the approximation gets monotonically worse when deviating from the trivial weight setting where soft Tversky equals soft Dice. We verify these results empirically in an extensive validation on six medical segmentation tasks and can confirm that metric-sensitive losses are superior to cross-entropy based loss functions in case of evaluation with Dice Score or Jaccard Index. This further holds in a multi-class setting, and across different object sizes and foreground/background ratios. These results encourage a wider adoption of metric-sensitive loss functions for medical segmentation tasks where the performance measure of interest is the Dice score or Jaccard index.