Computer-aided diagnosis (CADx) is anticipated to enhance the prediction of colorectal polyp histology. This study aims to compare the diagnostic accuracy of CADx in the optical diagnosis of colorectal polyps, evaluating its performance against that of both experienced and inexperienced endoscopists.

The protocol of this study was registered in the International Prospective Register of Systematic Reviews (PROSPERO) (ID: CRD42024585097). Three electronic databases including MEDLINE, Embase, and the Cochrane Central Register of Controlled Trials (CENTRAL) were searched in September 2024. A bivariate random effects model was employed. The primary outcome was the comparison of sensitivity and specificity between CADx and experienced endoscopists; the secondary outcome was the comparison between CADx and inexperienced endoscopists.

Twenty-one studies involving 5477 polyps were included. The pooled sensitivities of CADx and experienced endoscopists were 0.87 (95% confidence interval [CI] 0.82-0.91) and 0.88 (95% CI 0.83-0.91), respectively (P = 0.93). The pooled specificities of CADx and experienced endoscopists were 0.85 (95% CI 0.78-0.90) and 0.87 (95% CI 0.82-0.92), respectively (P = 0.53). In nine studies comparing CADx with inexperienced endoscopists, the pooled sensitivities were 0.88 (95% CI 0.82-0.92) for CADx and 0.85 (95% CI 0.78-0.90) for inexperienced endoscopists (P = 0.46). The pooled specificities were 0.84 (95% CI 0.78-0.88) for CADx and 0.77 (95% CI 0.70-0.83) for inexperienced endoscopists (P = 0.16).

Computer-aided diagnosis does not demonstrate superior diagnostic accuracy in optical diagnosis of colorectal polyps compared to endoscopists, regardless of their experience level.

Colorectal cancer is one of the most prevalent cancers worldwide, highlighting the critical need for early and accurate diagnosis to reduce patient risks. Inaccurate diagnoses not only compromise patient outcomes but also lead to increased costs and additional time burdens for clinicians. Enhancing diagnostic accuracy is essential, and this study focuses on improving the accuracy of polyp classification using the NICE classification, which evaluates three key features: colour, vessels, and surface pattern.

A multiclass classifier was developed and trained to independently classify each of the three features in the NICE classification. The approach prioritizes clinically relevant features rather than relying on handcrafted or obscure deep learning features, ensuring transparency and reliability for clinical use. The classifier was trained on internal datasets and tested on both internal and public datasets.

The classifier successfully classified the three polyp features, achieving an accuracy of over 92% on internal datasets and exceeding 88% on a public dataset. The high classification accuracy demonstrates the system's effectiveness in identifying the key features from the NICE classification.

This study underscores the potential of using an independent classification approach for NICE features to enhance clinical decision-making in colorectal cancer diagnosis. The method shows promise in improving diagnostic accuracy, which could lead to better patient outcomes and more efficient clinical workflows.

The diagnostic performance of artificial intelligence (AI)-assisted endoscopy for digestive tumors remains controversial. The objective of this umbrella review was to summarize the comprehensive evidence for the AI-assisted endoscopic diagnosis of digestive system tumors. We grouped the evidence according to the location of each digestive system tumor and performed separate subgroup analyses on the basis of the method of data collection and form of the data. We also compared the diagnostic performance of AI with that of experts and nonexperts. For early digestive system cancer and precancerous lesions, AI showed a high diagnostic performance in capsule endoscopy and esophageal squamous cell carcinoma. Additionally, AI-assisted endoscopic ultrasonography (EUS) had good diagnostic accuracy for pancreatic cancer. In the subgroup analysis, AI had a better diagnostic performance than experts for most digestive system tumors. However, the diagnostic performance of AI using video data requires improvement.

Since it is difficult for expert endoscopists to diagnose early gastric cancer in post-eradication status, it may be critical to evaluate the stratification of high-risk groups using the advance of gastric atrophy or intestinal metaplasia. We tried to determine whether the combination of endoscopic artificial intelligence (AI) diagnosis for the evaluation of gastric atrophy could be a useful tool in both pre- and post-eradication status.

270 Helicobacter pylori-positive outpatients (Study I) were enrolled and Study II was planned to compare patients (n = 72) with pre-eradication therapy with post-eradication therapy. Assessment of endoscopic appearance was evaluated by the Kyoto classification and Kimura-Takemoto classification. The trained neural network generated a continuous number between 0 and 1 for gastric atrophy.

There were significant associations between the severity of gastric atrophy determined by AI endoscopic diagnosis and not having a regular arrangement of collecting venules in angle, visibility of vascular pattern, and mucus using Kyoto classification in H. pylori-positive gastritis. There were significant differences (p = 0.037 and p = 0.014) in the severity of gastric atrophy between the high-risk group and low-risk group based on the combination of Kimura-Takemoto classification and endoscopic AI diagnosis in pre- and post-eradication status. The area under the curve values of the severity of gastric atrophy (0.674) determined by the combination of Kimura-Takemoto classification and gastric atrophy determined by AI diagnosis was higher than that determined by Kimura-Takemoto classification alone in post-eradication status.

A combination of gastric atrophy determined by AI endoscopic diagnosis and Kimura-Takemoto classification may be a useful tool for the prediction of high-risk groups in post-eradication status.

Adenomatous polyps, a common premalignant lesion, are often classified into villous adenoma (VA) and tubular adenoma (TA). VA has a higher risk of malignancy, whereas TA typically grows slowly and has a lower likelihood of cancerous transformation. Accurate classification is essential for tailored treatment. In this study, we develop a deep learning-based approach for the localization and classification of adenomatous polyps using endoscopic images. Specifically, a pre-trained EGE-UNet is first adopted to extract regions of interest from original images. Multi-level feature maps are then extracted by the feature extraction pipeline (FEP). The deep-level features are fed into the Pyramid Pooling Module (PPM) to capture global contextual information, and the squeeze body edge (SBE) module is then used to decouple the body and edge parts of features, enabling separate analysis of their distinct characteristics. The Group Aggregation Bridge (GAB) and Boundary Enhancement Module (BEM) are then applied to enhance the body features and edge features, respectively, emphasizing their structural and morphological characteristics. By combining the features of the body and edge parts, the final output can be obtained. Experiments show the proposed method achieved promising results on two private datasets. For adenoma vs. non-adenoma classification, It achieved a mIoU of 91.41%, mPA of 96.33%, mHD of 11.63, and mASD of 2.33. For adenoma subclassification (non-adenomas vs. villous adenomas vs. tubular adenomas), it achieved a mIoU of 91.21%, mPA of 94.83%, mHD of 13.75, and mASD of 2.56. These results demonstrate the potential of our approach for precise adenomatous polyp classification.

Machine learning and its specialized forms, such as Artificial Neural Networks and Convolutional Neural Networks, are increasingly being used for detecting and managing gastrointestinal conditions. Recent advancements involve using Artificial Neural Network models to enhance predictive accuracy for severe lower gastrointestinal (LGI) bleeding outcomes, including the need for surgery. To this end, artificial intelligence (AI)-guided predictive models have shown promise in improving management outcomes. While much literature focuses on AI in early neoplasia detection, this review highlights AI's role in managing LGI and small bowel disorders, including risk stratification for LGI bleeding, quality control, evaluation of inflammatory bowel disease, and video capsule endoscopy reading. Overall, the integration of AI into routine clinical practice is still developing, with ongoing research aimed at addressing current limitations and gaps in patient care.

Osteoarthritis (OA), characterized by progressive degeneration of cartilage and reactive proliferation of subchondral bone, stands as a prevalent condition in orthopedic clinics. However, the precise mechanisms underlying OA pathogenesis remain inadequately explored.

In this study, Random Forest (RF), Least Absolute Shrinkage and Selection Operator (LASSO), and Support Vector Machine-Recursive Feature Elimination (SVM-RFE) machine learning techniques were employed to identify hub genes. Based on these hub genes, an Artificial Neural Network (ANN) diagnostic model was constructed. The Drug Signatures Database (DSigDB) was utilized to screen small-molecule drugs targeting these hub genes, and molecular docking analyses and molecular dynamics simulations were employed to explore and validate the binding interactions between proteins and small-molecule drugs. Expression changes of the hub genes under inflammatory conditions were validated through in vitro experiments, including RT-qPCR and Western blotting, and the therapeutic effects of the identified small-molecule drug on chondrocytes under inflammatory conditions were further verified in vitro. Lastly, Mendelian randomization analysis was conducted to examine the causal association between progesterone levels and various OA phenotypes.

In this study, we identified three hub genes: interleukin 1 receptor-associated kinase 3 (IRAK3), integrin subunit beta-like 1 (ITGBL1), and Ras homolog family member U (RHOU). An Artificial Neural Network (ANN) diagnostic model constructed based on these hub genes demonstrated excellent performance in both training and validation phases. Screening with the Drug Signatures Database (DSigDB) identified progesterone as a small-molecule drug targeting these key proteins. Molecular docking analysis using AutoDock Vina revealed that progesterone exhibited binding energies of ≤ -7 kcal/mol with each of the key proteins, indicating strong binding affinity. Furthermore, molecular dynamics simulations validated the stability and strength of these interactions. RT-qPCR and Western blotting confirmed the downregulation of the hub genes in IL-1β-treated chondrocytes. Western blotting also demonstrated the potential therapeutic effects of progesterone on IL-1β-treated chondrocytes. Finally, Mendelian randomization analysis established a significant association between progesterone levels and multiple OA phenotypes.

In our study, IRAK3, ITGBL1, and RHOU were identified as potential novel diagnostic and therapeutic targets for OA. Progesterone was preliminarily validated as a small-molecule drug with potential effects on OA. Further research is crucial to elucidate the pathogenesis of OA and the specific therapeutic mechanisms involved.

Gastrointestinal endoscopic image analysis presents significant challenges, such as considerable variations in quality due to the challenging in-body imaging environment, the often-subtle nature of abnormalities with low interobserver agreement, and the need for real-time processing. These challenges pose strong requirements on the performance, generalization, robustness and complexity of deep learning-based techniques in such safety-critical applications. While Convolutional Neural Networks (CNNs) have been the go-to architecture for endoscopic image analysis, recent successes of the Transformer architecture in computer vision raise the possibility to update this conclusion. To this end, we evaluate and compare clinically relevant performance, generalization and robustness of state-of-the-art CNNs and Transformers for neoplasia detection in Barrett's esophagus. We have trained and validated several top-performing CNNs and Transformers on a total of 10,208 images (2,079 patients), and tested on a total of 7,118 images (998 patients) across multiple test sets, including a high-quality test set, two internal and two external generalization test sets, and a robustness test set. Furthermore, to expand the scope of the study, we have conducted the performance and robustness comparisons for colonic polyp segmentation (Kvasir-SEG) and angiodysplasia detection (Giana). The results obtained for featured models across a wide range of training set sizes demonstrate that Transformers achieve comparable performance as CNNs on various applications, show comparable or slightly improved generalization capabilities and offer equally strong resilience and robustness against common image corruptions and perturbations. These findings confirm the viability of the Transformer architecture, particularly suited to the dynamic nature of endoscopic video analysis, characterized by fluctuating image quality, appearance and equipment configurations in transition from hospital to hospital. The code is made publicly available at: https://github.com/BONS-AI-VCA-AMC/Endoscopy-CNNs-vs-Transformers.

The 6-min withdrawal time for colonoscopy is widely considered the standard of care. However, there may not be appropriate if the 6-min is equally divided into various colon segments. Since the adenoma detection in each colon segment is not the same, there may be differences with the withdrawal time in different colon segments. Our objective was to evaluate the relationships between adenoma detection rate (ADR) and respective withdrawal time in different colon segments.

Outpatients, age range 18-75 years, undertaking complete colonoscopy were enrolled in this study from November 2019 to November 2020 in the digestive endoscopy center. The entire colon was divided into four different segments: ascending colon, transverse colon, descending colon and rectosigmoid colon. The respective withdrawal time and ADR in each colon segment were recorded respectively.

A total of 586 outpatients (279 males, 307 females) enrolled in this study and the general ADR was 38.2%. The positive withdrawal time (adenomas detected) was longer than negative withdrawal time (non-adenomas detected) (334.04 ± 24.21 s vs. 303.65 ± 5.20 s, t = 1.26, p < 0.001). ADR in ascending colon, transverse colon, descending colon and rectosigmoid colon were respectively 30.5%, 2.9%, 3.1% and 7.5%. While all of their positive withdrawal time were longer than negative withdrawal time (94.34 ± 33.76 s vs. 70.40 ± 41.84 s, t = 3.31, p = 0.001; 85.40 ± 49.76 s vs. 71.66 ± 36.87 s, t = 1.95, p = 0.025; 80.29 ± 39.85 s vs. 69.73 ± 35.96 s, t = 1.40, p = 0.016;100.95 ± 55.92 s vs. 80.96 ± 42.87 s, t = 3.61; p < 0.001, respectively). The withdrawal time threshold in the ascending colon, transverse colon, descending colon, rectosigmoid colon determined by receiver operating characteristic (ROC) curve were 77, 61, 56 and 109 s, respectively. In the ascending colon, ADR was significantly higher (47.0% vs. 33.1%, p < 0.001) when the colonoscopy withdrawal time was ≥ 77 s. When the withdrawal time was ≥ 61 s in the transverse colon (42.7% vs. 32.7%, p = 0.013), ≥ 59 s in the descending colon (42.3% vs. 29.9%, p = 0.004) and ≥ 109 s in rectosigmoid colon (52.2% vs. 33.9%, p < 0.001), ADR was also significantly higher. After adjusting for age, sex and BMI, Logistic regression analysis showed that withdrawal time ≥ 77 s in the ascending colon (OR, 1.796; 95% CI, 1.273-2.532; p < 0.001), ≥ 61 s in the transverse colon (OR, 1.535; 95% CI, 1.094-2.155; p = 0.013), ≥ 56 s in the descending colon (OR, 1.722; 95% CI, 1.193-2.486; p = 0.004) and ≥ 109 s in the rectosigmoid colon (OR, 2.134; 95% CI, 1.446-2.350; p < 0.001) were independent risk factors for the increase of ADR.

ADR and withdrawal time are all various in individual colon segments. During the operation of colonoscopy, withdrawal time in the ascending colon may be shortened appropriately. The adenomas in the rectosigmoid colon are more likely to be detected and do not take longer withdrawal times. We need to choose the appropriate time according to different colon segments.

The methodology in colon polyp labeling in establishing database for ma-chine learning is not well-described and standardized. We aimed to find out the best annotation method to generate the most accurate model in polyp detection.

3542 colonoscopy polyp images were obtained from endoscopy database of a tertiary medical center. Two experienced endoscopists manually annotated the polyp with (1) exact outline segmentation and (2) using a standard rectangle box close to the polyp margin, and extending 10%, 20%, 30%, 40% and 50% longer in both width and length of the standard rectangle for AI modeling setup. The images were randomly divided into training and validation sets in 4:1 ratio. U-Net convolutional network architecture was used to develop automatic segmentation machine learning model. Another unrelated verification set was established to evaluate the performance of polyp detection by different segmentation methods.

Extending the bounding box to 20% of the polyp margin represented the best performance in accuracy (95.42%), sensitivity (94.84%) and F1-score (95.41%). Exact outline segmentation model showed the excellent performance in sensitivity (99.6%) and the worst precision (77.47%). The 20% model was the best among the 6 models. (confidence interval = 0.957-0.985; AUC = 0.971).

Labelling methodology affect the predictability of AI model in polyp detection. Extending the bounding box to 20% of the polyp margin would result in the best polyp detection predictive model based on AUC data. It is mandatory to establish a standardized way in colon polyp labeling for comparison of the precision of different AI models.

Artificial intelligence (AI) has significantly impacted medical imaging, particularly in gastrointestinal endoscopy. Computer-aided detection and diagnosis systems (CADe and CADx) are thought to enhance the quality of colonoscopy procedures.

Colonoscopy is essential for colorectal cancer screening but often misses a significant percentage of adenomas. AI-assisted systems employing deep learning offer improved detection and differentiation of colorectal polyps, potentially increasing adenoma detection rates by 8%-10%. The main benefit of CADe is in detecting small adenomas, whereas it has a limited impact on advanced neoplasm detection. Recent advancements include real-time CADe systems and CADx for histopathological predictions, aiding in the differentiation of neoplastic and nonneoplastic lesions. Biases such as the Hawthorne effect and potential overdiagnosis necessitate large-scale clinical trials to validate the long-term benefits of AI. Additionally, novel concepts such as computer-aided quality improvement systems are emerging to address limitations facing current CADe systems.

Despite the potential of AI for enhancing colonoscopy outcomes, its effectiveness in reducing colorectal cancer incidence and mortality remains unproven. Further prospective studies are essential to establish the overall utility and clinical benefits of AI in colonoscopy.

Recent advancements in artificial intelligence (AI) have significantly impacted the field of gastrointestinal (GI) endoscopy, with applications spanning a wide range of clinical indications. The central goals for AI in GI endoscopy are to improve endoscopic procedural performance and quality assessment, optimize patient outcomes, and reduce administrative burden. Despite early progress, such as Food and Drug Administration approval of the first computer-aided polyp detection system in 2021, there are numerous important challenges to be faced on the path toward broader adoption of AI algorithms in clinical endoscopic practice.

The development of new image enhancement modalities and improved endoscopic imaging quality has not led to increased adoption of resect-and-discard in routine practice. Studies have shown that endoscopists have the capacity to achieve quality thresholds to perform optical diagnosis; however, this has not led to acceptance of optical diagnosis as a replacement for pathology for diminutive (1-5 mm) polyps. In recent years, artificial intelligence (AI)-based computer-assisted characterization of diminutive polyps has recently emerged as a strategy that could potentially represent a breakthrough technology to enable widespread adoption of resect-and-discard. Recent evidence suggests that pathology-based diagnosis is suboptimal, as polyp nonretrieval, fragmentation, sectioning errors, incorrect diagnosis as "normal mucosa," and interpathologist variability limit the efficacy of pathology for the diagnosis of 1-5 mm polyps. New paradigms in performing polyp diagnosis with or without AI have emerged to compete with pathology in terms of efficacy. Strategies, such as autonomous AI, AI-assisted human diagnosis, AI-unassisted human diagnosis, and combined strategies have been proposed as potential paradigms for resect-and-discard, although further research is still required to determine the optimal strategy. Implementation studies with high patient acceptance, where polyps are truly being discarded without histologic diagnosis, are paving the way toward normalizing resect-and-discard in routine clinical practice. Ultimately the largest challenges for computer-assisted characterization remain liability perceptions from endoscopists. The potential benefits of AI-based resect-and-discard are many, with very little potential harm. Real-world implementation studies are therefore required to pave the way for the acceptability of such strategies in routine practice.

Artificial intelligence (AI) has emerged as a transformative tool across several specialties, namely gastroenterology, where it has the potential to optimize both diagnosis and treatment as well as enhance patient care. Coloproctology, due to its highly prevalent pathologies and tremendous potential to cause significant mortality and morbidity, has drawn a lot of attention regarding AI applications. In fact, its application has yielded impressive outcomes in various domains, colonoscopy being one prominent example, where it aids in the detection of polyps and early signs of colorectal cancer with high accuracy and efficiency. With a less explored path but equivalent promise, AI-powered capsule endoscopy ensures accurate and time-efficient video readings, already detecting a wide spectrum of anomalies. High-resolution anoscopy is an area that has been growing in interest in recent years, with efforts being made to integrate AI. There are other areas, such as functional studies, that are currently in the early stages, but evidence is expected to emerge soon. According to the current state of research, AI is anticipated to empower gastroenterologists in the decision-making process, paving the way for a more precise approach to diagnosing and treating patients. This review aims to provide the state-of-the-art use of AI in coloproctology while also reflecting on future directions and perspectives.

Colorectal cancer is a significant global health concern, necessitating effective screening strategies to reduce its incidence and mortality rates. Colonoscopy plays a crucial role in the detection and removal of colorectal neoplastic precursors. However, there are limitations and variations in the performance of endoscopists, leading to missed lesions and suboptimal outcomes. The emergence of artificial intelligence (AI) in endoscopy offers promising opportunities to improve the quality and efficacy of screening colonoscopies. In particular, AI applications, including computer-aided detection (CADe) and computer-aided characterization (CADx), have demonstrated the potential to enhance adenoma detection and optical diagnosis accuracy. Additionally, AI-assisted quality control systems aim to standardize the endoscopic examination process. This narrative review provides an overview of AI principles and discusses the current knowledge on AI-assisted endoscopy in the context of screening colonoscopies. It highlights the significant role of AI in improving lesion detection, characterization, and quality assurance during colonoscopy. However, further well-designed studies are needed to validate the clinical impact and cost-effectiveness of AI-assisted colonoscopy before its widespread implementation.

Precision endoscopy in the management of colorectal polyps and early colorectal cancer has emerged as the standard of care. It includes optical characterization of polyps and estimation of submucosal invasion depth of large nonpedunculated colorectal polyps to select the appropriate endoscopic resection modality. Over time, several imaging modalities have been implemented in endoscopic practice to improve optical performance. Among these, image-enhanced endoscopy systems and magnification endoscopy represent now well-established tools. New advanced technologies, such as endocytoscopy and confocal laser endomicroscopy, have recently shown promising results in predicting the histology of colorectal polyps. In recent years, artificial intelligence has continued to enhance endoscopic performance in the characterization of colorectal polyps, overcoming the limitations of other imaging modes. In this review we retrace the path of precision endoscopy, analyzing the yield of various endoscopic imaging techniques in personalizing management of colorectal polyps and early colorectal cancer.

PolyDeep is a computer-aided detection and classification (CADe/x) system trained to detect and classify polyps. During colonoscopy, CADe/x systems help endoscopists to predict the histology of colonic lesions.

To compare the diagnostic performance of PolyDeep and expert endoscopists for the optical diagnosis of colorectal polyps on still images.

PolyDeep Image Classification (PIC) is an in vitro diagnostic test study. The PIC database contains NBI images of 491 colorectal polyps with histological diagnosis. We evaluated the diagnostic performance of PolyDeep and four expert endoscopists for neoplasia (adenoma, sessile serrated lesion, traditional serrated adenoma) and adenoma characterization and compared them with the McNemar test. Receiver operating characteristic curves were constructed to assess the overall discriminatory ability, comparing the area under the curve of endoscopists and PolyDeep with the chi- square homogeneity areas test.

The diagnostic performance of the endoscopists and PolyDeep in the characterization of neoplasia is similar in terms of sensitivity (PolyDeep: 89.05%; E1: 91.23%, p=0.5; E2: 96.11%, p<0.001; E3: 86.65%, p=0.3; E4: 91.26% p=0.3) and specificity (PolyDeep: 35.53%; E1: 33.80%, p=0.8; E2: 34.72%, p=1; E3: 39.24%, p=0.8; E4: 46.84%, p=0.2). The overall discriminative ability also showed no statistically significant differences (PolyDeep: 0.623; E1: 0.625, p=0.8; E2: 0.654, p=0.2; E3: 0.629, p=0.9; E4: 0.690, p=0.09). In the optical diagnosis of adenomatous polyps, we found that PolyDeep had a significantly higher sensitivity and a significantly lower specificity. The overall discriminative ability of adenomatous lesions by expert endoscopists is significantly higher than PolyDeep (PolyDeep: 0.582; E1: 0.685, p < 0.001; E2: 0.677, p < 0.0001; E3: 0.658, p < 0.01; E4: 0.694, p < 0.0001).

PolyDeep and endoscopists have similar diagnostic performance in the optical diagnosis of neoplastic lesions. However, endoscopists have a better global discriminatory ability than PolyDeep in the optical diagnosis of adenomatous polyps.

Computer-aided detection and diagnosis systems (CADe/CADx) in endoscopy are commonly trained using high-quality imagery, which is not representative for the heterogeneous input typically encountered in clinical practice. In endoscopy, the image quality heavily relies on both the skills and experience of the endoscopist and the specifications of the system used for screening. Factors such as poor illumination, motion blur, and specific post-processing settings can significantly alter the quality and general appearance of these images. This so-called domain gap between the data used for developing the system and the data it encounters after deployment, and the impact it has on the performance of deep neural networks (DNNs) supportive endoscopic CAD systems remains largely unexplored. As many of such systems, for e.g. polyp detection, are already being rolled out in clinical practice, this poses severe patient risks in particularly community hospitals, where both the imaging equipment and experience are subject to considerable variation. Therefore, this study aims to evaluate the impact of this domain gap on the clinical performance of CADe/CADx for various endoscopic applications. For this, we leverage two publicly available data sets (KVASIR-SEG and GIANA) and two in-house data sets. We investigate the performance of commonly-used DNN architectures under synthetic, clinically calibrated image degradations and on a prospectively collected dataset including 342 endoscopic images of lower subjective quality. Additionally, we assess the influence of DNN architecture and complexity, data augmentation, and pretraining techniques for improved robustness. The results reveal a considerable decline in performance of 11.6% (±1.5) as compared to the reference, within the clinically calibrated boundaries of image degradations. Nevertheless, employing more advanced DNN architectures and self-supervised in-domain pre-training effectively mitigate this drop to 7.7% (±2.03). Additionally, these enhancements yield the highest performance on the manually collected test set including images with lower subjective quality. By comprehensively assessing the robustness of popular DNN architectures and training strategies across multiple datasets, this study provides valuable insights into their performance and limitations for endoscopic applications. The findings highlight the importance of including robustness evaluation when developing DNNs for endoscopy applications and propose strategies to mitigate performance loss.

The study introduces various methods of performing conventional ML and their implementation in surgical areas, and the need to move beyond these traditional approaches given the advent of big data.

Investigate current understanding and future directions of machine learning applications, such as risk stratification, clinical data analytics, and decision support, in surgical practice.

The advent of the electronic health record, near unlimited computing, and open-source computational packages have created an environment for applying artificial intelligence, machine learning, and predictive analytic techniques to healthcare. The "hype" phase has passed, and algorithmic approaches are being developed for surgery patients through all stages of care, involving preoperative, intraoperative, and postoperative components. Surgeons must understand and critically evaluate the strengths and weaknesses of these methodologies.

The current body of AI literature was reviewed, emphasizing on contemporary approaches important in the surgical realm.

The unrealized impacts of AI on clinical surgery and its subspecialties are immense. As this technology continues to pervade surgical literature and clinical applications, knowledge of its inner workings and shortcomings is paramount in determining its appropriate implementation.

Artificial intelligence (AI) may be practical for image classification of small bowel capsule endoscopy (CE). However, creating a functional AI model is challenging. We attempted to create a dataset and an object detection CE AI model to explore modeling problems to assist in reading small bowel CE.

We extracted 18,481 images from 523 small bowel CE procedures performed at Kyushu University Hospital from September 2014 to June 2021. We annotated 12,320 images with 23,033 disease lesions, combined them with 6161 normal images as the dataset, and examined the characteristics. Based on the dataset, we created an object detection AI model using YOLO v5 and we tested validation.

We annotated the dataset with 12 types of annotations, and multiple annotation types were observed in the same image. We test validated our AI model with 1396 images, and sensitivity for all 12 types of annotations was about 91%, with 1375 true positives, 659 false positives, and 120 false negatives detected. The highest sensitivity for individual annotations was 97%, and the highest area under the receiver operating characteristic curve was 0.98, but the quality of detection varied depending on the specific annotation.

Object detection AI model in small bowel CE using YOLO v5 may provide effective and easy-to-understand reading assistance. In this SEE-AI project, we open our dataset, the weights of the AI model, and a demonstration to experience our AI. We look forward to further improving the AI model in the future.

Artificial intelligence represents an emerging area with promising potential for improving colonoscopy quality.

To develop a colon polyp detection model using STFT and evaluate its performance through a randomized sample experiment.

Colonoscopy videos from the Digestive Endoscopy Center of the First Affiliated Hospital of Anhui Medical University, recorded between January 2018 and November 2022, were selected and divided into two datasets. To verify the model's practical application in clinical settings, 1500 colonoscopy images and 1200 polyp images of various sizes were randomly selected from the test set and compared with the STFT model's and endoscopists' recognition results with different years of experience.

In the randomized sample trial involving 1500 colonoscopy images, the STFT model demonstrated significantly higher accuracy and specificity compared to endoscopists with low years of experience (0.902 vs. 0.809, 0.898 vs. 0.826, respectively). Moreover, the model's sensitivity was 0.904, which was higher than that of endoscopists with low, medium, or high years of experience (0.80, 0.896, 0.895, respectively), with statistical significance (P < 0.05). In the randomized sample experiment of 1200 polyp images of different sizes, the accuracy of the STFT model was significantly higher than that of endoscopists with low years of experience when the polyp size was ≤ 0.5 cm and 0.6-1.0 cm (0.902 vs. 0.70, 0.953 vs. 0.865, respectively).

The STFT-based colon polyp detection model exhibits high accuracy in detecting polyps in colonoscopy videos, with a particular efficiency in detecting small polyps (≤ 0.5 cm)(0.902 vs. 0.70, P < 0.001).

Computer-aided characterization (CADx) may be used to implement optical biopsy strategies into colonoscopy practice; however, its impact on endoscopic diagnosis remains unknown. We aimed to evaluate the additional diagnostic value of CADx when used by endoscopists for assessing colorectal polyps.

This was a single-center, multicase, multireader, image-reading study using randomly extracted images of pathologically confirmed polyps resected between July 2021 and January 2022. Approved CADx that could predict two-tier classification (neoplastic or nonneoplastic) by analyzing narrow-band images of the polyps was used to obtain a CADx diagnosis. Participating endoscopists determined if the polyps were neoplastic or not and noted their confidence level using a computer-based, image-reading test. The test was conducted twice with a 4-week interval: the first test was conducted without CADx prediction and the second test with CADx prediction. Diagnostic performances for neoplasms were calculated using the pathological diagnosis as reference and performances with and without CADx prediction were compared.

Five hundred polyps were randomly extracted from 385 patients and diagnosed by 14 endoscopists (including seven experts). The sensitivity for neoplasia was significantly improved by referring to CADx (89.4% vs. 95.6%). CADx also had incremental effects on the negative predictive value (69.3% vs. 84.3%), overall accuracy (87.2% vs. 91.8%), and high-confidence diagnosis rate (77.4% vs. 85.8%). However, there was no significant difference in specificity (80.1% vs. 78.9%).

Computer-aided characterization has added diagnostic value for differentiating colorectal neoplasms and may improve the high-confidence diagnosis rate.

Patient care workflows are highly multimodal and intertwined: the intersection of data outputs provided from different disciplines and in different formats remains one of the main challenges of modern oncology. Artificial Intelligence (AI) has the potential to revolutionize the current clinical practice of oncology owing to advancements in digitalization, database expansion, computational technologies, and algorithmic innovations that facilitate discernment of complex relationships in multimodal data. Within oncology, radiation therapy (RT) represents an increasingly complex working procedure, involving many labor-intensive and operator-dependent tasks. In this context, AI has gained momentum as a powerful tool to standardize treatment performance and reduce inter-observer variability in a time-efficient manner. This review explores the hurdles associated with the development, implementation, and maintenance of AI platforms and highlights current measures in place to address them. In examining AI's role in oncology workflows, we underscore that a thorough and critical consideration of these challenges is the only way to ensure equitable and unbiased care delivery, ultimately serving patients' survival and quality of life.

Recognizing anatomical sections during colonoscopy is crucial for diagnosing colonic diseases and generating accurate reports. While recent studies have endeavored to identify anatomical regions of the colon using deep learning, the deformable anatomical characteristics of the colon pose challenges for establishing a reliable localization system. This study presents a system utilizing 100 colonoscopy videos, combining density clustering and deep learning. Cascaded CNN models are employed to estimate the appendix orifice (AO), flexures, and "outside of the body," sequentially. Subsequently, DBSCAN algorithm is applied to identify anatomical sections. Clustering-based analysis integrates clinical knowledge and context based on the anatomical section within the model. We address challenges posed by colonoscopy images through non-informative removal preprocessing. The image data is labeled by clinicians, and the system deduces section correspondence stochastically. The model categorizes the colon into three sections: right (cecum and ascending colon), middle (transverse colon), and left (descending colon, sigmoid colon, rectum). We estimated the appearance time of anatomical boundaries with an average error of 6.31 s for AO, 9.79 s for HF, 27.69 s for SF, and 3.26 s for outside of the body. The proposed method can facilitate future advancements towards AI-based automatic reporting, offering time-saving efficacy and standardization.

Background: Aldosterone-producing adenomas (APA) are a common cause of primary aldosteronism (PA), a clinical syndrome characterized by hypertension and electrolyte disturbances. If untreated, it may lead to serious cardiovascular complications. Therefore, there is an urgent need for potential biomarkers and targeted drugs for the diagnosis and treatment of aldosteronism. Methods: We downloaded two datasets (GSE156931 and GSE60042) from the GEO database and merged them by de-batch effect, then screened the top50 of differential genes using PPI and enriched them, followed by screening the Aldosterone adenoma-related genes (ARGs) in the top50 using three machine learning algorithms. We performed GSEA analysis on the ARGs separately and constructed artificial neural networks based on the ARGs. Finally, the Enrich platform was utilized to identify drugs with potential therapeutic effects on APA by tARGseting the ARGs. Results: We identified 190 differential genes by differential analysis, and then identified the top50 genes by PPI, and the enrichment analysis showed that they were mainly enriched in amino acid metabolic pathways. Then three machine learning algorithms identified five ARGs, namely, SST, RAB3C, PPY, CYP3A4, CDH10, and the ANN constructed on the basis of these five ARGs had better diagnostic effect on APA, in which the AUC of the training set is 1 and the AUC of the validation set is 0.755. And then the Enrich platform identified drugs tARGseting the ARGs with potential therapeutic effects on APA. Conclusion: We identified five ARGs for APA through bioinformatic analysis and constructed Artificial neural network (ANN) based on them with better diagnostic effects, and identified drugs with potential therapeutic effects on APA by tARGseting these ARGs. Our study provides more options for the diagnosis and treatment of APA.

With an increasing number of patients with gastrointestinal cancer, effective and accurate early diagnostic clinical tools are required provide better health care for patients with gastrointestinal cancer. Recent studies have shown that artificial intelligence (AI) plays an important role in the diagnosis and treatment of patients with gastrointestinal tumors, which not only improves the efficiency of early tumor screening, but also significantly improves the survival rate of patients after treatment. With the aid of efficient learning and judgment abilities of AI, endoscopists can improve the accuracy of diagnosis and treatment through endoscopy and avoid incorrect descriptions or judgments of gastrointestinal lesions. The present article provides an overview of the application status of various artificial intelligence in gastric and colorectal cancers in recent years, and the direction of future research and clinical practice is clarified from a clinical perspective to provide a comprehensive theoretical basis for AI as a promising diagnostic and therapeutic tool for gastrointestinal cancer.

Colonoscopy is crucial in preventing colorectal cancer (CRC) and reducing associated mortality. This comprehensive review examines the importance of high-quality colonoscopy and associated quality indicators, including bowel preparation, cecal intubation rate, withdrawal time, adenoma detection rate (ADR), complete resection, specimen retrieval, complication rates, and patient satisfaction, while also discussing other ADR-related metrics. Additionally, the review draws attention to often overlooked quality aspects, such as nonpolypoid lesion detection, as well as insertion and withdrawal skills. Moreover, it explores the potential of artificial intelligence in enhancing colonoscopy quality and highlights specific considerations for organized screening programs. The review also emphasizes the implications of organized screening programs and the need for continuous quality improvement. A high-quality colonoscopy is crucial for preventing postcolonoscopy CRC- and CRC-related deaths. Health-care professionals must develop a thorough understanding of colonoscopy quality components, including technical quality, patient safety, and patient experience. By prioritizing ongoing evaluation and refinement of these quality indicators, health-care providers can contribute to improved patient outcomes and develop more effective CRC screening programs.

Colorectal cancer (CRC) has become a severe global health concern, with the third‑high incidence and second‑high mortality rate of all cancers. The burden of CRC is expected to surge to 60% by 2030. Fortunately, effective early evidence‑based screening could significantly reduce the incidence and mortality of CRC. Colonoscopy is the core screening method for CRC with high popularity and accuracy. Yet, the accuracy of colonoscopy in CRC screening is related to the experience and state of operating physicians. It is challenging to maintain the high CRC diagnostic rate of colonoscopy. Artificial intelligence (AI)‑assisted colonoscopy will compensate for the above shortcomings and improve the accuracy, efficiency, and quality of colonoscopy screening. The unique advantages of AI, such as the continuous advancement of high‑performance computing capabilities and innovative deep‑learning architectures, which hugely impact the control of colorectal cancer morbidity and mortality expectancy, highlight its role in colonoscopy screening.

Artificial intelligence (AI) is a subfield of computer science that aims to implement computer systems that perform tasks that generally require human learning, reasoning, and perceptual abilities. AI is widely used in the medical field. The interpretation of medical images requires considerable effort, time, and skill. AI-aided interpretations, such as automated abnormal lesion detection and image classification, are promising areas of AI. However, when images with different characteristics are extracted, depending on the manufacturer and imaging environment, a so-called domain shift problem occurs in which the developed AI has a poor versatility. Domain adaptation is used to address this problem. Domain adaptation is a tool that generates a newly converted image which is suitable for other domains. It has also shown promise in reducing the differences in appearance among the images collected from different devices. Domain adaptation is expected to improve the reading accuracy of AI for heterogeneous image distributions in gastrointestinal (GI) endoscopy and medical image analyses. In this paper, we review the history and basic characteristics of domain shift and domain adaptation. We also address their use in gastrointestinal endoscopy and the medical field more generally through published examples, perspectives, and future directions.

Since the emergence of artificial intelligence (AI) in medicine, endoscopy applications in gastroenterology have been at the forefront of innovations. The ever-increasing number of studies necessitates the need to organize and classify applications in a useful way. Separating AI capabilities by computer aided detection (CADe), diagnosis (CADx), and quality assessment (CADq) allows for a systematic evaluation of each application. CADe studies have shown promise in accurate detection of esophageal, gastric and colonic neoplasia as well as identifying sources of bleeding and Crohn's disease in the small bowel. While more advanced CADx applications employ optical biopsies to give further information to characterize neoplasia and grade inflammatory disease, diverse CADq applications ensure quality and increase the efficiency of procedures. Future applications show promise in advanced therapeutic modalities and integrated systems that provide multimodal capabilities. AI is set to revolutionize clinical decision making and performance of endoscopy.

BACKGROUND : We aimed to compare the accuracy of the optical diagnosis of diminutive colorectal polyps, including sessile serrated lesions (SSLs), between a computer-aided diagnosis (CADx) system and endoscopists during real-time colonoscopy. METHODS : We developed the POLyp Artificial Recognition (POLAR) system, which was capable of performing real-time characterization of diminutive colorectal polyps. For pretraining, the Microsoft-COCO dataset with over 300 000 nonpolyp object images was used. For training, eight hospitals prospectively collected 2637 annotated images from 1339 polyps (i. e. publicly available online POLAR database). For clinical validation, POLAR was tested during colonoscopy in patients with a positive fecal immunochemical test (FIT), and compared with the performance of 20 endoscopists from eight hospitals. Endoscopists were blinded to the POLAR output. Primary outcome was the comparison of accuracy of the optical diagnosis of diminutive colorectal polyps between POLAR and endoscopists (neoplastic [adenomas and SSLs] versus non-neoplastic [hyperplastic polyps]). Histopathology served as the reference standard. RESULTS : During clinical validation, 423 diminutive polyps detected in 194 FIT-positive individuals were included for analysis (300 adenomas, 41 SSLs, 82 hyperplastic polyps). POLAR distinguished neoplastic from non-neoplastic lesions with 79 % accuracy, 89 % sensitivity, and 38 % specificity. The endoscopists achieved 83 % accuracy, 92 % sensitivity, and 44 % specificity. The optical diagnosis accuracy between POLAR and endoscopists was not significantly different (P = 0.10). The proportion of polyps in which POLAR was able to provide an optical diagnosis was 98 % (i. e. success rate). CONCLUSIONS : We developed a CADx system that differentiated neoplastic from non-neoplastic diminutive polyps during endoscopy, with an accuracy comparable to that of screening endoscopists and near-perfect success rate.

Endoscopic assessment of ulcerative colitis (UC) can be performed by using the Mayo Endoscopic Score (MES) or the Ulcerative Colitis Endoscopic Index of Severity (UCEIS). In this meta-analysis, we assessed the pooled diagnostic accuracy parameters of deep machine learning by means of convolutional neural network (CNN) algorithms in predicting UC severity on endoscopic images.

Databases including MEDLINE, Scopus, and Embase were searched in June 2022. Outcomes of interest were the pooled accuracy, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV). Standard meta-analysis methods used the random-effects model, and heterogeneity was assessed using the I2statistics.

Twelve studies were included in the final analysis. The pooled diagnostic parameters of CNN-based machine learning algorithms in endoscopic severity assessment of UC were as follows: accuracy 91.5% (95% confidence interval [CI], 88.3-93.8; I2 = 84%), sensitivity 82.8% (95% CI, 78.3-86.5; I2 = 89%), specificity 92.4% (95% CI, 89.4-94.6; I2 = 84%), PPV 86.6% (95% CI, 82.3-90; I2 = 89%), and NPV 88.6% (95% CI, 85.7-91; I2 = 78%). Subgroup analysis revealed significantly better sensitivity and PPV with the UCEIS scoring system compared with the MES (93.6% [95% CI, 87.5-96.8; I2 = 77%] vs 82% [95% CI, 75.6-87; I2 = 89%], P = .003, and 93.6% [95% CI, 88.7-96.4; I2 = 68%] vs 83.6% [95% CI, 76.8-88.8; I2 = 77%], P = .007, respectively).

CNN-based machine learning algorithms demonstrated excellent pooled diagnostic accuracy parameters in the endoscopic severity assessment of UC. Using UCEIS scores in CNN training might offer better results than the MES. Further studies are warranted to establish these findings in real clinical settings.

Differentiation of colorectal cancers (CRCs) with deep submucosal invasion (T1b) from CRCs with superficial invasion (T1a) or no invasion (Tis) is not straightforward. This study aimed to develop a computer-aided diagnosis (CADx) system to establish the diagnosis of early-stage cancers using nonmagnified endoscopic white-light images alone.

From 5108 images, 1513 lesions (Tis, 1074; T1a, 145; T1b, 294) were collected from 1470 patients at 10 academic hospitals and assigned to training and testing datasets (3:1). The ResNet-50 network was used as the backbone to extract features from images. Oversampling and focal loss were used to compensate class imbalance of the invasive stage. Diagnostic performance was assessed using the testing dataset including 403 CRCs with 1392 images. Two experts and 2 trainees read the identical testing dataset.

At a 90% cutoff for the per-lesion score, CADx showed the highest specificity of 94.4% (95% confidence interval [CI], 91.3-96.6), with 59.8% (95% CI, 48.3-70.4) sensitivity and 87.3% (95% CI, 83.7-90.4) accuracy. The area under the characteristic curve was 85.1% (95% CI, 79.9-90.4) for CADx, 88.2% (95% CI, 83.7-92.8) for expert 1, 85.9% (95% CI, 80.9-90.9) for expert 2, 77.0% (95% CI, 71.5-82.4) for trainee 1 (vs CADx; P = .0076), and 66.2% (95% CI, 60.6-71.9) for trainee 2 (P < .0001). The function was also confirmed on 9 short videos.

A CADx system developed with endoscopic white-light images showed excellent per-lesion specificity and accuracy for T1b lesion diagnosis, equivalent to experts and superior to trainees. (Clinical trial registration number: UMIN000037053.).

Artificial intelligence (AI) in gastrointestinal endoscopy includes systems designed to interpret medical images and increase sensitivity during examination. This may be a promising solution to human biases and may provide support during diagnostic endoscopy.

This review aims to summarize and evaluate data supporting AI technologies in lower endoscopy, addressing their effectiveness, limitations, and future perspectives.

Computer-aided detection (CADe) systems have been studied with promising results, allowing for an increase in adenoma detection rate (ADR), adenoma per colonoscopy (APC), and a reduction in adenoma miss rate (AMR). This may lead to an increase in the sensitivity of endoscopic examinations and a reduction in the risk of interval-colorectal cancer. In addition, computer-aided characterization (CADx) has also been implemented, aiming to distinguish adenomatous and non-adenomatous lesions through real-time assessment using advanced endoscopic imaging techniques. Moreover, computer-aided quality (CADq) systems have been developed with the aim of standardizing quality measures in colonoscopy (e.g. withdrawal time and adequacy of bowel cleansing) both to improve the quality of examinations and set a reference standard for randomized controlled trials.

Image-enhanced endoscopy (IEE) has been used in the differentiation between neoplastic and non-neoplastic colorectal lesions through microvasculature analysis. This study aimed to evaluate the computer-aided diagnosis (CADx) mode of the CAD EYE system for the optical diagnosis of colorectal lesions and compare it with the performance of an expert, in addition to evaluating the computer-aided detection (CADe) mode in terms of polyp detection rate (PDR) and adenoma detection rate (ADR).

A prospective study was conducted to evaluate the performance of CAD EYE using blue light imaging (BLI), dichotomizing lesions into hyperplastic and neoplastic, and of an expert based on the Japan Narrow-Band Imaging Expert Team (JNET) classification for the characterization of lesions. After white light imaging (WLI) diagnosis, magnification was used on all lesions, which were removed and examined histologically. Diagnostic criteria were evaluated, and PDR and ADR were calculated.

A total of 110 lesions (80 (72.7%) dysplastic lesions and 30 (27.3%) nondysplastic lesions) were evaluated in 52 patients, with a mean lesion size of 4.3 mm. Artificial intelligence (AI) analysis showed 81.8% accuracy, 76.3% sensitivity, 96.7% specificity, 98.5% positive predictive value (PPV), and 60.4% negative predictive value (NPV). The kappa value was 0.61, and the area under the receiver operating characteristic curve (AUC) was 0.87. Expert analysis showed 93.6% accuracy, 92.5% sensitivity, 96.7% specificity, 98.7% PPV, and 82.9% NPV. The kappa value was 0.85, and the AUC was 0.95. Overall, PDR was 67.6% and ADR was 45.9%.

The CADx mode showed good accuracy in characterizing colorectal lesions, but the expert assessment was superior in almost all diagnostic criteria. PDR and ADR were high.