TOXRIC: a comprehensive database of toxicological data and benchmarks.

The toxic effects of compounds on environment, humans, and other organisms have been a major focus of many research areas, including drug discovery and ecological research. Identifying the potential toxicity in the early stage of compound/drug discovery is critical. The rapid development of computational methods for evaluating various toxicity categories has increased the need for comprehensive and system-level collection of toxicological data, associated attributes, and benchmarks. To contribute toward this goal, we proposed TOXRIC (https://toxric.bioinforai.tech/), a database with comprehensive toxicological data, standardized attribute data, practical benchmarks, informative visualization of molecular representations, and an intuitive function interface. The data stored in TOXRIC contains 113 372 compounds, 13 toxicity categories, 1474 toxicity endpoints covering in vivo/in vitro endpoints and 39 feature types, covering structural, target, transcriptome, metabolic data, and other descriptors. All the curated datasets of endpoints and features can be retrieved, downloaded and directly used as output or input to Machine Learning (ML)-based prediction models. In addition to serving as a data repository, TOXRIC also provides visualization of benchmarks and molecular representations for all endpoint datasets. Based on these results, researchers can better understand and select optimal feature types, molecular representations, and baseline algorithms for each endpoint prediction task. We believe that the rich information on compound toxicology, ML-ready datasets, benchmarks and molecular representation distribution can greatly facilitate toxicological investigations, interpretation of toxicological mechanisms, compound/drug discovery and the development of computational methods.

An enhanced cascade-based deep forest model for drug combination prediction.

Combination therapy has shown an obvious curative effect on complex diseases, whereas the search space of drug combinations is too large to be validated experimentally even with high-throughput screens. With the increase of the number of drugs, artificial intelligence techniques, especially machine learning methods, have become applicable for the discovery of synergistic drug combinations to significantly reduce the experimental workload. In this study, in order to predict novel synergistic drug combinations in various cancer cell lines, the cell line-specific drug-induced gene expression profile (GP) is added as a new feature type to capture the cellular response of drugs and reveal the biological mechanism of synergistic effect. Then, an enhanced cascade-based deep forest regressor (EC-DFR) is innovatively presented to apply the new small-scale drug combination dataset involving chemical, physical and biological (GP) properties of drugs and cells. Verified by the dataset, EC-DFR outperforms two state-of-the-art deep neural network-based methods and several advanced classical machine learning algorithms. Biological experimental validation performed subsequently on a set of previously untested drug combinations further confirms the performance of EC-DFR. What is more prominent is that EC-DFR can distinguish the most important features, making it more interpretable. By evaluating the contribution of each feature type, GP feature contributes 82.40%, showing the cellular responses of drugs may play crucial roles in synergism prediction. The analysis based on the top contributing genes in GP further demonstrates some potential relationships between the transcriptomic levels of key genes under drug regulation and the synergism of drug combinations.

Machine learning methods, databases and tools for drug combination prediction.

Combination therapy has shown an obvious efficacy on complex diseases and can greatly reduce the development of drug resistance. However, even with high-throughput screens, experimental methods are insufficient to explore novel drug combinations. In order to reduce the search space of drug combinations, there is an urgent need to develop more efficient computational methods to predict novel drug combinations. In recent decades, more and more machine learning (ML) algorithms have been applied to improve the predictive performance. The object of this study is to introduce and discuss the recent applications of ML methods and the widely used databases in drug combination prediction. In this study, we first describe the concept and controversy of synergism between drug combinations. Then, we investigate various publicly available data resources and tools for prediction tasks. Next, ML methods including classic ML and deep learning methods applied in drug combination prediction are introduced. Finally, we summarize the challenges to ML methods in prediction tasks and provide a discussion on future work.

Systematic optimization of host-directed therapeutic targets and preclinical validation of repositioned antiviral drugs.

Inhibition of host protein functions using established drugs produces a promising antiviral effect with excellent safety profiles, decreased incidence of resistant variants and favorable balance of costs and risks. Genomic methods have produced a large number of robust host factors, providing candidates for identification of antiviral drug targets. However, there is a lack of global perspectives and systematic prioritization of known virus-targeted host proteins (VTHPs) and drug targets. There is also a need for host-directed repositioned antivirals. Here, we integrated 6140 VTHPs and grouped viral infection modes from a new perspective of enriched pathways of VTHPs. Clarifying the superiority of nonessential membrane and hub VTHPs as potential ideal targets for repositioned antivirals, we proposed 543 candidate VTHPs. We then presented a large-scale drug-virus network (DVN) based on matching these VTHPs and drug targets. We predicted possible indications for 703 approved drugs against 35 viruses and explored their potential as broad-spectrum antivirals. In vitro and in vivo tests validated the efficacy of bosutinib, maraviroc and dextromethorphan against human herpesvirus 1 (HHV-1), hepatitis B virus (HBV) and influenza A virus (IAV). Their drug synergy with clinically used antivirals was evaluated and confirmed. The results proved that low-dose dextromethorphan is better than high-dose in both single and combined treatments. This study provides a comprehensive landscape and optimization strategy for druggable VTHPs, constructing an innovative and potent pipeline to discover novel antiviral host proteins and repositioned drugs, which may facilitate their delivery to clinical application in translational medicine to combat fatal and spreading viral infections.

Computational methods, databases and tools for synthetic lethality prediction.

Synthetic lethality (SL) occurs between two genes when the inactivation of either gene alone has no effect on cell survival but the inactivation of both genes results in cell death. SL-based therapy has become one of the most promising targeted cancer therapies in the last decade as PARP inhibitors achieve great success in the clinic. The key point to exploiting SL-based cancer therapy is the identification of robust SL pairs. Although many wet-lab-based methods have been developed to screen SL pairs, known SL pairs are less than 0.1% of all potential pairs due to large number of human gene combinations. Computational prediction methods complement wet-lab-based methods to effectively reduce the search space of SL pairs. In this paper, we review the recent applications of computational methods and commonly used databases for SL prediction. First, we introduce the concept of SL and its screening methods. Second, various SL-related data resources are summarized. Then, computational methods including statistical-based methods, network-based methods, classical machine learning methods and deep learning methods for SL prediction are summarized. In particular, we elaborate on the negative sampling methods applied in these models. Next, representative tools for SL prediction are introduced. Finally, the challenges and future work for SL prediction are discussed.

Multistep validation of a post-ERCP pancreatitis prediction system integrating multimodal data: a multicenter study.

BACKGROUND AND AIMS: The impact of various categories of information on the prediction of post-ERCP pancreatitis (PEP) remains uncertain. We comprehensively investigated the risk factors associated with PEP by constructing and validating a model incorporating multimodal data through multiple steps. METHODS: Cases (n = 1916) of ERCP were retrospectively collected from multiple centers for model construction. Through literature research, 49 electronic health record (EHR) features and 1 image feature related to PEP were identified. The EHR features were categorized into baseline, diagnosis, technique, and prevention strategies, covering pre-ERCP, intra-ERCP, and peri-ERCP phases. We first incrementally constructed models 1 to 4 incorporating these 4 feature categories and then added the image feature into models 1 to 4 and developed models 5 to 8. All models underwent testing and comparison using both internal and external test sets. Once the optimal model was selected, we conducted comparisons among multiple machine learning algorithms. RESULTS: Compared with model 2 that incorporated baseline and diagnosis features, adding technique and prevention strategies (model 4) greatly improved the sensitivity (63.89% vs 83.33%, P < .05) and specificity (75.00% vs 85.92%, P < .001). A similar tendency was observed in the internal and external tests. In model 4, the top 3 features ranked by weight were previous pancreatitis, nonsteroidal anti-inflammatory drug use, and difficult cannulation. The image-based feature has the highest weight in models 5 to 8. Finally, model 8 used a random forest algorithm and showed the best performance. CONCLUSIONS: We first developed a multimodal prediction model for identifying PEP with a clinical-acceptable performance. The image and technique features are crucial for PEP prediction.

A real-time deep learning-based system for colorectal polyp size estimation by white-light endoscopy: development and multicenter prospective validation.

BACKGROUND: The choice of polypectomy device and surveillance intervals for colorectal polyps are primarily decided by polyp size. We developed a deep learning-based system (ENDOANGEL-CPS) to estimate colorectal polyp size in real time. METHODS: ENDOANGEL-CPS calculates polyp size by estimating the distance from the endoscope lens to the polyp using the parameters of the lens. The depth estimator network was developed on 7297 images from five virtually produced colon videos and tested on 730 images from seven virtual colon videos. The performance of the system was first evaluated in nine videos of a simulated colon with polyps attached, then tested in 157 real-world prospective videos from three hospitals, with the outcomes compared with that of nine endoscopists over 69 videos. Inappropriate surveillance recommendations caused by incorrect estimation of polyp size were also analyzed. RESULTS: The relative error of depth estimation was 11.3% (SD 6.0%) in successive virtual colon images. The concordance correlation coefficients (CCCs) between system estimation and ground truth were 0.89 and 0.93 in images of a simulated colon and multicenter videos of 157 polyps. The mean CCC of ENDOANGEL-CPS surpassed all endoscopists (0.89 vs. 0.41 [SD 0.29]; P<0.001). The relative accuracy of ENDOANGEL-CPS was significantly higher than that of endoscopists (89.9% vs. 54.7%; P<0.001). Regarding inappropriate surveillance recommendations, the system's error rate is also lower than that of endoscopists (1.5% vs. 16.6%; P<0.001). CONCLUSIONS: ENDOANGEL-CPS could potentially improve the accuracy of colorectal polyp size measurements and size-based surveillance intervals.

A newly developed deep learning-based system for automatic detection and classification of small bowel lesions during double-balloon enteroscopy examination.

BACKGROUND: Double-balloon enteroscopy (DBE) is a standard method for diagnosing and treating small bowel disease. However, DBE may yield false-negative results due to oversight or inexperience. We aim to develop a computer-aided diagnostic (CAD) system for the automatic detection and classification of small bowel abnormalities in DBE. DESIGN AND METHODS: A total of 5201 images were collected from Renmin Hospital of Wuhan University to construct a detection model for localizing lesions during DBE, and 3021 images were collected to construct a classification model for classifying lesions into four classes, protruding lesion, diverticulum, erosion & ulcer and angioectasia. The performance of the two models was evaluated using 1318 normal images and 915 abnormal images and 65 videos from independent patients and then compared with that of 8 endoscopists. The standard answer was the expert consensus. RESULTS: For the image test set, the detection model achieved a sensitivity of 92% (843/915) and an area under the curve (AUC) of 0.947, and the classification model achieved an accuracy of 86%. For the video test set, the accuracy of the system was significantly better than that of the endoscopists (85% vs. 77 ± 6%, p < 0.01). For the video test set, the proposed system was superior to novices and comparable to experts. CONCLUSIONS: We established a real-time CAD system for detecting and classifying small bowel lesions in DBE with favourable performance. ENDOANGEL-DBE has the potential to help endoscopists, especially novices, in clinical practice and may reduce the miss rate of small bowel lesions.

Gastric neoplasm detection of computer-aided detection-assisted esophagogastroduodenoscopy changes with implement scenarios: a real-world study.

BACKGROUND AND AIM: The implementation of computer-aided detection (CAD) devices in esophagogastroduodenoscopy (EGD) could autonomously identify gastric precancerous lesions and neoplasms and reduce the miss rate of gastric neoplasms in prospective trials. However, there is still insufficient evidence of their use in real-life clinical practice. METHODS: A real-world, two-center study was conducted at Wenzhou Central Hospital (WCH) and Renmin Hospital of Wuhan University (RHWU). High biopsy rate and low biopsy rate strategies were adopted, and CAD devices were applied in 2019 and 2021 at WCH and RHWU, respectively. We compared differences in gastric precancerous and neoplasm detection of EGD before and after the use of CAD devices in the first half of the year. RESULTS: A total of 33 885 patients were included and 32 886 patients were ultimately analyzed. In WCH of which biopsy rate >95%, with the implementation of CAD, more the number of early gastric cancer divided by all gastric neoplasm (EGC/GN) (0.35% vs 0.59%, P = 0.028, OR [95% CI] = 1.65 [1.0-2.60]) was found, while gastric neoplasm detection rate (1.39% vs 1.36%, P = 0.897, OR [95% CI] = 0.98 [0.76-1.26]) remained stable. In RHWU of which biopsy rate <20%, the gastric neoplasm detection rate (1.78% vs 3.23%, P < 0.001, OR [95% CI] = 1.84 [1.33-2.54]) nearly doubled after the implementation of CAD, while there was no significant change in the EGC/GN. CONCLUSION: The application of CAD devices devoted to distinct increases in gastric neoplasm detection according to different biopsy strategies, which implied that CAD devices demonstrated assistance on gastric neoplasm detection while varied effectiveness according to different implementation scenarios.

The effect of incorporating domain knowledge with deep learning in identifying benign and malignant gastric whitish lesions: A retrospective study.

BACKGROUND AND AIM: Early whitish gastric neoplasms can be easily misdiagnosed; differential diagnosis of gastric whitish lesions remains a challenge. We aim to build a deep learning (DL) model to diagnose whitish gastric neoplasms and explore the effect of adding domain knowledge in model construction. METHODS: We collected 4558 images from two institutions to train and test models. We first developed two sole DL models (1 and 2) using supervised and semi-supervised algorithms. Then we selected diagnosis-related features through literature research and developed feature-extraction models to determine features including boundary, surface, roundness, depression, and location. Then predictions of the five feature-extraction models and sole DL model were combined and inputted into seven machine-learning (ML) based fitting-diagnosis models. The optimal model was selected as ENDOANGEL-WD (whitish-diagnosis) and compared with endoscopists. RESULTS: Sole DL 2 had higher sensitivity (83.12% vs 68.67%, Bonferroni adjusted P = 0.024) than sole DL 1. Adding domain knowledge, the decision tree performed best among the seven ML models, achieving higher specificity than DL 1 (84.38% vs 72.27%, Bonferroni adjusted P < 0.05) and higher accuracy than DL 2 (80.47%, Bonferroni adjusted P < 0.001) and was selected as ENDOANGEL-WD. ENDOANGEL-WD showed better accuracy compared with 10 endoscopists (75.70%, P < 0.001). CONCLUSIONS: We developed a novel system ENDOANGEL-WD combining domain knowledge and traditional DL to detect gastric whitish neoplasms. Adding domain knowledge improved the performance of traditional DL, which provided a novel solution for establishing diagnostic models for other rare diseases potentially.

EDST: a decision stump based ensemble algorithm for synergistic drug combination prediction.

INTRODUCTION: There are countless possibilities for drug combinations, which makes it expensive and time-consuming to rely solely on clinical trials to determine the effects of each possible drug combination. In order to screen out the most effective drug combinations more quickly, scholars began to apply machine learning to drug combination prediction. However, most of them are of low interpretability. Consequently, even though they can sometimes produce high prediction accuracy, experts in the medical and biological fields can still not fully rely on their judgments because of the lack of knowledge about the decision-making process. RELATED WORK: Decision trees and their ensemble algorithms are considered to be suitable methods for pharmaceutical applications due to their excellent performance and good interpretability. We review existing decision trees or decision tree ensemble algorithms in the medical field and point out their shortcomings. METHOD: This study proposes a decision stump (DS)-based solution to extract interpretable knowledge from data sets. In this method, a set of DSs is first generated to selectively form a decision tree (DST). Different from the traditional decision tree, our algorithm not only enables a partial exchange of information between base classifiers by introducing a stump exchange method but also uses a modified Gini index to evaluate stump performance so that the generation of each node is evaluated by a global view to maintain high generalization ability. Furthermore, these trees are combined to construct an ensemble of DST (EDST). EXPERIMENT: The two-drug combination data sets are collected from two cell lines with three classes (additive, antagonistic and synergistic effects) to test our method. Experimental results show that both our DST and EDST perform better than other methods. Besides, the rules generated by our methods are more compact and more accurate than other rule-based algorithms. Finally, we also analyze the extracted knowledge by the model in the field of bioinformatics. CONCLUSION: The novel decision tree ensemble model can effectively predict the effect of drug combination datasets and easily obtain the decision-making process.

Effect of a deep learning-based automatic upper GI endoscopic reporting system: a randomized crossover study (with video).

BACKGROUND AND AIMS: EGD is essential for GI disorders, and reports are pivotal to facilitating postprocedure diagnosis and treatment. Manual report generation lacks sufficient quality and is labor intensive. We reported and validated an artificial intelligence-based endoscopy automatic reporting system (AI-EARS). METHODS: The AI-EARS was designed for automatic report generation, including real-time image capturing, diagnosis, and textual description. It was developed using multicenter datasets from 8 hospitals in China, including 252,111 images for training, 62,706 images, and 950 videos for testing. Twelve endoscopists and 44 endoscopy procedures were consecutively enrolled to evaluate the effect of the AI-EARS in a multireader, multicase, crossover study. The precision and completeness of the reports were compared between endoscopists using the AI-EARS and conventional reporting systems. RESULTS: In video validation, the AI-EARS achieved completeness of 98.59% and 99.69% for esophageal and gastric abnormality records, respectively, accuracies of 87.99% and 88.85% for esophageal and gastric lesion location records, and 73.14% and 85.24% for diagnosis. Compared with the conventional reporting systems, the AI-EARS achieved greater completeness (79.03% vs 51.86%, P < .001) and accuracy (64.47% vs 42.81%, P < .001) of the textual description and completeness of the photo-documents of landmarks (92.23% vs 73.69%, P < .001). The mean reporting time for an individual lesion was significantly reduced (80.13 ± 16.12 seconds vs 46.47 ± 11.68 seconds, P < .001) after the AI-EARS assistance. CONCLUSIONS: The AI-EARS showed its efficacy in improving the accuracy and completeness of EGD reports. It might facilitate the generation of complete endoscopy reports and postendoscopy patient management. (Clinical trial registration number: NCT05479253.).

An artificial intelligence difficulty scoring system for stone removal during ERCP: a prospective validation.

BACKGROUND: A computer-assisted (CAD) system was developed to assess, score, and classify the technical difficulty of common bile duct (CBD) stone removal during endoscopic retrograde cholangiopancreatography (ERCP). The efficacy of the CAD system was subsequently assessed through a multicenter, prospective, observational study. METHOD: All patients who met the inclusion criteria were included. Based on cholangiogram images, the CAD system analyzed the level of difficulty of stone removal and classified it into "difficult" and "easy" groups. Subsequently, differences in clinical endpoints, including attempts at stone extraction, stone extraction time, total operation time, and stone clearance rates were compared between the two groups. RESULTS: 173 patients with CBD stones from three hospitals were included in the study. The group classified as difficult by CAD had more extraction attempts (7.20 vs. 4.20, P < 0.001), more frequent machine lithotripsy (30.4 % vs. 7.1 %, P < 0.001), longer stone extraction time (16.59 vs. 7.69 minutes, P < 0.001), lower single-session stone clearance rate (73.9 % vs. 94.5 %, P < 0.001), and lower total stone clearance rate (89.1 % vs. 97.6 %, P = 0.019) compared with the group classified as easy by CAD. CONCLUSION: The CAD system effectively assessed and classified the degree of technical difficulty in endoscopic stone extraction during ERCP. In addition, it automatically provided a quantitative evaluation of CBD and stones, which in turn could help endoscopists to apply suitable procedures and interventional methods to minimize the possible risks associated with endoscopic stone removal.

Effect of an automated flexible endoscope channel brushing system on improving reprocessing quality: a randomized controlled study.

BACKGROUND: Qualified reprocessing, of which meticulous channel brushing is the most crucial step, is essential for prevention and control of endoscopy-associated infections. However, channel brushing is often omitted in practice. This study aimed to evaluate the effect of an automated flexible endoscope channel brushing system (AECBS) on improving the quality of endoscope reprocessing. METHODS: This prospective, randomized controlled study was conducted between 24 November 2021 and 22 January 2022 at Renmin Hospital of Wuhan University, China. Eligible endoscopes were randomly allocated to the auto group (channels brushed by AECBS) or the manual group (channels brushed manually), with sampling and culturing after high-level disinfection and drying. The primary end point was the proportion of endoscopes with positive cultures. RESULTS: 204 endoscopes in the auto group and 205 in the manual group were analyzed. The proportion of endoscopes with positive cultures was significantly lower in the auto group (15.2 % [95 %CI 10.7 %-21.0 %]) than in the manual group (23.4 % [95 %CI 17.9 %-29.9 %]). CONCLUSIONS: AECBS could effectively reduce bioburden and improve reprocessing quality of gastroscopes and colonoscopes. AECBS has the potential to replace manual brushing and lower the risk of endoscopy-associated infections, providing a new option for the optimization of reprocessing.

A Genetic Algorithm-Based Ensemble Learning Framework for Drug Combination Prediction.

Combination therapy is a promising clinical treatment strategy for cancer and other complex diseases. Multiple drugs can target multiple proteins and pathways, greatly improving the therapeutic effect and slowing down drug resistance. To narrow the search space of synergistic drug combinations, many prediction models have been developed. However, drug combination datasets always have the characteristics of class imbalance. Synergistic drug combinations receive the most attention in clinical application but are in small numbers. To predict synergistic drug combinations in different cancer cell lines, in this study, we propose a genetic algorithm-based ensemble learning framework, GA-DRUG, to address the problems of class imbalance and high dimensionality of input data. The cell-line-specific gene expression profiles under drug perturbations are used to train GA-DRUG, which contains imbalanced data processing and the search of global optimal solutions. Compared to 11 state-of-the-art algorithms, GA-DRUG achieves the best performance and significantly improves the prediction performance in the minority class (Synergy). The ensemble framework can effectively correct the classification results of a single classifier. In addition, the cellular proliferation experiment performed on several previously unexplored drug combinations further confirms the predictive ability of GA-DRUG.

A deep-learning based system using multi-modal data for diagnosing gastric neoplasms in real-time (with video).

BACKGROUND: White light (WL) and weak-magnifying (WM) endoscopy are both important methods for diagnosing gastric neoplasms. This study constructed a deep-learning system named ENDOANGEL-MM (multi-modal) aimed at real-time diagnosing gastric neoplasms using WL and WM data. METHODS: WL and WM images of a same lesion were combined into image-pairs. A total of 4201 images, 7436 image-pairs, and 162 videos were used for model construction and validation. Models 1-5 including two single-modal models (WL, WM) and three multi-modal models (data fusion on task-level, feature-level, and input-level) were constructed. The models were tested on three levels including images, videos, and prospective patients. The best model was selected for constructing ENDOANGEL-MM. We compared the performance between the models and endoscopists and conducted a diagnostic study to explore the ENDOANGEL-MM's assistance ability. RESULTS: Model 4 (ENDOANGEL-MM) showed the best performance among five models. Model 2 performed better in single-modal models. The accuracy of ENDOANGEL-MM was higher than that of Model 2 in still images, real-time videos, and prospective patients. (86.54 vs 78.85%, P = 0.134; 90.00 vs 85.00%, P = 0.179; 93.55 vs 70.97%, P < 0.001). Model 2 and ENDOANGEL-MM outperformed endoscopists on WM data (85.00 vs 71.67%, P = 0.002) and multi-modal data (90.00 vs 76.17%, P = 0.002), significantly. With the assistance of ENDOANGEL-MM, the accuracy of non-experts improved significantly (85.75 vs 70.75%, P = 0.020), and performed no significant difference from experts (85.75 vs 89.00%, P = 0.159). CONCLUSIONS: The multi-modal model constructed by feature-level fusion showed the best performance. ENDOANGEL-MM identified gastric neoplasms with good accuracy and has a potential role in real-clinic.

DEML: Drug Synergy and Interaction Prediction Using Ensemble-Based Multi-Task Learning.

Synergistic drug combinations have demonstrated effective therapeutic effects in cancer treatment. Deep learning methods accelerate identification of novel drug combinations by reducing the search space. However, potential adverse drug-drug interactions (DDIs), which may increase the risks for combination therapy, cannot be detected by existing computational synergy prediction methods. We propose DEML, an ensemble-based multi-task neural network, for the simultaneous optimization of five synergy regression prediction tasks, synergy classification, and DDI classification tasks. DEML uses chemical and transcriptomics information as inputs. DEML adapts the novel hybrid ensemble layer structure to construct higher order representation using different perspectives. The task-specific fusion layer of DEML joins representations for each task using a gating mechanism. For the Loewe synergy prediction task, DEML overperforms the state-of-the-art synergy prediction method with an improvement of 7.8% and 13.2% for the root mean squared error and the R2 correlation coefficient. Owing to soft parameter sharing and ensemble learning, DEML alleviates the multi-task learning 'seesaw effect' problem and shows no performance loss on other tasks. DEML has a superior ability to predict drug pairs with high confidence and less adverse DDIs. DEML provides a promising way to guideline novel combination therapy strategies for cancer treatment.

Real-time use of artificial intelligence for diagnosing early gastric cancer by magnifying image-enhanced endoscopy: a multicenter diagnostic study (with videos).

BACKGROUND AND AIMS: Endoscopy is a pivotal method for detecting early gastric cancer (EGC). However, skill among endoscopists varies greatly. Here, we proposed a deep learning-based system named ENDOANGEL-ME to diagnose EGC in magnifying image-enhanced endoscopy (M-IEE). METHODS: M-IEE images were retrospectively obtained from 6 hospitals in China, including 4667 images for training and validation, 1324 images for internal tests, and 4702 images for external tests. One hundred eighty-seven stored videos from 2 hospitals were used to evaluate the performance of ENDOANGEL-ME and endoscopists and to assess the effect of ENDOANGEL-ME on improving the performance of endoscopists. Prospective consecutive patients undergoing M-IEE were enrolled from August 17, 2020 to August 2, 2021 in Renmin Hospital of Wuhan University to assess the applicability of ENDOANGEL-ME in clinical practice. RESULTS: A total of 3099 patients undergoing M-IEE were enrolled in this study. The diagnostic accuracy of ENDOANGEL-ME for diagnosing EGC was 88.44% and 90.49% in internal and external images, respectively. In 93 internal videos, ENDOANGEL-ME achieved an accuracy of 90.32% for diagnosing EGC, significantly superior to that of senior endoscopists (70.16% ± 8.78%). In 94 external videos, with the assistance of ENDOANGEL-ME, endoscopists showed improved accuracy and sensitivity (85.64% vs 80.32% and 82.03% vs 67.19%, respectively). In 194 prospective consecutive patients with 251 lesions, ENDOANGEL-ME achieved a sensitivity of 92.59% (25/27) and an accuracy of 83.67% (210/251) in real clinical practice. CONCLUSIONS: This multicenter diagnostic study showed that ENDOANGEL-ME can be well applied in the clinical setting. (Clinical trial registration number: ChiCTR2000035116.).

Real-time artificial intelligence for detecting focal lesions and diagnosing neoplasms of the stomach by white-light endoscopy (with videos).

BACKGROUND AND AIMS: White-light endoscopy (WLE) is the most pivotal tool to detect gastric cancer in an early stage. However, the skill among endoscopists varies greatly. Here, we aim to develop a deep learning-based system named ENDOANGEL-LD (lesion detection) to assist in detecting all focal gastric lesions and predicting neoplasms by WLE. METHODS: Endoscopic images were retrospectively obtained from Renmin Hospital of Wuhan University (RHWU) for the development, validation, and internal test of the system. Additional external tests were conducted in 5 other hospitals to evaluate the robustness. Stored videos from RHWU were used for assessing and comparing the performance of ENDOANGEL-LD with that of experts. Prospective consecutive patients undergoing upper endoscopy were enrolled from May 6, 2021 to August 2, 2021 in RHWU to assess clinical practice applicability. RESULTS: Over 10,000 patients undergoing upper endoscopy were enrolled in this study. The sensitivities were 96.9% and 95.6% for detecting gastric lesions and 92.9% and 91.7% for diagnosing neoplasms in internal and external patients, respectively. In 100 videos, ENDOANGEL-LD achieved superior sensitivity and negative predictive value for detecting gastric neoplasms from that of experts (100% vs 85.5% ± 3.4% [P = .003] and 100% vs 86.4% ± 2.8% [P = .002], respectively). In 2010 prospective consecutive patients, ENDOANGEL-LD achieved a sensitivity of 92.8% for detecting gastric lesions with 3.04 ± 3.04 false positives per gastroscopy and a sensitivity of 91.8% and specificity of 92.4% for diagnosing neoplasms. CONCLUSIONS: Our results show that ENDOANGEL-LD has great potential for assisting endoscopists in screening gastric lesions and suspicious neoplasms in clinical work. (Clinical trial registration number: ChiCTR2100045963.).

Deep learning system compared with expert endoscopists in predicting early gastric cancer and its invasion depth and differentiation status (with videos).

BACKGROUND AND AIMS: We aimed to develop and validate a deep learning-based system that covers various aspects of early gastric cancer (EGC) diagnosis, including detecting gastric neoplasm, identifying EGC, and predicting EGC invasion depth and differentiation status. Herein, we provide a state-of-the-art comparison of the system with endoscopists using real-time videos in a nationwide human-machine competition. METHODS: This multicenter, prospective, real-time, competitive comparative, diagnostic study enrolled consecutive patients who received magnifying narrow-band imaging endoscopy at the Peking University Cancer Hospital from June 9, 2020 to November 17, 2020. The offline competition was conducted in Wuhan, China, and the endoscopists and the system simultaneously read patients' videos and made diagnoses. The primary outcomes were sensitivity in detecting neoplasms and diagnosing EGCs. RESULTS: One hundred videos, including 37 EGCs and 63 noncancerous lesions, were enrolled; 46 endoscopists from 44 hospitals in 19 provinces in China participated in the competition. The sensitivity rates of the system for detecting neoplasms and diagnosing EGCs were 87.81% and 100%, respectively, significantly higher than those of endoscopists (83.51% [95% confidence interval [CI], 81.23-85.79] and 87.13% [95% CI, 83.75-90.51], respectively). Accuracy rates of the system for predicting EGC invasion depth and differentiation status were 78.57% and 71.43%, respectively, slightly higher than those of endoscopists (63.75% [95% CI, 61.12-66.39] and 64.41% [95% CI, 60.65-68.16], respectively). CONCLUSIONS: The system outperformed endoscopists in identifying EGCs and was comparable with endoscopists in predicting EGC invasion depth and differentiation status in videos. This deep learning-based system could be a powerful tool to assist endoscopists in EGC diagnosis in clinical practice.