Application of Artificial Intelligence to Gastroenterology and Hepatology

// Dong Ke 1, 2, 3 , Hanwei Li 2 , Yi Zhang 2, 3 , Yinghong An 4 , Hanjiang Fu 2 , Xuedong Fang 1, 3 , Xiaofei Zheng 2 1 General Surgery, The Second Hospital of Jilin University, Changchun, 130041, China 2 Beijing Key Laboratory for Radiobiology, Beijing institute of Radiation Medicine, Beijing, 100850, China 3 Gastrointestinal Colorectal and Anal Surgery, The China-Japan Union Hospital of Jilin University, Changchun, 130033, China 4 Clinical Laboratory Center, Chinese PLA Air Force General Hospital, Beijing, 100142, China Correspondence to: Hanjiang Fu, email: Fuhj75@126.com Xuedong Fang, email: fangxuedong@medmail.com.cn Xiaofei Zheng, email: xfzheng100@126.com Keywords: gastric cancer, long noncoding RNA, plasma, biomarker, RNA fragments Received: November 30, 2016      Accepted: February 14, 2017      Published: February 22, 2017 ABSTRACT Background: Suitable diagnostic markers for cancers are urgently required in clinical practice. Long non-coding RNAs, which have been reported in many cancer types, are a potential new class of biomarkers for tumor diagnosis. Results: Five lncRNAs, including AK001058, INHBA-AS1, MIR4435-2HG, UCA1 and CEBPA-AS1 were validated to be increased in gastric cancer tissues. Furthermore, we found that plasma level of these five lncRNAs were significantly higher in gastric cancer patients compared with normal controls. By receiver operating characteristic analysis, we found that the combination of plasma lncRNAs with the area under the curve up to 0.921, including AK001058, INHBA-AS1, MIR4435-2HG, and CEBPA-AS1, is a better indicator of gastric cancer than their individual levels or other lncRNA combinations. Simultaneously, we found that the expression levels of a series of MIR4435-2HG fragments are different in gastric cancer plasma samples, but most of them higher than that in healthy control plasma samples. Materials and Methods: LncRNA gene expression profiles were analyzed in two pairs of human gastric cancer and adjacent non-tumor tissues by microarray analysis. Nine gastric cancer-associated lncRNAs were selected and assessed by quantitative real-time polymerase chain reaction in gastric tissues, and 5 of them were further analyzed in gastric cancer patients’ plasma. Conclusions: Our results demonstrate that certain lncRNAs, such as AK001058, INHBA-AS1, MIR4435-2HG, and CEBPA-AS1, are enriched in human gastric cancer tissues and significantly elevated in the plasma of patients with gastric cancer. These findings indicate that the combination of these four lncRNAs might be used as diagnostic or prognostic markers for gastric cancer patients.

/pdf/the-combination-of-circulating-long-noncoding-rnas-ak001058-2q2vejnxvu.pdf

The combination of circulating long noncoding RNAs AK001058, INHBA-AS1, MIR4435-2HG, and CEBPA-AS1 fragments in plasma serve as diagnostic markers for gastric cancer.

Gastrointestinal (GI) disease is one of the most common diseases and primarily examined by GI endoscopy. Recently, deep learning (DL), in particular convolutional neural networks (CNNs) have made achievements in GI endoscopy image analysis. This review focuses on the applications of DL methods in the analysis of GI images. We summarized and compared the latest published literature related to the common clinical GI diseases and covers the key applications of DL in GI image detection, classification, segmentation, recognition, location, and other tasks. At the end, we give a discussion on the challenges and the research directions of GI image analysis based on DL in the future.

/pdf/review-on-the-applications-of-deep-learning-in-the-analysis-5dwidlyuoh.pdf

Review on the Applications of Deep Learning in the Analysis of Gastrointestinal Endoscopy Images

Gastric cancer is the fourth leading cause of cancer-related mortality across the globe, with a 5-year survival rate of less than 40%. In recent years, several applications of artificial intelligence (AI) have emerged in the gastric cancer field based on its efficient computational power and learning capacities, such as image-based diagnosis and prognosis prediction. AI-assisted diagnosis includes pathology, endoscopy, and computerized tomography, while researchers in the prognosis circle focus on recurrence, metastasis, and survival prediction. In this review, a comprehensive literature search was performed on articles published up to April 2020 from the databases of PubMed, Embase, Web of Science, and the Cochrane Library. Thereby the current status of AI-applications was systematically summarized in gastric cancer. Moreover, future directions that target this field were also analyzed to overcome the risk of overfitting AI models and enhance their accuracy as well as the applicability in clinical practice.

Artificial intelligence in gastric cancer: Application and future perspectives

Fine needle aspiration cytology (FNAC) entails using a narrow gauge (25-22 G) needle to collect a sample of a lesion for microscopic examination. It allows a minimally invasive, rapid diagnosis of tissue but does not preserve its histological architecture. FNAC is commonly used for diagnosis of breast cancer, with traditional practice being based on the subjective visual assessment of the breast cytopathology cell samples under a microscope to evaluate the state of various cytological features. Therefore, there are many challenges in maintaining consistency and reproducibility of findings. However, the advent of digital imaging and computational aid in diagnosis can improve the diagnostic accuracy and reduce the effective workload of pathologists. This paper presents a comparison of various deep convolutional neural network (CNN) based fine-tuned transfer learned classification approach for the diagnosis of the cell samples. The proposed approach has been tested using VGG16, VGG19, ResNet-50 and GoogLeNet-V3 (aka Inception V3) architectures of CNN on an image dataset of 212 images (99 benign and 113 malignant), later augmented and cleansed to 2120 images (990 benign and 1130 malignant), where the network was trained using images of 80% cell samples and tested on the rest. This paper presents a comparative assessment of the models giving a new dimension to FNAC study where GoogLeNet-V3 (fine-tuned) achieved an accuracy of 96.25% which is highly satisfactory.

Comparative assessment of CNN architectures for classification of breast FNAC images.

Identification of lesion images from gastrointestinal endoscope based on feature extraction of combinational methods with and without learning process

The invention relates to the field of electrocardiosignal processing and provides an electrocardiogram pseudo-difference signal identification method and an electrocardiogram pseudo-difference signal identification device for solving the problems such as high calculation complexity and low robustness and adaptability existing in the conventional ECG pseudo-difference signal identification method The method mainly comprises the following steps: initializing an ECG signal; respectively calculating the first information parameter Rmax, the second information parameter Plambda 1 and the third information parameter Psn of the selected segmentation ECG signal by virtue of correlation analysis, principal component analysis and frequency-domain analysis; and identifying the ECG segmented signal according to the previous three information parameter values of the segment The experimental result proves that the accuracy, sensitivity and positive predictive value of the identified ECG pseudo-difference signal respectively reach 9742 percent, 6921 percent and 9206 percent; and therefore, the identification accuracy is high In addition, according to the technical scheme provided by the invention, prior information is not needed, and the method and device have excellent real-time property, robustness and adaptability

Electrocardiogram pseudo-difference signal identification method and device

The invention relates to the field of electrocardiograph monitoring, and provides a wireless remote electrocardiograph monitoring system. The problems that an electrocardiograph monitoring system at present is weak in anti-jamming capability, large in power dissipation and larger in occupancy of network resources are solved. The wireless remote electrocardiograph monitoring system comprises an electrocardiograph signal collecting and transmitting device, an edge router and a server end monitoring device, wherein the electrocardiograph signal collecting and transmitting device comprises an electrocardiograph signal pre-processing module, an electrocardiograph collecting controller, a dual-port RAM, a wireless transmission controller and a wireless transmission module. According to the technical scheme, a ping-pong storage structure based on the dual-port RAM is adopted, the data transmission bandwidth is improved, and invalid occupation of a wireless channel is reduced; in addition, the wireless transmission mode based on the 6LoWPAN protocol is adopted, so that complicated networking protocol conversion is effectively avoided while the anti-jamming capability of wireless transmission is improved and the power consumption of the system is reduced.

Wan Yongli

Papers

Identification of lesion images from gastrointestinal endoscope based on feature extraction of combinational methods with and without learning process

Electrocardiogram pseudo-difference signal identification method and device

Wireless remote electrocardiograph monitoring system