Abstract
Gastric cancer (GC) is one of the commonest cancers with high morbidity and mortality in the world. How to realize precise diagnosis and therapy of GC owns great clinical requirement. In recent years, artificial intelligence (AI) has been actively explored to apply to early diagnosis and treatment and prognosis of gastric carcinoma. Herein, we review recent advance of AI in early screening, diagnosis, therapy and prognosis of stomach carcinoma. Especially AI combined with breath screening early GC system improved 97.4 % of early GC diagnosis ratio, AI model on stomach cancer diagnosis system of saliva biomarkers obtained an overall accuracy of 97.18 %, specificity of 97.44 %, and sensitivity of 96.88 %. We also discuss concept, issues, approaches and challenges of AI applied in stomach cancer. This review provides a comprehensive view and roadmap for readers working in this field, with the aim of pushing application of AI in theranostics of stomach cancer to increase the early discovery ratio and curative ratio of GC patients.
Introduction
Up to date, Gastric cancer (GC) is the fourth commonest carcinoma and the third primary cause of cancer associated deaths in the world [1]. As the rapid expansion of precise medicine, the mortality of stomach carcinoma has begun to decrease in some countries. However, so far China is still one of the top five countries with high morbidity and mortality of GC [2]. GC is closely associated with those factors such as lifestyle, heredity and environment [3]. GC is divided into two stages: (1) EGC: Early Gastric Cancer; (2) AGC: Advanced Gastric Cancer. In hospital, endoscopy and pathological biopsy are general means for diagnosis of EGC and AGC [4]. Up to date, discovery rate of EGC in China is still less than 15 % due to vague symptoms of EGC [5]. The patients with AGC have 24 % five-year survival rate, whereas the patients with EGC have more than 90 % [6]. To screen out EGC patients owns huge clinical requirement.
In order to solve problem of detection of EGC, prewarning and early diagnosis system of GC has been being developed since 1999 [7]. Gene expression profile chip was used to screen out differently expressed genes associated with EGC, AGC and normal gastric tissues respectively, and GC prewarning gene chips with primary diagnosis standard was developed, GC pre-warning database and information analysis platform was established [8].
GC prewarning and early theranostics system based on biomarkes was also studied [9]. Some GC biomarkers associated with prewarning, diagnosis and staging of GC were screened and identified [10], [11], [12], [13], [14]. For example, two novel plasma microRNA biomarkers such as miR-16-5p and miR-19b-3p were identified to be capable of distinguishing GC patients with different tumor node metastasis (TNM) stages and differentiation grades [10]. Four key GC circulating exosomal microRNAs such as hsa-miR-130b-3p, hsa-miR-151a-3p, hsa-miR-15b-3p and hsa-miR-1246 were identified [11], a suspension array with aggregation-induced emission luminogen barcodes microspheres was developed, and was successfully used for multiplex detection of GC microRNAs [12, 13].
Fourteen of volatile organic compound (VOC) breath biomarkers were screened out to distinguish EGC and AGC patients from healthy persons, a graphene oxide-gold nanoparticles-based sensor was prepared to detect these VOC biomarkers. This method was successfully used to detect 200 breath samples from clinical patients with a sensitivity of 83 % and a specificity of 92 % [14].
Ten kinds of amino acids in human saliva were screened out as metabolic biomarkers to differentiate EGC and AGC patients from healthy people [15]. Then, the ultrasensitive sensors based on graphene oxide nanoscrolls wrapped with gold nanoparticles were fabricated to measure amino acid biomarkers in saliva, which successfully distinguished EGC and AGC patients from healthy population by using 220 clinical saliva samples with excellent performance (specificity>87.7 % and sensitivity>80 %) [15, 16]. The salivary detection method based on the surface enhanced raman scattering (SERS) sensors will revolutionize the technique of screening EGC and AGC patients from population.
As rapid development of molecular imaging, multi-functional nanoprobes used for theranostics of GC were developed, realizing simultaneous imaging and treatment of GC [17], [18], [19], [20], [21], [22], [23], [24], [25]. Magnetic nanoparticles-labeled lateral flow test chips and quantitative devices were developed [26], [27], [28], and a series of biosensors and microfluidic chip were developed for fast ultrasensitive detection of GC biomarkers [26, 29], [30], [31], [32], [33], [34], including carcinoembryonic antigen (CEA), carbohydrate antigen 19-9 (CA19-9), Helicobacter pylori CagA protein (H.P.), P53 oncoprotein (P53), pepsinogen I and II (PG I, PG-II), HAI-178, GC circulation cancer cells (CTC) [35, 36].
Up to date, GC therapeutic methods mainly include surgery, radiation and chemotherapies, which are generally very effective for early and in situ GC, but advanced and metastatic cases do not produce effective respond to chemotherapy or radiation therapy [37]. Resistance to chemotherapy-induced apoptosis is one main factor for the failure of conventional therapies. The current prognosis of GC patients is very poor with 5-year survivals of less than 24 % [38]. Therefore, how to recognize, track or kill EGC cells is still a great challenge.
GC is not very sensitive to traditional treatment such as chemotherapy, and radiation therapy, which may be closely associated with many intrinsic or acquired properties of gastric cancer stem cells (CSCs) [39, 40]. GC stem cells can result in GC and are main cause for invasion, metastasis, and resistance to conventional therapies [41]. Therefore, GCSC-based targeted therapy is one innovative effective therapeutic direction, multifunctional nanoprobes for targeted imaging and therapy of GCSCs were developed [42].
The factors for recurrence and metastasis of GC are summarized as follows [43]: (I) The intrinsic antigenicity weakness of tumor cells, immunological surveillance of the host cannot identify and eliminate the malignant cells that are distributed out of the resection field and the peripheral lymphoid organs. (II) Immunological surveillance defect or dysfunction of the host. (III) Most tumor therapeutic drugs own serious toxicities, may cause extreme weak immunity of the patients. Therefore, it is very necessary to develop new therapeutic strategies that could enhance the host immunosurveilliance and/or improve immunogenicity of the tumor cells. We also developed GC antigens fused vaccine and achieve better therapeutic efficacy [44].
Artificial intelligence (AI) means the ability of a machine to learn and display intelligence [45]. Machine learning (ML) is an implementation method of AI, is also the core of AI, and is the study of learning algorithms. Common ML algorithms include supervised learning, unsupervised learning, semi-supervised learning, reinforcement learning, deep learning (DL), transfer learning, etc. [46]. For example, DL, which is an innovative method of ML, is capable of keeping machines to analyze various training images and use backpropagation algorithms to extract specific image features [47]. This approach is inspired by the biological neural network of the human brain, and uses a layered structure of algorithms called multi-layered artificial neural networks. In addition, just like our brains, DL models can use logic to analyze data, recognize patterns, draw conclusions, and make decisions [48].
In the past decade, AI has been widely explored to apply for medical engineering [48, 49], which is displayed as the increasing number of medical devices with embedded AI algorithms on the market, and the increasing number of AI papers published in journals.
AI has three obvious technology advantages. Firstly, AI is easy to optimize and can perform cost-effective and flexible nonlinear modeling of large data sets. Secondly, these models can make knowledge dissemination easier by providing explanations. For example, using rule extraction or sensitivity analysis [49]. Thirdly, unlike machines, the performance of the human brain may be affected by fatigue, stress, or limited experience. AI technology will make up for the limited capabilities of humans, prevent human error, give machines some reliable autonomy, and improve work productivity and efficiency.
There are different types of AI computer systems to achieve different functions in cancer management. The two main categories of AI systems are computer-aided detection (CADe) for lesion detection and computer-aided diagnosis (CADx) for optical biopsy and lesion characterization. Other AI systems provide treatment assistance, such as personalized therapy training and surgical skills training and assessment [50]. In addition, there are other AI systems that provide technical support for disease prediction based on patient data. Some examples of applying AI in cancer management include image interpretation [51], surgical interventions [52], drug discovery [52], hospital-wide data analysis [53], and personalized treatment [54].
Herein, we review the recent advances of AI-assisted GC diagnosis, especially in the early screening and diagnosis of GC, and GC treatment and prognosis (Scheme 1). We also discuss the concept, issue and approaches and challenges, with the aim of attracting a lot of scientists to use AI to solve precise theranostics of GC to increase the early discovery ratio and curative ratio of GC patients in near future.
Application of AI in GC. AI, artificial intelligence; GC, gastric cancer.
Advance of AI in GC
The 5-year survival rate for advanced GC is 5 %–25 %, while it can be 90 % for EGC [55]. Early detection and curative treatment are the best strategies to improve the survival rate of patients. For example, Japan’s nationwide large-scale GC screening program has reduced related mortality [56]. However, so far morbidity and mortality of GC in the world is still very high, how to realize precise theranostics of GC has great clinical requirement. In recent years, AI are actively applied for clinical GC research, great advances of AI application for GC have been achieved, and deeply improve precise theranostics of GC. Up to date, AI can offer invaluable assistance in the management of EGC from following seven different levels:
Screening: Deep staked sparse autoencoder neural network (DSSAENN) for EGC screening based on VOC biomarkers.
Screening: Support vector machine (SVM) for EGC screening via Saliva biomarkers.
Diagnosis: CADx for EGC detection in esophagogastroduodenoscopy (EGD).
Diagnosis: CADx for invasion depth estimation and cancer staging.
Diagnosis: AI systems for prediction of H. pylori infection.
Treatment: AI systems for GC treatment.
Prognosis: AI Systems for lymph node metastasis (LNM) prediction and survival prediction.
Herein, we review the main advances of AI in GC from above-mentioned seven levels.
AI in screening EGC via VOC biomarkers
Muhammad et al. constructed a CADe system which used a softmax classifier and a DSSAENN (DSSAENN is a network model that automates the process of encoding and decoding operations) to classify GC based on VOC breath biomarkers to distinguish EGC, advanced GC and healthy persons [14]. All the breath samples were collected from the Shanghai Tongren Hospital, Shanghai, China [57]. They got an overall accuracy of 89.7 % for AGC and 97.3 % for EGC detection [57].
As shown in Figure 1, the second model was developed with [100 40] size of autoencoder, 100 and 40 are the number of neurons in the 1st and 2nd hidden layers respectively. This model produces an overall accuracy of 96.3 %, this model misclassified only four samples of early-stage gastric cancer. Moreover, this model produces very good accuracy for predicting healthy person and advanced GC as well. This model yields an accuracy of 97.4 %, 98.0 % and 93.3 % for EGC, Healthy and AGC patients respectively [57]. Up to date, this model is being tested and optimized in multi-clinical centers, own clinical translation prospects.
![Figure 1:
Breath analysis based EGC classification from DSSAENN (Image source: Reproduced from Aslan et al. [57], 2021 with permission of Springer Nature). DSSAENN, deep staked sparse autoencoder neural network; EGC, early gastric cancer.](/document/doi/10.1515/mr-2022-0042/asset/graphic/j_mr-2022-0042_fig_001.jpg)
Breath analysis based EGC classification from DSSAENN (Image source: Reproduced from Aslan et al. [57], 2021 with permission of Springer Nature). DSSAENN, deep staked sparse autoencoder neural network; EGC, early gastric cancer.
SVM based classification system for GC via saliva
Based on our previous report [15], 220 saliva samples were collected and analyzed, 10 kinds of amino acid biomarkers were identified to distinguish the patients with GC from persons without GC in the saliva samples with dominant peaks. As shown in Figure 2, the SVM was used for binary classification. The learning algorithm of SVM is the optimization algorithm for solving convex quadratic programming. The processed Raman dataset was used to train and test the established model. SVM based neural networks were developed using different kernels. Accuracy, specificity, sensitivity, and receiver operating characteristics (ROC) were applied to evaluate the classification model, along with mean average error (MAE), mean square Error (MSE), sum average error (SAE), and sum square error (SSE). Finally, we obtained an overall accuracy of 97.18 %, specificity of 97.44 %, and sensitivity of 96.88 % for the proposed model [58, 59].
![Figure 2:
Methodology of GC diagnosis model using SVM (Image source: Aslam et al. [58]). GC, gastric cancer; SVM, support vector machine. (Reproduced from Aslan et al. [58], 2015 with permission of NBE).](/document/doi/10.1515/mr-2022-0042/asset/graphic/j_mr-2022-0042_fig_002.jpg)
Our proposed method for the classification of GC is non-invasive, cheap, and faster. With the combination of SERS sensors, our proposed model has provided us an entirely new diagnostic way of GC. The proposed model is capable of playing an important role in clinics. This established method owns the prospect of clinical translation.
CADx for EGC detection in EGD
EGD and biopsy are two methods that are mostly used for the diagnosis of EGC. Although EGD is the standard procedure for diagnosing GC, the false negative rate of EGD for GC is 4.6 %–25.8 % [60–65]. According to recent studies [66, 67], the missed diagnosis rate of gastroscopy doctors with less than 10 years of work experience is about 25 %. In addition, even experienced Chinese gastroscopists, due to the heavy burden of medical image analysis, that is, each gastroscopist has about 50 patients per day, so it is almost inevitable that they will encounter missed or misdiagnosed.
The application of CADx technology in EGD has many advantages. It can reduce the variability between operators, improve the accuracy of diagnosis, and help make treatment decisions quickly and accurately on site. In addition, CADx will reduce the time, cost and burden of endoscopic surgery [68].
Hirasawa et al. constructed a convolutional neural network (CNN)-based diagnostic system for detecting GC using stored images of EGD [69]. CNN is to convert a picture into a feature vector, through multiple layers of convolution, pooling, full connection, reduce the dimension of the picture, and finally transform it into a one-dimensional vector, this vector contains the characteristics of the picture. The CNN required 47 s to analyze 2296 test images obtained from two hospitals (Cancer Institute Hospital Ariake, Tokyo, Japan, and TokatsuTsujinaka Hospital, Chiba, Japan) and two clinics (Tada Tomohiro Institute of Gastroenterology and Proctology, Saitama, Japan, and Lalaport Yokohama Clinic, Kanagawa, Japan) and it correctly diagnosed 71 of 77 GC lesions with an overall sensitivity of 92.2 %, and 161 non-cancerous lesions were detected as GC, resulting in a positive predictive value of 30.6 %. Seventy of the 71 lesions (98.6 %) with a diameter of 6 mm or more as well as all invasive cancers were correctly detected. As shown in Figure 3, Li et al. developed a new system based on CNN to analyze gastric mucosal lesions observed by magnifying endoscopy with narrow-band imaging (M-NBI). They concluded that there is no significant difference in diagnostic specificity and accuracy between CNN and experts, but the diagnostic sensitivity of CNN is significantly higher than that of experts. However, this study excluded type 0-I and type 0-III lesions, and the scope of application of the CNN system was limited [51].
![Figure 3:
The endoscopist uses the green rectangle to manually mark the location of the cancer in each image. The yellow rectangle is generated by CNN to identify suspected lesions and indicate the degree of GC. Although CNN did not identify GC in the distant view (a) it correctly located the GC in the near view (b) (Image source: Reproduced from Hirasawa et al. [69], 2018 with permission of Springer). (c) The Structure of SVM that Cheng et al. developed (Image source: Reproduced from Cheng et al. [70], 2007 with permission of IEEE). (d) Inception v3 model architecture that Li et al. develop (Image source: Reproduced from Li et al. [51], 2019 with permission of Springer).](/document/doi/10.1515/mr-2022-0042/asset/graphic/j_mr-2022-0042_fig_003.jpg)
The endoscopist uses the green rectangle to manually mark the location of the cancer in each image. The yellow rectangle is generated by CNN to identify suspected lesions and indicate the degree of GC. Although CNN did not identify GC in the distant view (a) it correctly located the GC in the near view (b) (Image source: Reproduced from Hirasawa et al. [69], 2018 with permission of Springer). (c) The Structure of SVM that Cheng et al. developed (Image source: Reproduced from Cheng et al. [70], 2007 with permission of IEEE). (d) Inception v3 model architecture that Li et al. develop (Image source: Reproduced from Li et al. [51], 2019 with permission of Springer).
Shibata et al. developed a method to detect and segment EGC regions from gastrointestinal endoscopic images using Mask-CNN [71]. Ikenoyama et al. evaluated whether CNN is better than endoscopists in detecting EGC [72].To sum up, AI based on deep learning through CNN has made significant progress in the field of gastroenterology.
In addition to CNN can be used for the diagnosis of GC, SVM is also often used in GC. Miyaki et al. designed an SVM-based analysis system to be used with an endoscope system [73]. It can quantitatively identify GC through blue laser imaging (BLI) magnified images obtained by endoscopy. The SVM output value for cancerous lesions being significantly greater than that for reddened lesions or surrounding tissue [73]. Cheng et al. developed a fourier transform infrared (FTIR) feature extraction method. It used continuous wavelet transform (CWT) analysis and SVM for classification to improve the accuracy of the EGC diagnosis rate by FTIR [70]. Kanesaka et al. trained two SVM to develop a CADx system and to help endoscopists identify and describe EGC. But their study focused on the small depression type EGC. Other EGCs may require different algorithms [74].
Podder et al. considered random forest (RF), decision tree (DT), k-nearest neighbor (KNN), and adaptive boost (AdaBoost) classifiers for the diagnosis of GC. The results show that when the test sample is 20 , 30, and 40 % of the data sample, RF is superior to DT, AdaBoost, and KNN in terms of accuracy, precision, and recall. For example, when testing a data set of 30 %, the accuracy values obtained by RF, DT, AdaBoost, and KNN were 86.67 , 83.33, 86.57, and 83.33 %, respectively [75].
CADx for invasion depth estimation and cancer staging
Invasion depth was classified into M, SM1, or SM2 (cancer with submucosal invasion ≥ 500 μm), MP (cancer invading the muscularis propria), SS (cancer invading the subserosa), SE (cancer invasion contiguous to the serosa or penetrating the serosa, and exposed to the peritoneal cavity), or SI (cancer invading adjacent structures) [76].
Endoscopic submucosal dissection (ESD) or endoscopic mucosal resection (EMR) can often be used for intramucosal carcinoma (M) and cancers with submucosal invasion <500 μm (SM1), while GC with deeper invasion requires surgical resection [77]. Therefore, accurate prediction of invasion depth is crucial for screening patients for endoscopic resection.
However, there is currently no reliable method to determine the invasion depth. In current clinical practice, experienced endoscopists usually use conventional endoscopes or use endoscopic ultrasound (EUS) to evaluate macroscopic features to diagnose the depth of EGC infiltration. According to reports, the accuracy of these two methods is limited, and no significant differences have been found in the diagnostic accuracy and macroscopic characteristics of EUS [78, 79]. The overall accuracy of conventional endoscopy is 69–79 % [78, 80], depending on the doctor’s experience in identifying endoscopic features. Therefore, a more accurate and objective method is needed to diagnose the invasive depth of GC.
CNN as a deep learning algorithm is often used to detect the depth of cancer invasion [81]. As shown in Figure 4, Zhu et al. constructed a CNN-CAD system, which uses ResNet50 to determine the depth of invasion and screen patients for endoscopic resection. The result is that the area under the receiver operating characteristic curve of the CNN-CAD system is 0.94 (95 % confidence interval [CI], 0.90–0.97). At a threshold of 0.5, the sensitivity is 76.47 % and the specificity is 95.56 %. The overall accuracy rate is 89.16 %. The positive and negative predictive values were 89.66 and 88.97 %, respectively. The CNN-CAD system achieves higher accuracy (17.25 %; 95 % CI, 11.63–22.59) and specificity (32.21 %; 95 % CI, 26.78–37.44) than human endoscopists [82].
![Figure 4:
Endoscopic images: (a) A tumor invasion depth restricted to the M or SM1. (b) A tumor invasion depth deeper than the SM1 (Image source: Reproduced from Zhu et al. [82], 2015 with permission of Elsevier). M, mucosa; SM1, submucosa.](/document/doi/10.1515/mr-2022-0042/asset/graphic/j_mr-2022-0042_fig_004.jpg)
Endoscopic images: (a) A tumor invasion depth restricted to the M or SM1. (b) A tumor invasion depth deeper than the SM1 (Image source: Reproduced from Zhu et al. [82], 2015 with permission of Elsevier). M, mucosa; SM1, submucosa.
Yoon et al. developed a visual Geometry Group (VGG)-16 model to classify endoscopic images as EGC (T1a or T1b) or non-EGC [83]. Nagao et al. devised three AI systems that use ResNet50 to predict the invasion depth of GC using white-light imaging (WLI), non-magnifying narrow-band imaging, and indigo-carmine dye contrast imaging respectively [76]. Kubota et al. developed a CNN system that uses the backpropagation (BP) algorithm to calculate the accuracy of diagnosing the T1, T2, T3, and T4 stagings of GC. As shown in Figure 5, the discrimination accuracy of T1–T4 stages is 77.2 %, 49.1 %, 51.0 %, and 55.3 % (the accuracy of distinguishing T1a [restricted mucosa] and T1b [submucosal invasion] is 68.9 %). The accuracy of their system can be improved by modifying the program [84]. Cho et al. established a DL algorithm for accurately predicting submucosal infiltration in endoscopic images of gastric tumors. In internal testing, the average area under the curve for the DenseNet-161 network was 0.887 (95 % CI: 0.849–0.924). In external tests, the average area under the curve reached 0.887 (0.863–0.910) [85].
![Figure 5:
Deep learning algorithm. (a) A scheme of the computer learning program with the BP algorithm that Kubota et al. proposed (Image source: Reproduced from Kubota et al. [84], 2012 with permission of Springer). (b) Schematic diagram of computer-aided diagnosis system that Li et al. proposed to diagnose the invasion depth of GC (Image source: Reproduced from Li et al. [86], 2015 with permission of Elsevier). (c) Convolutional neural network computer-aided detection system architecture (Image source: Reproduced from Zhu et al. [82], 2019 with permission of Elsevier).](/document/doi/10.1515/mr-2022-0042/asset/graphic/j_mr-2022-0042_fig_005.jpg)
Deep learning algorithm. (a) A scheme of the computer learning program with the BP algorithm that Kubota et al. proposed (Image source: Reproduced from Kubota et al. [84], 2012 with permission of Springer). (b) Schematic diagram of computer-aided diagnosis system that Li et al. proposed to diagnose the invasion depth of GC (Image source: Reproduced from Li et al. [86], 2015 with permission of Elsevier). (c) Convolutional neural network computer-aided detection system architecture (Image source: Reproduced from Zhu et al. [82], 2019 with permission of Elsevier).
SVM is also often used to diagnose the depth of GC invasion. Jiang et al. developed an SVM-based GC prognostic classifier (GC-SVM) based on a variety of clinicopathological features and immune markers to predict overall survival (OS) and disease-free survival (DFS), and explore whether the GC-SVM classifier can identify patients with stage II and stage III of GC [87]. Li et al. evaluated the accuracy of dual-energy spectrum computed tomography (DESCT) imaging with the aid of a CADx system, which uses the SVM classification method to assess the serosal infiltration of GC patients. The total classification accuracy rate reaches 90 % [86].
KNN is also a good way to diagnose the depth of GC invasion. Li et al. used the KNN classifier to distinguish between lymph node metastasis and non-lymph node metastasis, and the overall accuracy rate was 96.33 %. Compared with traditional diagnostic methods such as spiral CT (sensitivity 75.2 %, specificity 41.8 %) and multi-slice spiral CT (82.09 %), the diagnostic accuracy of lymph node metastasis (LNM) is higher [88]. As shown in Figure 6, Li et al. proposed an improved Citation-KNN method to identify the depth of tumor invasion of GC through dual-energy CT imaging. The total accuracy is 0.7692 [89].
![Figure 6:
The flowchart of the multiple instance learning based CADx system that Li et al. proposed (Image source: Reproduced from Li et al. [89], 2015 with permission of Elsevier).](/document/doi/10.1515/mr-2022-0042/asset/graphic/j_mr-2022-0042_fig_006.jpg)
The flowchart of the multiple instance learning based CADx system that Li et al. proposed (Image source: Reproduced from Li et al. [89], 2015 with permission of Elsevier).
AI systems for prediction of H. pylori infection
As shown in Figure 7, H. pylori infection can induce atrophic gastritis and intestinal metaplasia, and eventually develop into GC [90–93]. According to reports, patients with H. pylori infection have an increased risk of GC, and the incidence of H. pylori infection has decreased after H. pylori eradication. This fact has led the International Agency for Cancer Research to classify H. pylori as a clear carcinogen [94–98], which in turn has led to an increase in the prevalence of eradication therapy [99–101].
![Figure 7:
Representative endoscopic images of H. pylori positive, negative and eradicated stomach. Atrophy and diffuse redness can be seen during infection. The regular arrangement of collecting venules (RAC) can be seen in the uninfected stomach. A map-like red appears in the stomach where H. pylori has been eradicated (Image source: Reproduced from de Vrles et al. [102], 2009 with permission of WWW).](/document/doi/10.1515/mr-2022-0042/asset/graphic/j_mr-2022-0042_fig_007.jpg)
Representative endoscopic images of H. pylori positive, negative and eradicated stomach. Atrophy and diffuse redness can be seen during infection. The regular arrangement of collecting venules (RAC) can be seen in the uninfected stomach. A map-like red appears in the stomach where H. pylori has been eradicated (Image source: Reproduced from de Vrles et al. [102], 2009 with permission of WWW).
Endoscopy can help diagnose H. pylori infection. Accurate endoscopic diagnosis of H. pylori positive should be confirmed by various tests, such as blood or urine anti-H. pylori IgG and fecal antigen, urease breath, or rapid urease tests, followed by eradication therapy. In addition, patients who have been eradicated from H. pylori have a moderate risk of developing GC. Even if the eradication is successful, GC may still occur [98, 102, 103], which is why post-eradication status should be distinguished from H. pylori-negativity.
Shichijo et al. constructed a CNN and evaluated its ability to diagnose H. pylori infection. Compared with the manual diagnosis of endoscopists, CNN has higher accuracy and shorter time [104]. However, the study only included H. pylori positive and negative patients, and excluded patients after H. pylori eradication. So Shichijo et al. constructed a CNN and evaluated its ability to determine the status of all H. pylori infections [105]. Mohan et al. evaluated the comprehensive performance of CNN-based AI in diagnosing H. pylori infection. Compared to physicians, CNN seemed to perform equivalently [106]. Zheng et al. used CNN to achieve high diagnostic accuracy for evaluating H. pylori infection. The area under the curve of the poly gastric image of each patient is 0.97, and the sensitivity, specificity, and accuracy are 91.6 %, 98.6 % and 93.8 %. Their system achieved a high degree of accuracy [107]. Itoh et al. developed a CNN to detect H. pylori infection and it turned out that CNN-assisted H. pylori infection diagnosis is feasible [108].
SVM is also commonly used to diagnose H. pylori infection. Ishihara et al. proposed a method for detecting H. pylori infection based on the combination of SVM and multiple kernel learning (MKL), and the experimental results obtained by applying this method to real X-ray images proved its effectiveness [109]. Huang et al. proposed a computer-aided diagnosis system that uses sequential forward floating selection (SFFS) and SVM to diagnose H. pylori gastric histology from endoscopic images [110]. Ishihara et al. constructed an SVM to automatically detect H. pylori infection in multi-gastric X-ray images [111].
There are other algorithms that can detect H. pylori infection. Ishihara et al. used MKL to build a system that can automatically detect H. pylori infection [112]. The MKL algorithm essentially defines M base kernel functions and uses a weighted linear combination of the base functions as the kernel function of the SVM.
Treatment: AI systems for GC treatment
Despite advances in computer systems and simulation methods, surgical training is still based on the direct observation of expert surgeons [113]. Due to the subjectivity of human observation, these methods lack consistency, reliability and efficiency [114]. In addition, the widespread implementation of robotic surgery has led to an increasing need for appropriate structured training models and objective evaluation tools for clinical capabilities. The ability of ML methods to find hidden patterns in large data sets (such as sports and video data) provides the possibility to better understand and simulate surgical data to evaluate surgeons’ skills and personalized training.
Fard et al. introduced a framework for objective skill assessment and prediction based on motion trajectory data. Their goal is to establish a classification framework to automatically assess the performance of surgeons with different levels of expertise. They used three classification methods: KNN, logistics regression and SVM and they proved that this system can classify the expertise of surgeons as novices or experts [50]. Wang et al. proposed an analytical deep learning framework for surgical training skills assessment. Implement a deep CNN to map multivariate time series data of kinematics to a single skill level. The results are that in-depth architecture has great potential for efficient online skill assessment in modern surgical training [115].
As shown in Figure 8, Ershad et al. proposed a sparse coding framework. Compared with the use of principal component analysis (PCA) features or raw data, the proposed dictionary learning method can use the user’s joint position data to evaluate the performance of stylistic behavior in near real-time, and has higher accuracy [116].
![Figure 8:
Feature reduction and basis vector transformation process of Ershad’s model (Image source: Reproduced from Ershad et al. [116], 2015 with permission of Springer).](/document/doi/10.1515/mr-2022-0042/asset/graphic/j_mr-2022-0042_fig_008.jpg)
Feature reduction and basis vector transformation process of Ershad’s model (Image source: Reproduced from Ershad et al. [116], 2015 with permission of Springer).
Prognosis: AI systems for LNM prediction and survival prediction
Predicting LNM is important for clinical decisions including endoscopic mucosal resection, neoadjuvant chemotherapy, or radical surgery. In many cases, the preoperative staging of N status is not satisfactory [117]. Recently, artificial neural network (ANN) has been used to predict LNM and have significantly improved accuracy. Gao et al. used a deep neural network (DNN) to calculate a CT of the proximal metastatic lymph node (PGMLN) to simulate the radiologist’s recognition of the lymph node and obtain more accurate recognition results [118]. Matsumoto et al. used DNN analysis to test the accuracy, effectiveness and practicality of deep-UV-laser-induced fluorescence imaging in detecting LNM in patients with GC [119]. Hensler et al. used the QUEEN system to predict the LNM of GC [120]. Su et al. developed a DL system to achieve early diagnosis of peritoneal metastasis (PM). The detection network based on fast Region-CNN uses pre-trained Resnet18 to achieve cell detection with an average accuracy of 0.8316. The classification network based on Resnet50 achieved an area under curve (AUC) of 0.8851, an accuracy of 96.80 %, and an FNR of 4.73 % in cell classification [121]. As shown in Figure 9, Li et al. established a deep learning radiology nomogram based on dual-energy CT (DECT) for the prediction of LNM in GC [122].
![Figure 9:
Workflow of the dual-energy CT–based radiomics nomogram (Image source: Reproduced from Li et al. [122], 2020 with permission of Springer).](/document/doi/10.1515/mr-2022-0042/asset/graphic/j_mr-2022-0042_fig_009.jpg)
Workflow of the dual-energy CT–based radiomics nomogram (Image source: Reproduced from Li et al. [122], 2020 with permission of Springer).
Most GC patients have metastatic disease when they relapse, and the overall prognosis is still poor. The expected survival period after recurrence is less than 1 year. The high tumor recurrence rate in patients with advanced GC highlights the importance of considering adjuvant therapy. Based on the survival prediction of GC patients, as shown in Figure 10, Li et al. developed a new DL prediction algorithm (survival recurrence network [SRN]). Their SRN predicted a high survival rate, reaching 92 % in the 5th year after surgery [54]. Biglarian et al. used the Cox proportional hazard and ANN model to predict the survival rate of GC patients. Compared with the Cox proportional hazard regression model, the neural network model is a more powerful tool for predicting the survival rate of GC patients [123]. Li et al. used an SVM-based method to generate radiological signature (RS), which can predict disease-free survival (DFS) and chemotherapy response in stage II/III GC [124].
![Figure 10:
DL prediction algorithm. (a) Use the corresponding test patients to perform iterative tests on 100 well-trained SRNs, and generate ROC curves to evaluate the average predictive power at each time point. (b) The AUC of the test data set (Image source: Reproduced from Lee et al. [54], 2018 with permission of ASCO).](/document/doi/10.1515/mr-2022-0042/asset/graphic/j_mr-2022-0042_fig_010.jpg)
DL prediction algorithm. (a) Use the corresponding test patients to perform iterative tests on 100 well-trained SRNs, and generate ROC curves to evaluate the average predictive power at each time point. (b) The AUC of the test data set (Image source: Reproduced from Lee et al. [54], 2018 with permission of ASCO).
Technique challenges and prospects
AI has achieved big success in screening, diagnosis, therapy and prognosis of GC. AI can handle successfully complex nonlinear relationships, fault tolerance, parallel distributed processing and learning [125]. In view of its advantages in the simultaneous processing of adaptive, quantitative and qualitative knowledge, as well as the verification results of multiple clinical studies in multiple fields [126], AI has multiple uses in the field of clinical medicine [127]. It not only makes full use of all aspects of clinical diversity [128, 129], but also helps to solve the problem of lack of objectivity and universality in the current expert system [130]. DL techniques, in particular, are transforming our ability to interpret imaging data [131, 132].
AI technology faces some great challenges, which must be solved to ensure its application in cancer screening, diagnosis, therapy and prognosis [133]. First, medical imaging data cannot be used as input data directly. To solve this problem, it is necessary to improve the perception of the machine so that it can get real images and process them. Second, they also have the risk of overfitting the training data, resulting in a sharp drop in performance under certain settings [134]. Continuous algorithm optimization can solve this challenge. Third, Cabitza et al. [135] puts forward a more pessimistic view, which mentions the inherent uncertainty in medicine and the “black box” of neural network/ML applications that may reduce the skills of doctors and may soon change certain aspects of healthcare. Deep learning, a seemingly practical and economical method that will bring unexpected negative consequences. It is necessary to focus on a wider range of AI application scenarios, and pay attention to various issues such as data algorithm security, technology adaptation, organizational reengineering, social risk and governance that may arise from technology application before AI has a large-scale irreversible impact on human society, so as to form a systematic understanding and forward-looking prediction of the social impact characteristics and dynamics of AI and other technologies as early as possible and provide timely feedback on the optimization of the technology development path. Fourth, there are some ethical and safety issues, such as using AI after obtaining patient consent and determining who is responsible for misdiagnosis or wrong treatment [55]. In addition, AI cannot determine causality, and predictions generated by AI must be critically evaluated and interpreted by doctors in a clinically meaningful way [55]. To solve the safety and ethical issues, the design must be continuously optimized and improved, and the ethical design can be introduced to make the AI have the ability to judge the behavior similar to human through ethical design. Fifth, reflection makes people improve, and so should the progress of machine. But AI cannot reflect on experience and learn from them. Using language to guide computers to think is a good way to solve this problem. This process is similar to humans learning to think by receiving language instruction.
In the era of precision medicine, the predictability of AI in GC management is promising in the future. But we believe that AI cannot completely replace doctors, and the way to achieve the best performance is through human–machine collaboration. In the era of precision medicine, combining omics data with clinical information is necessary for future clinical practice. To this end, AI must be combined with multiple disciplines. In addition, larger and higher-quality random events need to be used to test the AI model.
In the near future, we can further study the application of AI in the whole process of GC management. In particular, some possible areas include early detection, stage recognition and prognosis. For GC detection, it has now entered an accuracy platform. To transfer CNN from the research environment to clinical practice, this accuracy platform must be established. For GC treatment, future research will focus on applying the proposed method to data from other parts of the robot, and be able to provide relevant feedback for trainees to improve performance in the case of poor detection performance. Regarding the prognosis of GC, in the future, it is necessary to study a system that can predict not only one type of cancer, but also predict other types of cancer.
Conclusions
In summary, AI has been actively explored in prewarning, early screening, diagnosis, therapy and prognosis of clinical GC, and has achieved great advances. However, how to use AI to realize precise theranostics of GC still face some big challenges, further works will focus on innovative algorithm, optimized models and validating models with big data to suit for requirement of clinical GC theranostics in near future.
Funding source: National Key Research and Development Program of China
Award Identifier / Grant number: No. 2017YFA0205301 and 2017YFA0205304
Funding source: National Natural Science Foundation of China
Award Identifier / Grant number: No. 82073380, 81921002, 82020108017
-
Research funding: This work was financially supported by the National Key Research and Development Program of China (Grant No. 2017YFA0205301 and 2017YFA0205304), National Natural Science Foundation of China (Grant No. 82073380, 81921002, 82020108017), National Postdoctoral Program for Innovative Talents (Grant No. BX20190205), China Postdoctoral Science Foundation (Grant No. 2020M671130), Projects of Shanghai Science and Technology Commission (21DZ2203200, and No. 20142201300).
-
Author contribution: Yiqing Zha and Cuili Xue wrote this manuscript, Yanlei Liu selected the figures in manuscript, Jesus M.De La Fuente read and revised it, DaxiangCui suggested to write this review mansucript, and finally corrected this mansucript.
-
Conflict of interest: The authors declare to have no conflicts of interest.
-
Ethics declarations: This article is no associated with ethics.
-
Human participants or animals: This article does not contain any studies with human participants or animals performed by any of the authors.
References
1. Johnston, FM, Beckman, M. Updates on management of gastric cancer. Curr Oncol Rep 2019;21:67. https://doi.org/10.1007/s11912-019-0820-4.Search in Google Scholar PubMed
2. Bray, F, Ferlay, J, Soerjomataram, I, Siegel, RL, Torre, LA, Jemal, A. Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA A Cancer J Clin 2018;68:394–424. https://doi.org/10.3322/caac.21492.Search in Google Scholar PubMed
3. Smyth, EC, Nilsson, M, Grabsch, HI, van Grieken, NC, Lordick, F. Gastric cancer. Lancet 2020;396:635–48. https://doi.org/10.1016/s0140-6736(20)31288-5.Search in Google Scholar PubMed
4. Wadhwa, R, Song, S, Lee, JS, Yao, Y, Wei, Q, Ajani, JA. Gastric cancer-molecular and clinical dimensions. Nat Rev Clin Oncol 2013;10:643–55. https://doi.org/10.1038/nrclinonc.2013.170.Search in Google Scholar PubMed PubMed Central
5. Pourhoseingholi, MA, Vahedi, M, Baghestani, AR. Burden of gastrointestinal cancer in Asia; an overview. Gastroenterol Hepatol Bed Bench 2015;8:19–27.Search in Google Scholar
6. Tan, Z. Recent advances in the surgical treatment of advanced gastric cancer: a review. Med Sci Mon Int Med J Exp Clin Res 2019;25:3537–41. https://doi.org/10.12659/msm.916475.Search in Google Scholar PubMed PubMed Central
7. Cui, D, Ma, L, Zhi, X, Zhang, C. Advance and prospects of nanotheranostic technology for gastric cancer. Nano Biomed Eng 2015;3:274–80.Search in Google Scholar
8. Cui, DX, Zhang, L, Yan, XJ, Zhang, LX, Xu, JR, Guo, YH, et al.. A microarray-based gastric carcinoma prewarning system. World J Gastroenterol 2005;11:1273–82. https://doi.org/10.3748/wjg.v11.i9.1273.Search in Google Scholar PubMed PubMed Central
9. Cui, D. Gastric cancer prewarning and early theranostics system based on nanotechnology. J Shanghai Jiaot Univ 2019;52:1396–403.Search in Google Scholar
10. Zhang, J, Song, Y, Zhang, C, Zhi, X, Fu, H, Ma, Y, et al.. Circulating MiR-16-5p and MiR-19b-3p as two novel potential biomarkers to indicate progression of gastric cancer. Theranostics 2015;5:733–45. https://doi.org/10.7150/thno.10305.Search in Google Scholar PubMed PubMed Central
11. Qian, X, Xie, F, Wei, H, Cui, D. Identification of key circulating exosomal microRNAs in gastric cancer. Front Oncol 2021;11:693360. https://doi.org/10.3389/fonc.2021.693360.Search in Google Scholar PubMed PubMed Central
12. Zou, D, Wu, W, Zhang, J, Ma, Q, Fan, S, Cheng, J, et al.. Multiplex detection of miRNAs based on aggregation-induced emission luminogen encoded microspheres. RSC Adv 2019;9:39976–85. https://doi.org/10.1039/c9ra07680h.Search in Google Scholar PubMed PubMed Central
13. Liu, X, Wu, W, Cui, D, Chen, X, Li, W. Functional micro-/nanomaterials for multiplexed biodetection. Adv Mater 2021;33:e2004734. https://doi.org/10.1002/adma.202004734.Search in Google Scholar PubMed
14. Chen, Y, Zhang, Y, Pan, F, Liu, J, Wang, K, Zhang, C, et al.. Breath analysis based on surface-enhanced Raman scattering sensors distinguishes early and advanced gastric cancer patients from healthy persons. ACS Nano 2016;10:8169–79. https://doi.org/10.1021/acsnano.6b01441.Search in Google Scholar PubMed
15. Chen, Y, Cheng, S, Zhang, A, Song, J, Chang, J, Wang, K, et al.. Salivary analysis based on surface enhanced Raman scattering sensors distinguishes early and advanced gastric cancer patients from healthy persons. J Biomed Nanotechnol 2018;14:1773–84. https://doi.org/10.1166/jbn.2018.2621.Search in Google Scholar PubMed
16. Zhang, A, Chang, J, Chen, Y, Huang, Z, Alfranca, G, Cui, D, et al.. Spontaneous implantation of gold nanoparticles on graphene oxide for salivary SERS sensing. Anal Methods 2019;11:5089–97. https://doi.org/10.1039/c9ay01500k.Search in Google Scholar
17. Wang, L, Zhang, C, Hong, Y, Li, X, Li, T, Cui, D, et al.. Integrating epigenetic modulators in nanofibers for synergistic gastric cancer therapy via epigenetic reprogramming. Nano Lett 2021;21:298–307. https://doi.org/10.1021/acs.nanolett.0c03665.Search in Google Scholar PubMed
18. Pan, S, Pei, L, Zhang, A, Zhang, Y, Zhang, C, Cui, D, et al.. Passion fruit-like exosome-PMA/Au-BSA@Ce6 nanovehicles for real-time fluorescence imaging and enhanced targeted photodynamic therapy with deep penetration and superior retention behavior in tumor. Biomaterials 2020;230:119606. https://doi.org/10.1016/j.biomaterials.2019.119606.Search in Google Scholar PubMed
19. Zhi, X, Liu, Y, Lin, L, Yang, M, Zhang, L, Cui, D, et al.. Oral pH sensitive GNS@ab nanoprobes for targeted therapy of Helicobacter pylori without disturbance gut microbiome. Nanomedicine 2019;20:102019. https://doi.org/10.1016/j.nano.2019.102019.Search in Google Scholar PubMed
20. Wang, K, Ruan, J, Qian, Q, Song, H, Bao, C, Cui, D, et al.. BRCAA1 monoclonal antibody conjugated fluorescent magnetic nanoparticles for in vivo targeted magnetofluorescent imaging of gastric cancer. J Nanobiotechnol 2011;9:23. https://doi.org/10.1186/1477-3155-9-23.Search in Google Scholar PubMed PubMed Central
21. Ruan, J, Song, H, Qian, Q, Li, C, Wang, K, Cui, D, et al.. HER2 monoclonal antibody conjugated RNase-A-associated CdTe quantum dots for targeted imaging and therapy of gastric cancer. Biomaterials 2012;33:7093–102. https://doi.org/10.1016/j.biomaterials.2012.06.053.Search in Google Scholar PubMed
22. He, M, Huang, P, Zhang, C, Hu, H, Bao, C, Cui, D, et al.. Dual phase-controlled synthesis of uniform lanthanide-doped NaGdF4 upconversion nanocrystals via an OA/ionic liquid two-phase system for in vivo. Dual-Modal Imag 2011;21:4470–7. https://doi.org/10.1002/adfm.201101040.Search in Google Scholar
23. Li, Z, Huang, P, Zhang, X, Lin, J, Yang, S, Cui, D, et al.. RGD-conjugated dendrimer-modified gold nanorods for in vivo tumor targeting and photothermal therapy. Mol Pharm 2009;7:94–104. https://doi.org/10.1021/mp9001415.Search in Google Scholar PubMed
24. Huang, P, Lin, J, Wang, X, Wang, Z, Zhang, C, Cui, D, et al.. Light-triggered theranostics based on photosensitizer-conjugated carbon dots for simultaneous enhanced-fluorescence imaging and photodynamic therapy. Adv Mater 2012;24:5104–10. https://doi.org/10.1002/adma.201200650.Search in Google Scholar PubMed PubMed Central
25. Yin, T, Wu, H, Zhang, Q, Gao, G, Shapter, JG, Cui, D, et al.. In vivo targeted therapy of gastric tumors via the mechanical rotation of a flower-like Fe3O4@Au nanoprobe under an alternating magnetic field. NPG Asia Mater 2017;9:e408. https://doi.org/10.1038/am.2017.117.Search in Google Scholar
26. Yan, X, Wang, K, Lu, W, Qin, W, Cui, D, He, J. CdSe/ZnS quantum dot-labeled lateral flow strips for rapid and quantitative detection of gastric cancer carbohydrate antigen 72-4. Nanoscale Res Lett 2016;11:138. https://doi.org/10.1186/s11671-016-1355-3.Search in Google Scholar PubMed PubMed Central
27. Gui, C, Wang, K, Li, C, Dai, X, Cui, D. A CCD-based reader combined with CdS quantum dot-labeled lateral flow strips for ultrasensitive quantitative detection of CagA. Nanoscale Res Lett 2014;9:57. https://doi.org/10.1186/1556-276x-9-57.Search in Google Scholar
28. Mou, X, Li, T, Wang, J, Ali, Z, Zhang, Y, Cui, D, et al.. Genetic variation of BCL2 (rs2279115), NEIL2 (rs804270), LTA (rs909253), PSCA (rs2294008) and PLCE1 (rs3765524, rs10509670) genes and their correlation to gastric cancer risk based on universal tagged arrays and Fe3O4 magnetic nanoparticles. J Biomed Nanotechnol 2015;11:2057–66. https://doi.org/10.1166/jbn.2015.2113.Search in Google Scholar PubMed
29. Xie, Y, Zhi, X, Su, H, Wang, K, Yan, Z, Cui, D, et al.. A novel electrochemical microfluidic chip combined with multiple biomarkers for early diagnosis of gastric cancer. Nanoscale Res Lett 2015;10:477. https://doi.org/10.1186/s11671-015-1153-3.Search in Google Scholar PubMed PubMed Central
30. Gao, S, Kang, L, Deng, M, Ji, B, Liu, J, Yang, H, et al.. A giant magnetoimpedance-based microfluidic system for multiplex immunological assay. Nano Biomed Eng 2016;8:240–5. https://doi.org/10.5101/nbe.v8i4.p240-245.Search in Google Scholar
31. Wang, K, Yang, J, Xu, H, Cao, B, Qin, Q, Cui, D, et al.. Smartphone-imaged multilayered paper-based analytical device for colorimetric analysis of carcinoembryonic antigen. Anal Bioanal Chem 2020;412:2517–28. https://doi.org/10.1007/s00216-020-02475-1.Search in Google Scholar PubMed
32. Liu, P, Qian, X, Li, X, Fan, L, Li, X, Cui, D, et al.. Enzyme-free electrochemical biosensor based on localized DNA cascade displacement reaction and versatile DNA nanosheets for ultrasensitive detection of exosomal MicroRNA. ACS Appl Mater Interfaces 2020;12:45648–56. https://doi.org/10.1021/acsami.0c14621.Search in Google Scholar PubMed
33. Zhang, J, Li, C, Zhi, X, Ramón, GA, Liu, Y, Cui, D, et al.. Hairpin DNA-templated silver nanoclusters as novel beacons in strand displacement amplification for MicroRNA detection. Anal Chem 2016;88:1294–302. https://doi.org/10.1021/acs.analchem.5b03729.Search in Google Scholar PubMed
34. Zhang, J, Liu, Y, Zhi, X, Zhang, C, Liu, TF, Cui, D. DNA-templated silver nanoclusters locate microRNAs in the nuclei of gastric cancer cells. Nanoscale 2018;10:11079–90. https://doi.org/10.1039/c8nr02634c.Search in Google Scholar PubMed
35. Yu, Z, Lin, S, Xia, F, Liu, Y, Zhang, D, Chen, D, et al.. ExoSD chips for high-purity immunomagnetic separation and high-sensitivity detection of gastric cancer cell-derived exosomes. Biosens Bioelectron 2021;194:113594. https://doi.org/10.1016/j.bios.2021.113594.Search in Google Scholar PubMed
36. Chang, J, Zhang, A, Huang, Z, Chen, Y, Zhang, Q, Cui, D. Monodisperse Au@Ag core-shell nanoprobes with ultrasensitive SERS-activity for rapid identification and Raman imaging of living cancer cells. Talanta 2019;198:45–54. https://doi.org/10.1016/j.talanta.2019.01.085.Search in Google Scholar PubMed
37. Sexton, RE, Al Hallak, MN, Diab, M, Azmi, AS. Gastric cancer: a comprehensive review of current and future treatment strategies. Cancer Metastasis Rev 2020;39:1179–203. https://doi.org/10.1007/s10555-020-09925-3.Search in Google Scholar PubMed PubMed Central
38. Strong, VE. Progress in gastric cancer. Updates Surg 2018;70:157–9. https://doi.org/10.1007/s13304-018-0543-3.Search in Google Scholar PubMed
39. Brabletz, T, Jung, A, Spaderna, S, Hlubek, F, Kirchner, T. Migrating cancer stem cells – an integrated concept of malignant tumour progression. Nat Rev Cancer 2005;5:744–9. https://doi.org/10.1038/nrc1694.Search in Google Scholar PubMed
40. Vinogradov, S, Wei, X. Cancer stem cells and drug resistance: the potential of nanomedicine. Nanomedicine (London, England) 2012;7:597–615. https://doi.org/10.2217/nnm.12.22.Search in Google Scholar PubMed PubMed Central
41. Gilbertson, RJ, Graham, TA. Cancer: resolving the stem-cell debate. Nature 2012;488:462–3. https://doi.org/10.1038/nature11480.Search in Google Scholar PubMed
42. Liang, S, Li, C, Zhang, C, Chen, Y, Xu, L, Cui, D, et al.. CD44v6 monoclonal antibody-conjugated gold nanostars for targeted photoacoustic imaging and plasmonic photothermal therapy of gastric cancer stem-like cells. Theranostics 2015;5:970–84. https://doi.org/10.7150/thno.11632.Search in Google Scholar PubMed PubMed Central
43. Zhang, D, Fan, D. New insights into the mechanisms of gastric cancer multidrug resistance and future perspectives. Future Oncol 2010;6:527–37. https://doi.org/10.2217/fon.10.21.Search in Google Scholar PubMed
44. Li, C, Liang, S, Zhang, C, Liu, Y, Yang, M, Cui, D, et al.. Allogenic dendritic cell and tumor cell fused vaccine for targeted imaging and enhanced immunotherapeutic efficacy of gastric cancer. Biomaterials 2015;54:177–87. https://doi.org/10.1016/j.biomaterials.2015.03.024.Search in Google Scholar PubMed
45. Colom, R, Karama, S, Jung, RE, Haier, RJ. Human intelligence and brain networks. Dialogues Clin Neurosci 2010;12:489–501. https://doi.org/10.31887/dcns.2010.12.4/rcolom.Search in Google Scholar
46. Sharma, H, Zerbe, N, Klempert, I, Hellwich, O, Hufnagl, P. Deep convolutional neural networks for automatic classification of gastric carcinoma using whole slide images in digital histopathology. Comput Med Imag Graph 2017;61:2–13. https://doi.org/10.1016/j.compmedimag.2017.06.001.Search in Google Scholar PubMed
47. Krizhevsky, A, Sutskever, I, Hinton, G. ImageNet classification with deep convolutional neural networks. Adv Neural Inf Process Syst 2012;25:1097–105.Search in Google Scholar
48. Hamet, P, Tremblay, J. Artificial intelligence in medicine. Metabolism 2017;69S:S36–40. https://doi.org/10.1016/j.metabol.2017.01.011.Search in Google Scholar PubMed
49. Lisboa, PJG. A review of evidence of health benefit from artificial neural networks in medical intervention. Neural Network 2002;15:11–39. https://doi.org/10.1016/s0893-6080(01)00111-3.Search in Google Scholar PubMed
50. Fard, MJ, Ameri, S, Darin Ellis, R, Chinnam, RB, Pandya, AK, Klein, MD. Automated robot-assisted surgical skill evaluation. Predict Anal Approach 2018;14:e1850. https://doi.org/10.1002/rcs.1850.Search in Google Scholar PubMed
51. Li, L, Chen, Y, Shen, Z, Zhang, X, Yu, C, Ding, Y, et al.. Convolutional neural network for the diagnosis of early gastric cancer based on magnifying narrow band imaging. Gastric Cancer 2019;23:126–32. https://doi.org/10.1007/s10120-019-00992-2.Search in Google Scholar PubMed PubMed Central
52. Hashimoto, DA, Rosman, G, Rus, D, Meireles, OR. Artificial intelligence in surgery: promises and perils. Ann Surg 2018;268:70–6. https://doi.org/10.1097/sla.0000000000002693.Search in Google Scholar
53. Biglarian, A, Hajizadeh, E, Kazemnejad, A, Zayeri, F. Determining of prognostic factors in gastric cancer PatientsUsing artificial neural networks. Asian Pac J Cancer Prev APJCP 2010;11:533–6.Search in Google Scholar
54. Lee, J, An, JY, Choi, MG, Park, SH, Kim, ST, Lee, JH, et al.. Deep learning–based survival analysis identified associations between molecular subtype and optimal adjuvant treatment of patients with gastric cancer. JCO Clin Cancer Inform 2018;2:1–14. https://doi.org/10.1200/cci.17.00065.Search in Google Scholar PubMed PubMed Central
55. Jin, P, Ji, X, Kang, W, Li, Y, Tian, Y, Ma, F, et al.. Artificial intelligence in gastric cancer: a systematic review. J Cancer Res Clin Oncol 2020;146:2339–50. https://doi.org/10.1007/s00432-020-03304-9.Search in Google Scholar PubMed
56. Karim-Kos, HE, Vries, ED, Soerjomataram, I, Lemmens, V, Siesling, S, Coebergh. Recent trends of cancer in Europe: a combined approach of incidence, survival and mortality for 17 cancer sites since the 1990s. Eur J Cancer 2008;44:1345–89. https://doi.org/10.1016/j.ejca.2007.12.015.Search in Google Scholar PubMed
57. Aslam, MA, Xue, C, Chen, Y, Zhang, A, Liu, M, Cui, D, et al.. Breath analysis based early gastric cancer classification from deep stacked sparse autoencoder neural network. Sci Rep 2021;11:4014. https://doi.org/10.1038/s41598-021-83184-2.Search in Google Scholar PubMed PubMed Central
58. Aslam, M, Cui, D, Song, J, Chen, Y, Zhang, A, Cai, W, et al.. SVM based classification and prediction system for gastric cancer using dominant features of saliva. Nano Biomed Eng 2020;12:1–13. https://doi.org/10.5101/nbe.v12i1.p1-13.Search in Google Scholar
59. Aslam, M, Xue, C, Liu, M, Wang, K, Cui, D. Classification and prediction of gastric cancer from saliva diagnosis using artificial neural network. Eng Lett 2021;29:10–24.Search in Google Scholar
60. Menon, S. Trudgill. How commonly is upper gastrointestinal cancer missed at endoscopy? A meta-analysis. Endosc Int Open 2014;2:E46–50. https://doi.org/10.1055/s-0034-1365524.Search in Google Scholar PubMed PubMed Central
61. Hosokawa, O, Hattori, M, Douden, K, Hayashi, H, Kaizaki, Y. Difference in accuracy between gastroscopy and colonoscopy for detection of cancer. J Hyg 2007;54:442–4.Search in Google Scholar
62. Hosokawa, O, Tsuda, S, Kidani, E, Watanabe, K, Tanigawa, Y, Shirasaki, S, et al.. Diagnosis of gastric cancer up to three years after negative upper gastrointestinal endoscopy. Endoscopy 1998;30:669–74. https://doi.org/10.1055/s-2007-1001386.Search in Google Scholar PubMed
63. Amin, A, Gilmour, H, Graham, L, Patersonbrown, S, Terrace, J, Crofts, T. Gastric adenocarcinoma missed at endoscopy. J R Coll Surg Edinb 2002;47:681–4.Search in Google Scholar
64. Yalamarthi, S, Witherspoon, P, McCole, D, Auld, CD. Missed diagnoses in patients with upper gastrointestinal cancers. Endoscopy 2004;36:874–9. https://doi.org/10.1055/s-2004-825853.Search in Google Scholar PubMed
65. Voutilainen, ME, Juhola, MT. Evaluation of the diagnostic accuracy of gastroscopy to detect gastric tumours: clinicopathological features and prognosis of patients with gastric cancer missed on endoscopy. Eur J Gastroenterol Hepatol 2005;17:1345–9. https://doi.org/10.1097/00042737-200512000-00013.Search in Google Scholar PubMed
66. Wang, SM, Qiao, YL. Implementation of cervical cancer screening and prevention in China – challenges and reality. Jpn J Clin Oncol 2015;45:7–11. https://doi.org/10.1093/jjco/hyu188.Search in Google Scholar PubMed
67. Sugano, K. Screening of gastric cancer in Asia. Best Pract Res Clin Gastroenterol 2015;29:895–905. https://doi.org/10.1016/j.bpg.2015.09.013.Search in Google Scholar PubMed
68. El Hajjar, A, Rey, J-F. Artificial intelligence in gastrointestinal endoscopy: general overview. Chinese Med J 2020;133:326–34. https://doi.org/10.1097/cm9.0000000000000623.Search in Google Scholar PubMed PubMed Central
69. Hirasawa, T, Aoyama, K, Tanimoto, T, Ishihara, S, Shichijo, S, Ozawa, T, et al.. Application of artificial intelligence using a convolutional neural network for detecting gastric cancer in endoscopic images. Gastric Cancer 2018;21:653–60. https://doi.org/10.1007/s10120-018-0793-2.Search in Google Scholar PubMed
70. Cheng, C, Cheng, L, Xu, R. Classification of FTIR gastric cancer data using wavelets and SVM. In: Proceedings third international conference on natural computation 2007, 1:543–7 pp.10.1109/ICNC.2007.299Search in Google Scholar
71. Shibata, T, Teramoto, A, Yamada, H, Ohmiya, N, Saito, K, Fujita, H. Automated detection and segmentation of early gastric cancer from endoscopic images using Mask R-CNN. Appl Sci 2020;10:3842. https://doi.org/10.3390/app10113842.Search in Google Scholar
72. Ikenoyama, Y, Hirasawa, T, Ishioka, M, Namikawa, K, Yoshimizu, S, Horiuchi, Y, et al.. Detecting early gastric cancer: comparison between the diagnostic ability of convolutional neural networks and endoscopists. Dig Endosc 2021;33:141–50. https://doi.org/10.1111/den.13688.Search in Google Scholar PubMed PubMed Central
73. Miyaki, R, Yoshida, S, Tanaka, S, Kominami, Y, Sanomura, Y, Matsuo, T, et al.. A computer system to be used with laser-based endoscopy for quantitative diagnosis of early gastric cancer. J Clin Gastroenterol 2015;49:108–15. https://doi.org/10.1097/mcg.0000000000000104.Search in Google Scholar PubMed
74. Kanesaka, T, Lee, T-C, Uedo, N, Lin, K-P, Chen, H-Z, Lee, J-Y, et al.. Computer-aided diagnosis for identifying and delineating early gastric cancers in magnifying narrow-band imaging. Gastrointest Endosc 2018;87:1339–44. https://doi.org/10.1016/j.gie.2017.11.029.Search in Google Scholar PubMed
75. Podder, P, Bharati, S, Mondal, MRH. Artificial intelligence for data-driven medical diagnosis. In: Deepak, G, Utku, K, Bao Le, N, Siddhartha, B, editors. 10 Automated gastric cancer detection and classification using machine learning. De Gruyter; 2021:207–24 pp.10.1515/9783110668322-010Search in Google Scholar
76. Nagao, S, Tsuji, Y, Sakaguchi, Y, Takahashi, Y, Minatsuki, C, Niimi, K, et al.. Highly accurate artificial intelligence systems to predict the invasion depth of gastric cancer: efficacy of conventional white-light imaging, nonmagnifying narrow-band imaging, and indigo-carmine dye contrast imaging. Gastrointest Endosc 2020;92:866–73. https://doi.org/10.1016/j.gie.2020.06.047.Search in Google Scholar PubMed
77. Japanese Gastric Cancer, A. Japanese gastric cancer treatment guidelines 2014 (ver. 4). Gastric Cancer 2017;20:1–19. https://doi.org/10.1007/s10120-016-0622-4.Search in Google Scholar PubMed PubMed Central
78. Choi, J, Kim, SG, Im, JP, Kim, JS, Jung, HC, Song, IS. Comparison of endoscopic ultrasonography and conventional endoscopy for prediction of depth of tumor invasion in early gastric cancer. Endoscopy 2010;42:705–13. https://doi.org/10.1055/s-0030-1255617.Search in Google Scholar PubMed
79. Tsujii, Y, Kato, M, Inoue, T, Yoshii, S, Nagai, K, Fujinaga, T, et al.. Integrated diagnostic strategy for the invasion depth of early gastric cancer by conventional endoscopy and EUS. Gastrointest Endosc 2015;82:452–9. https://doi.org/10.1016/j.gie.2015.01.022.Search in Google Scholar PubMed
80. Sano, T, Okuyama, Y, Kobori, O, Shimizu, T, Morioka, Y. Early gastric cancer. Endoscopic diagnosis of depth of invasion. Dig Dis Sci 1990;35:1340–4. https://doi.org/10.1007/bf01536738.Search in Google Scholar
81. Wu, J, Chen, J, Cai, J. Application of artificial intelligence in gastrointestinal endoscopy. J Clin Gastroenterol 2021;55:110–20. https://doi.org/10.1097/mcg.0000000000001423.Search in Google Scholar
82. Zhu, Y, Wang, QC, Xu, MD, Zhang, Z, Cheng, J, Zhong, YS, et al.. Application of convolutional neural network in the diagnosis of the invasion depth of gastric cancer based on conventional endoscopy. Gastrointest Endosc 2019;89:806–15. https://doi.org/10.1016/j.gie.2018.11.011.Search in Google Scholar PubMed
83. Yoon, HJ, Kim, S, Kim, J-H, Keum, J-S, Oh, S-I, Jo, J, et al.. A lesion-based convolutional neural network improves endoscopic detection and depth prediction of early gastric cancer. J Clin Med 2019;8:1310. https://doi.org/10.3390/jcm8091310.Search in Google Scholar PubMed PubMed Central
84. Kubota, K, Kuroda, J, Yoshida, M, Ohta, K, Kitajima, M. Medical image analysis: computer-aided diagnosis of gastric cancer invasion on endoscopic images. Surg Endosc 2012;26:1485–9. https://doi.org/10.1007/s00464-011-2036-z.Search in Google Scholar PubMed
85. Cho, B-J, Bang, CS, Lee, JJ, Seo, CW, Kim, JH. Prediction of submucosal invasion for gastric neoplasms in endoscopic images using deep-learning. J Clin Med 2020;9:1858. https://doi.org/10.3390/jcm9061858.Search in Google Scholar PubMed PubMed Central
86. Li, C, Shi, C, Zhang, H, Hui, C, Lam, KM, Zhang, S. Computer-aided diagnosis for preoperative invasion depth of gastric cancer with dual-energy spectral CT imaging. Acad Radiol 2015;22:149–57. https://doi.org/10.1016/j.acra.2014.08.006.Search in Google Scholar PubMed
87. Jiang, Y, Xie, J, Han, Z, Liu, W, Xi, S, Huang, L, et al.. Immunomarker support vector machine classifier for prediction of gastric cancer survival and adjuvant chemotherapeutic benefit. Clin Cancer Res 2018;24:5574–84. https://doi.org/10.1158/1078-0432.ccr-18-0848.Search in Google Scholar PubMed
88. Li, C, Zhang, S, Zhang, H, Pang, L, Lam, K, Hui, C, et al.. Using the K-nearest neighbor algorithm for the classification of lymph node metastasis in gastric cancer. Comput Math Methods Med 2012;2012:876545. https://doi.org/10.1155/2012/876545.Search in Google Scholar PubMed PubMed Central
89. Li, C, Shi, C, Zhang, H, Chen, Y, Zhang, S. Multiple instance learning for computer aided detection and diagnosis of gastric cancer with dual-energy CT imaging. J Biomed Inf 2015;57:358–68. https://doi.org/10.1016/j.jbi.2015.08.017.Search in Google Scholar PubMed
90. Tsukamoto, T, Toyoda, T, Mizoshita, T, Tatematsu, M. Helicobacter pylori infection and gastric carcinogenesis in rodent models. Semin Immunopathol 2013;35:177–90. https://doi.org/10.1007/s00281-012-0357-1.Search in Google Scholar PubMed
91. Correa, P, Houghton, J. Carcinogenesis of Helicobacter pylori. Gastroenterology 2007;133:659–72. https://doi.org/10.1053/j.gastro.2007.06.026.Search in Google Scholar PubMed
92. Take, S, Mizuno, M, Ishiki, K, Hamada, F, Yoshida, T, Yokota, K, et al.. Seventeen-year effects of eradicating Helicobacter pylori on the prevention of gastric cancer in patients with peptic ulcer; a prospective cohort study. J Gastroenterol 2015;50:638–44. https://doi.org/10.1007/s00535-014-1004-5.Search in Google Scholar PubMed
93. Shichijo, S, Hirata, Y, Sakitani, K, Yamamoto, S, Serizawa, T, Niikura, R, et al.. Distribution of intestinal metaplasia as a predictor of gastric cancer development. J Gastroenterol Hepatol 2015;30:1260–4. https://doi.org/10.1111/jgh.12946.Search in Google Scholar PubMed
94. Fukase, K, Kato, M, Kikuchi, S, Inoue, K, Uemura, N, Okamoto, S, et al.. Effect of eradication of Helicobacter pylori on incidence of metachronous gastric carcinoma after endoscopic resection of early gastric cancer: an open-label, randomised controlled trial. Lancet 2008;372:392–7. https://doi.org/10.1016/s0140-6736(08)61159-9.Search in Google Scholar
95. Ford, AC, Forman, D, Hunt, RH, Yuan, Y, Moayyedi, P. Helicobacter pylori eradication therapy to prevent gastric cancer in healthy asymptomatic infected individuals: systematic review and meta-analysis of randomised controlled trials. BMJ 2014;348:g3174. https://doi.org/10.1136/bmj.g3174.Search in Google Scholar PubMed PubMed Central
96. Sugano, K. Effect of Helicobacter pylori eradication on the incidence of gastric cancer: a systematic review and meta-analysis. Gastric Cancer 2019;22:435–45. https://doi.org/10.1007/s10120-018-0876-0.Search in Google Scholar PubMed
97. Choi, IJ, Kook, MC, Kim, YI, Cho, SJ, Lee, JY, Kim, CG, et al.. Helicobacter pylori therapy for the prevention of metachronous gastric cancer. N Engl J Med 2018;378:1085–95. https://doi.org/10.1056/nejmoa1708423.Search in Google Scholar
98. Wong, BC-Y, Lam, SK, Wong, WM, Chen, JS, Zheng, TT, Feng, RE, et al.. Helicobacter pylori eradication to prevent gastric cancer in a high-risk region of China: a randomized controlled trial. JAMA 2004;291:187–94. https://doi.org/10.1001/jama.291.2.187.Search in Google Scholar PubMed
99. Asaka, M, Mabe, K, Matsushima, R, Tsuda, M. Helicobacter pylori eradication to eliminate gastric cancer: the Japanese strategy. Gastroenterol Clin North Am 2015;44:639–48. https://doi.org/10.1016/j.gtc.2015.05.010.Search in Google Scholar PubMed
100. Suzuki, H, Mori, H. World trends for H. pylori eradication therapy and gastric cancer prevention strategy by H. pylori test-and-treat. J Gastroenterol 2018;53:354–61. https://doi.org/10.1007/s00535-017-1407-1.Search in Google Scholar PubMed PubMed Central
101. Asaka, M, Kato, M, Takahashi, S-i., Fukuda, Y, Sugiyama, T, Ota, H, et al.. Guidelines for the management of Helicobacter pylori infection in Japan: 2009 revised edition. Helicobacter 2010;15:1–20. https://doi.org/10.1111/j.1523-5378.2009.00738.x.Search in Google Scholar PubMed
102. de Vries, AC, Kuipers, EJ, Rauws, EA. Helicobacter pylori eradication and gastric cancer: when is the horse out of the barn? Am J Gastroenterol 2009;104:1342–5. https://doi.org/10.1038/ajg.2008.15.Search in Google Scholar PubMed
103. Shichijo, S, Hirata, Y, Niikura, R, Hayakawa, Y, Yamada, A, Ushiku, T, et al.. Histologic intestinal metaplasia and endoscopic atrophy are predictors of gastric cancer development after Helicobacter pylori eradication. Gastrointest Endosc 2016;84:618–24. https://doi.org/10.1016/j.gie.2016.03.791.Search in Google Scholar PubMed
104. Shichijo, S, Nomura, S, Aoyama, K, Nishikawa, Y, Miura, M, Shinagawa, T, et al.. Application of convolutional neural networks in the diagnosis of Helicobacter pylori infection based on endoscopic images. EBioMedicine 2017;25:106–11. https://doi.org/10.1016/j.ebiom.2017.10.014.Search in Google Scholar PubMed PubMed Central
105. Shichijo, S, Endo, Y, Aoyama, K, Takeuchi, Y, Ozawa, T, Takiyama, H, et al.. Application of convolutional neural networks for evaluating Helicobacter pylori infection status on the basis of endoscopic images. Scand J Gastroenterol 2019;54:158–63. https://doi.org/10.1080/00365521.2019.1577486.Search in Google Scholar PubMed
106. Mohan, BP, Khan, SR, Kassab, LL, Ponnada, S, Mohy-Ud-Din, N, Chandan, S, et al.. Convolutional neural networks in the computer-aided diagnosis of Helicobacter pylori infection and non-causal comparison to physician endoscopists: a systematic review with meta-analysis. Ann Gastroenterol 2021;34:20–5. https://doi.org/10.20524/aog.2020.0542.Search in Google Scholar PubMed PubMed Central
107. Zheng, W, Zhang, X, Kim, JJ, Zhu, X, Ye, G, Ye, B, et al.. High accuracy of convolutional neural network for evaluation of Helicobacter pylori infection based on endoscopic images: preliminary experience. Clin Transl Gastroenterol 2019;10:e00109. https://doi.org/10.14309/ctg.0000000000000109.Search in Google Scholar PubMed PubMed Central
108. Itoh, T, Kawahira, H, Nakashima, H, Yata, N. Deep learning analyzes Helicobacter pylori infection by upper gastrointestinal endoscopy images. Endosc Int Open 2018;6:E139–44. https://doi.org/10.1055/s-0043-120830.Search in Google Scholar PubMed PubMed Central
109. Ishihara, K, Ogawa, T, Haseyama, M. Helicobacter pylori infection detection from multiple x-ray images based on combination use of support vector machine and multiple kernel learning. In: Proceedings 2015 IEEE international conference on image processing 2015:4728–32 pp.10.1109/ICIP.2015.7351704Search in Google Scholar
110. Huang, CR, Chung, PC, Sheu, BS, Kuo, HJ, Popper, M. Helicobacter pylori-related gastric histology classification using support-vector-machine-based feature selection. IEEE Trans Inf Technol Biomed 2008;12:523–31. https://doi.org/10.1109/titb.2007.913128.Search in Google Scholar PubMed
111. Ishihara, K, Ogawa, T, Haseyama, M. Helicobacter pylori infection detection from multiple X-ray images based on decision level fusion. In: Proceedings 2014 IEEE international conference on image processing 2014:2769–73 pp.10.1109/ICIP.2014.7025560Search in Google Scholar
112. Ishihara, K, Ogawa, T, Haseyama, M. Helicobacter Pylori infection detection from gastric X-ray images based on feature fusion and decision fusion. Comput Biol Med 2017;84:69–78. https://doi.org/10.1016/j.compbiomed.2017.03.007.Search in Google Scholar PubMed
113. Grantcharov, TP, Bardram, L, Funch-Jensen, P, Rosenberg, J. Assessment of technical surgical skills. Eur J Surg 2002;168:139–44. https://doi.org/10.1080/110241502320127739.Search in Google Scholar PubMed
114. Andras, I, Mazzone, E, van Leeuwen, FWB, De Naeyer, G, van Oosterom, MN, Beato, S, et al.. Artificial intelligence and robotics: a combination that is changing the operating room. World J Urol 2020;38:2359–66. https://doi.org/10.1007/s00345-019-03037-6.Search in Google Scholar PubMed
115. Wang, Z, Majewicz Fey, A. Deep learning with convolutional neural network for objective skill evaluation in robot-assisted surgery. Int J Comput Assist Radiol Surg 2018;13:1959–70. https://doi.org/10.1007/s11548-018-1860-1.Search in Google Scholar PubMed
116. Ershad, M, Rege, R, Majewicz Fey, A. Automatic and near real-time stylistic behavior assessment in robotic surgery. Int J Comput Assist Radiol Surg 2019;14:635–43. https://doi.org/10.1007/s11548-019-01920-6.Search in Google Scholar PubMed
117. Jin, P, Ji, X, Kang, W, Li, Y, Liu, H, Ma, F, et al.. Artificial intelligence in gastric cancer: a systematic review. J Cancer Res Clin Oncol 2020;146:2339–50. https://doi.org/10.1007/s00432-020-03304-9.Search in Google Scholar PubMed
118. Gao, Y, Zhang, ZD, Li, S, Guo, YT, Wu, QY, Liu, SH, et al.. Deep neural network-assisted computed tomography diagnosis of metastatic lymph nodes from gastric cancer. Chinese Med J 2019;132:2804–11. https://doi.org/10.1097/cm9.0000000000000532.Search in Google Scholar
119. Matsumoto, T, Niioka, H, Kumamoto, Y, Sato, J, Inamori, O, Nakao, R, et al.. Deep-UV excitation fluorescence microscopy for detection of lymph node metastasis using deep neural network. Sci Rep 2009;9:16912. https://doi.org/10.1038/s41598-019-53405-w.Search in Google Scholar PubMed PubMed Central
120. Hensler, K, Waschulzik, T, Mönig, SP, Maruyama, K, Hölscher, AH, Bollschweiler, E. Quality-assured efficient engineering of feedforward neural networks (QUEEN). Methods Inf Med 2005;44:647–54. https://doi.org/10.1055/s-0038-1634021.Search in Google Scholar
121. Su, F, Sun, Y, Hu, Y, Yuan, P, Wang, X, Wang, Q, et al.. Development and validation of a deep learning system for ascites cytopathology interpretation. Gastric Cancer 2020;23:1041–50. https://doi.org/10.1007/s10120-020-01093-1.Search in Google Scholar PubMed
122. Li, J, Dong, D, Fang, M, Wang, R, Tian, J, Li, H, et al.. Dual-energy CT–based deep learning radiomics can improve lymph node metastasis risk prediction for gastric cancer. Eur Radiol 2020;30:2324–33. https://doi.org/10.1007/s00330-019-06621-x.Search in Google Scholar PubMed
123. Biglarian, A, Hajizadeh, E, Kazemnejad, A, Zali, M. Application of artificial neural network in predicting the survival rate of gastric cancer patients. Iran J Public Health 2011;40:80–6.Search in Google Scholar
124. Li, J, Zhang, C, Wei, J, Zheng, P, Zhang, H, Xie, Y, et al.. Intratumoral and peritumoral radiomics of contrast-enhanced CT for prediction of disease-free survival and chemotherapy response in stage II/III gastric cancer. SSRN Electron J 2020;10:552270. https://doi.org/10.3389/fonc.2020.552270.Search in Google Scholar PubMed PubMed Central
125. He, J, Baxter, SL, Xu, J, Xu, J, Zhou, X, Zhang, K. The practical implementation of artificial intelligence technologies in medicine. Nat Med 2019;25:30–6. https://doi.org/10.1038/s41591-018-0307-0.Search in Google Scholar PubMed PubMed Central
126. Jin, S, Wang, B, Zhu, Y, Dai, W, Xu, P, Yang, C, et al.. Log odds could better predict survival in muscle-invasive bladder cancer patients compared with pN and lymph node ratio. J Cancer 2019;10:249–56. https://doi.org/10.7150/jca.27399.Search in Google Scholar PubMed PubMed Central
127. You, R, Liu, YP, Lin, M, Huang, P-Y, Tang, LQ, Zhang, YN, et al.. Relationship of circulating tumor cells and Epstein–Barr virus DNA to progression-free survival and overall survival in metastatic nasopharyngeal carcinoma patients. Int J Cancer 2019;145:2873–83. https://doi.org/10.1002/ijc.32380.Search in Google Scholar PubMed
128. Haag, GM, Czink, E, Ahadova, A, Schmidt, T, Sisic, L, Blank, S, et al.. Prognostic significance of microsatellite-instability in gastric and gastroesophageal junction cancer patients undergoing neoadjuvant chemotherapy. Int J Cancer 2019;144:1697–703. https://doi.org/10.1002/ijc.32030.Search in Google Scholar PubMed
129. Alabi, RO, Elmusrati, M, Sawazaki-Calone, I, Kowalski, LP, Haglund, C, Coletta, RD, et al.. Machine learning application for prediction of locoregional recurrences in early oral tongue cancer: a Web-based prognostic tool. Virchows Arch 2019;475:489–97. https://doi.org/10.1007/s00428-019-02642-5.Search in Google Scholar PubMed PubMed Central
130. Pedersen, MH, Hood, BL, Ehmsen, S, Beck, HC, Conrads, TP, Bak, M, et al.. CYPOR is a novel and independent prognostic biomarker of recurrence-free survival in triple-negative breast cancer patients. Int J Cancer 2019;144:631–40. https://doi.org/10.1002/ijc.31798.Search in Google Scholar PubMed
131. Chartrand, G, Cheng, PM, Vorontsov, E, Drozdzal, M, Turcotte, S, Pal, CJ, et al.. Deep learning: a primer for radiologists. Deep Learning 2017;37:2113–31. https://doi.org/10.1148/rg.2017170077.Search in Google Scholar PubMed
132. Lan, K, Wang, DT, Fong, S, Liu, LS, Wong, KKL, Dey, N. A survey of data mining and deep learning in bioinformatics. J Med Syst 2018;42:139. https://doi.org/10.1007/s10916-018-1003-9.Search in Google Scholar PubMed
133. Thurtle, DR, Greenberg, DC, Lee, LS, Huang, HH, Pharoah, PD, Gnanapragasam, VJ. Individual prognosis at diagnosis in nonmetastatic prostate cancer: development and external validation of the PREDICT Prostate multivariable model. PLoS Med 2019;16:e1002758. https://doi.org/10.1371/journal.pmed.1002758.Search in Google Scholar PubMed PubMed Central
134. Chen, JH, Asch, SM. Machine learning and prediction in medicine–beyond the peak of inflated expectations. N Engl J Med 2017;376:2507–9. https://doi.org/10.1056/nejmp1702071.Search in Google Scholar
135. Cabitza, F, Rasoini, R, Gensini, GF. Unintended consequences of machine learning in medicine. JAMA 2017;318:517–8. https://doi.org/10.1001/jama.2017.7797.Search in Google Scholar PubMed
© 2023 the author(s), published by De Gruyter, Berlin/Boston
This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
