人工智能在胶囊内镜中的应用现状与前景北大核心CSCD

以深度学习中卷积神经网络模型为代表的人工智能(artificial intelligence,AI)技术为胶囊内镜(capsule endoscopy,CE)提供了一种高效而准确的自动图像识别方法,可辅助临床医生诊断,在多项研究中具有较好的应用效果。同时,人工智能在多病变分类诊断、胶囊内镜定位、内镜质量控制、胶囊内镜教学培训以及多场景应用等方面均具有较大的发展潜力。     

Clinicopathologic study of 208 patients with early gastric cancer in Taiwan: A comparison between Eastern and Western countries

Abstract To investigate whether there are any differences between the clinicopathologic characteristics of early gastric cancer (EGC) patients in Eastern and Western countries, 208 Taiwanese patients with EGC were reviewed between 1964 and 1992. The incidence of cancer has increased slightly over the 29-year period. Men were diagnosed with EGC frequently than women, and their mean age was 56 years. Epigastralgia (58.2%) was the most common symptom, whereas 5.8% of cancers were incidentally detected by endoscopy. Physical signs and laboratory tests were of limited value in making the diagnosis. Endoscopy was a better diagnostic aid than radiology. Tumours were frequently located in the lower third (53.2%) and middle third (43.3%) of the stomach. Cancers of the elevated type (17.8%) were less frequent than the depressed type (82.2%). Type II_c (31.2%) was the most common macroscopic type. The frequency of mucosal carcinoma (51.0%) was similar to submucosal carcinoma (49.0%). Mucosal carcinoma had less lymph node metastases (3.1%) than submucosal carcinoma (12.2%; P < 0.05), with an overall frequency of metastases of 7.5% (14/186). The 5-year survival rate was 90.8%. The clinicopathologic characteristics of EGC in Taiwan were similar to those of Western countries and other Eastern countries. Improvement of diagnostic examinations and endoscopic surveillance of asymptomatic subjects may lead to early diagnosis and thus ensure a more favourable outcome.     

Rib fracture detection in computed tomography images using deep convolutional neural networks

To evaluate the rib fracture detection performance in computed tomography (CT) images using a software based on a deep convolutional neural network (DCNN) and compare it with the rib fracture diagnostic performance of doctors.We included CT images from 39 patients with thoracic injuries who underwent CT scans. In these images, 256 rib fractures were detected by two radiologists. This result was defined as the gold standard. The performances of rib fracture detection by the software and two interns were compared via the McNemar test and the jackknife alternative free-response receiver operating characteristic (JAFROC) analysis.The sensitivity of the DCNN software was significantly higher than those of both Intern A (0.645 vs 0.313; P < .001) and Intern B (0.645 vs 0.258; P < .001). Based on the JAFROC analysis, the differences in the figure-of-merits between the results obtained via the DCNN software and those by Interns A and B were 0.057 (95% confidence interval: −0.081, 0.195) and 0.071 (−0.082, 0.224), respectively. As the non-inferiority margin was set to −0.10, the DCNN software is non-inferior to the rib fracture detection performed by both interns.In the detection of rib fractures, detection by the DCNN software could be an alternative to the interpretation performed by doctors who do not have intensive training experience in image interpretation.     

Application of convolutional neural networks for evaluating Helicobacter pylori infection status on the basis of endoscopic images

Background and aim: We recently reported the role of artificial intelligence in the diagnosis of Helicobacter pylori (H. pylori) gastritis on the basis of endoscopic images. However, that study included only H. pylori-positive and -negative patients, excluding patients after H. pylori-eradication. In this study, we constructed a convolutional neural network (CNN) and evaluated its ability to ascertain all H. pylori infection statuses.

Methods: A deep CNN was pre-trained and fine-tuned on a dataset of 98,564 endoscopic images from 5236 patients (742 H. pylori-positive, 3649 -negative, and 845 -eradicated). A separate test data set (23,699 images from 847 patients; 70 positive, 493 negative, and 284 eradicated) was evaluated by the CNN.

Results: The trained CNN outputs a continuous number between 0 and 1 as the probability index for H. pylori infection status per image (Pp, H. pylori-positive; Pn, negative; Pe, eradicated). The most probable (largest number) of the three infectious statuses was selected as the ‘CNN diagnosis’. Among 23,699 images, the CNN diagnosed 418 images as positive, 23,034 as negative, and 247 as eradicated. Because of the large number of H. pylori negative findings, the probability of H. pylori-negative was artificially re-defined as Pn ?0.9, after which 80% (465/582) of negative diagnoses were accurate, 84% (147/174) eradicated, and 48% (44/91) positive. The time needed to diagnose 23,699 images was 261?seconds.

Conclusion: We used a novel algorithm to construct a CNN for diagnosing H. pylori infection status on the basis of endoscopic images very quickly.

Abbreviations: H. pylori: Helicobacter pylori; CNN: convolutional neural network; AI: artificial intelligence; EGD: esophagogastroduodenoscopies.     

增强放大内镜对早期胃癌的诊断价值

目的评价冰醋酸．靛胭酯增强的放大内镜（ME-AIM）和窄带成像增强的放大内镜（ME—NBI）对早期胃癌的诊断价值。方法对普通白光内镜（WLE）依据黏膜形态或色泽发生异常改变筛查出的317处局灶性病变先后行ME—NBI和ME-AIM检查,以活检病理学诊断结果为金标准,统计分析各种检查方法诊断早期胃癌的准确性。结果对于早期胃癌的诊断,ME—NBI的敏感度、特异度和准确率分别为100．0％（11／11）、99．7％（305／-306）和99．7％（316／317）,ME-AIM分别为90．9％（10／11）、99．0％（303／306）”和98．7％（313／317）,WLE分别为72．7％（8／11）、88．2％（270／306）和87．7％（278／317）。ME—NBI和ME．AIM诊断早期胃癌的特异度和准确率显著高于wLE,差异均有统计学意义（P〈0．05）;尽管ME-NBI‘和ME=AIM诊断早期胃癌的敏感度也高于WLE,但差异无统计学意义（P〉0．05）。而ME-NBI与ME—AIM在诊断早期胃癌的敏感度、特异度和准确率方面差异均无统计学意义（P〉0．05）。结论wLE仍然是早期胃癌筛查首选的检查方法。ME-NBI和ME-AIM均能提高诊断的准确性,且两者诊断价值相似。     

Diagnostic outcomes of esophageal cancer by artificial intelligence using convolutional neural networks

1. Download high-res image (405KB)
2. Download full-size image

     

Magnifying endoscopy simple diagnostic algorithm for early gastric cancer (MESDA‐G)

Gastric cancer is the third leading cause of cancer death worldwide. Early detection and accurate diagnosis of mucosal cancer is desirable in order to achieve decreased mortality; cause‐specific survival of patients with early gastric cancer is reported to exceed 95%. Endoscopy is the functional modality to detect early cancer; however, the procedure is not definitive when using conventional white‐light imaging. In contrast, magnifying narrow‐band imaging (M‐NBI), a novel endoscopic technology, is a powerful tool for characterizing gastric mucosal lesions because it can visualize the microvascular architecture and microsurface structure. To date, many reports on the diagnosis of early gastric cancer by M‐NBI, including multicenter prospective randomized studies conducted in Japan, have been published in peer‐reviewed international journals. Based on these published data, we devised a proposal for a diagnostic strategy for gastric mucosal cancer using M‐NBI to simplify the process of diagnosis and improve accuracy. Herein, we recommend a diagnostic algorithm for early gastric cancer using magnifying endoscopy.     

提高早期胃癌诊断水平的几项关键点

长期以来，胃癌一直是我国发病率和病死率最高的恶性肿瘤。患者就诊晚、临床总体诊断水平欠佳等造成众多病人确诊时已属进展期，且常伴有胃周、腹腔甚至远处转移。自从证实仅浸润黏膜或黏膜下层的早期胃癌预后良好以来，30多年间我国从事胃肠病、外科学和肿瘤学的专业人员努力进取，在早期胃癌的诊断方面开展了大量研究，但因种种原因早期胃癌的诊断率依然很低，早期胃癌手术率〈10％，这与日本的早期胃癌占50％以上相比差之甚远。据统计，日本1975年早期胃癌占所有接受治疗胃癌病例的20．9％，1990年迅速升至43．4％。15年间日本TNM分期中T2的发生率无明显变化，而T4的发生率从22．3％下降至11．1％。2004年以来在日本早期胃癌检诊协会所属医疗机构中，检出的胃癌中超过70％为早期胃癌，约50％的早期胃癌患者接受了内镜下黏膜切除术（EMR）或黏膜下剥离术（ESD）。如此高的早期胃癌检出率得益于对无症状的日本人群进行胃癌筛选。     

色素内镜诊断早期胃癌组织和癌旁组织基因表达差异

目的探讨早期胃癌组织及癌旁组织ZO-1和ZO-2表达差异。方法将活检的早期胃癌组织和癌旁组织液氮冷冻保存后,分别进行RT-PCR、Western blot来观察ZO-1和ZO-2表达。结果ZO-1和ZO-2的表达广泛存在于人类胃癌和癌旁组织中,ZO-1和ZO-2在人类胃癌组织的表达量均较癌旁组织有显著下降。结论ZO-1和ZO-2可能与人类胃癌的发生和转移有密切的关系。     

Optical diagnosis of colorectal polyps using convolutional neural networks

Colonoscopy remains the gold standard investigation for colorectal cancer screening as it offers the opportunity to both detect and resect pre-malignant and neoplastic polyps. Although technologies for image-enhanced endoscopy are widely available, optical diagnosis has not been incorporated into routine clinical practice, mainly due to significant inter-operator variability. In recent years, there has been a growing number of studies demonstrating the potential of convolutional neural networks (CNN) to enhance optical diagnosis of polyps. Data suggest that the use of CNNs might mitigate the inter-operator variability amongst endoscopists, potentially enabling a “resect and discard“ or ”leave in“ strategy to be adopted in real-time. This would have significant financial benefits for healthcare systems, avoid unnecessary polypectomies of non-neoplastic polyps and improve the efficiency of colonoscopy. Here, we review advances in CNN for the optical diagnosis of colorectal polyps, current limitations and future directions.     

Study on minute surface structures of the depressed‐type early gastric cancer with magnifying endoscopy

Background: Gastric surface patterns and morphology of minute surface vessels in depressed lesions were analyzed using a magnifying endoscope with high resolving power to contribute to qualitative diagnosis of gastric cancer. Methods: Subjects were diagnosed with depressed‐type early gastric cancer (pT1), there were 63 lesions, 38 differentiated‐type lesions, and 25 undifferentiated‐type lesions. There were also 40 benign depressed lesions found. After routine observations with an endoscope, amplifying observations of lesions were made by EG‐410CR (Fuji Photo Optical; Saitama, Japan) (CR). The images were compared with macroscopic patterns and histopathological patterns of the surgical specimens and endoscopic mucosal resection specimens. Results: Surface patterns of gastric depressed lesions were classified as irregular protrusion, normal papilla, pseudopapilla and amorphia. Irregular protrusion was found only in cancerous lesions. Characteristic minute vessels were observed in amorphia. Their patterns were classified into the following six types: sand, fence, round net, flat net, branch and coil. Irregular protrusion and minute vessels in amorphia (round net, flat net, branch and coil) were specific to cancers. There was a tendency for round net and flat net patterns to be found often in differentiated cancers and for branch and coil patterns to be found often in undifferentiated cancers. Conclusion: This magnifying endoscopic classification is considered useful for the qualitative diagnosis of depressed‐type early gastric cancer.