作者：张秋盛，林祺，李广华，林昭宇，王昭

单位：中山大学附属第一医院 胃肠外科中心，广东 广州 510080

Authors: Zhang Qiusheng, Lin Qi, Li Guanghua, Lin Zhaoyu, Wang Zhao

Unit:  Gastrointestinal Surgery Unit, the First Affiliated Hospital of Sun Yat-sen University, Guangzhou 510080, Guangdong, China

摘要：

近年来人工智能（artificial intelligence, AI）技术发展迅速，是目前各大行业的研究热点，尤其是在医学方面，其中深度学习已广泛应用于医学图像处理和临床肿瘤学分析。根据10年来的研究结果，AI与医学结合能够极大提升医生的工作效率，减轻工作负担，对疾病诊断和分级有更高的敏感度和准确度，对指导手术治疗和提高患者生存率具有重要临床意义。在影像组学方面，AI可以从标准医疗成像中提取并分析图像的向量特征，用于胃癌的淋巴结转移和早期复发情况的预测。同样的，在数字病理学方面，AI也可以通过计算机扫描标准的苏木精-伊红染色组织切片，进行数字化处理，进而运用卷积神经网络来处理图像特征，用于胃癌的病理诊断和治疗前的疗效预测等。内镜技术和AI结合近年来也表现出良好的前景，AI能从内镜视频流和静止图片抓取病变特征和确定病变部位。总而言之，目前在肿瘤领域，AI技术可以协助医生更准确地进行诊断，减少人为误差；也可以判断患者是否能从某种治疗中获益，减轻治疗负担。AI的出现对于医生和患者来说都是一种福音。

关键词： 人工智能；深度学习；胃癌；诊断；预测

Abstract：

In recent years, the rapid advancement of artificial intelligence (AI) technologies has taken center stage across various industries, particularly in the field of medicine where deep learning has been extensively applied to medical image processing and oncological clinical analysis. Studies over the past decade indicate that the integration of AI with medicine significantly boosts physicians' work efficiency and alleviates their workload. It also provides higher sensitivity and accuracy in disease diagnosis and staging, which plays a critical role in guiding surgical interventions and improving patients' survival rates. In radiomics, AI algorithms are capable of extracting and analyzing vector features from standard medical imaging modalities, which are instrumental in predicting lymph node metastasis and early recurrence of gastric cancer. Similarly, in digital pathology, AI can digitize and process standard hematoxylin-eosin stained tissue sections scanned by computers, employing convolutional neural networks to handle image features for pathologic diagnosis of gastric cancer and prediction of therapeutic efficacy pre -treatment. Moreover, the synergy between endoscopic technology and AI has shown promising prospects recently, with AI's ability to capture pathological characteristics and determine lesion locations from endoscopic video streams and still images. In summary, AI technology aids oncologists in making more accurate diagnoses, reducing human error, and assessing whether patients will benefit from certain treatments, thereby reducing treatment burdens. The emergence of AI represents a boon for both clinicians and patients, enhancing the quality of healthcare delivery and paving the way for more precise and personalized treatment approaches.

Key Words:  Artificial intelligence; Deep learning; Gastric cancer; Diagnosis; Forecast

注：网络优先发布