重视人工智能在葡萄膜恶性黑色素瘤中的应用和挑战  

Focusing on the application and challenges of artificial intelligence in uveal malignant melanoma

作　　者：丁运刚 李永平[2] Ding Yungang;Li Yongping(Affiliated Eye Hospital of Shandong University of Traditional Chinese Medicine,Jinan 250002,China;Zhongshan Ophthalmic Center,Sun Yat-sen University,Guangzhou 510060,China)

机构地区：[1]山东中医药大学附属眼科医院,济南250002 [2]中山大学中山眼科中心,广州510060

出　　处：《中华实验眼科杂志》2024年第12期1084-1089,共6页Chinese Journal Of Experimental Ophthalmology

摘　　要：葡萄膜恶性黑色素瘤是成人常见的原发性眼内恶性肿瘤,具有高度隐匿性和转移性,具有高致盲和致死风险。随着机器学习和深度学习技术的发展,人工智能(AI)在葡萄膜恶性黑色素瘤的诊断、治疗和预后评估中展现出相当的应用潜力,能够深入挖掘临床影像、病理及基因组等多维数据,辅助临床医师进行诊断和治疗决策。AI可以分析眼部照相和放射学图像数据,辅助鉴别诊断;预测放射治疗不良反应和效果,辅助优化治疗方案;基于临床特征和数字病理图像,构建精准的预后预测模型,其准确率可以媲美基因表达谱分析。尽管AI在葡萄膜恶性黑色素瘤的临床应用面临数据、技术和人机协作等挑战,但是随着对葡萄膜恶性黑色素瘤研究的深入和AI技术的发展,AI有望更精准、高效地助力葡萄膜恶性黑色素瘤患者的诊疗和预后评估,最终改善患者预后。Uveal malignant melanoma is one of the common primary intraocular malignancies in adults.Its high concealment and significant metastatic potential lead to a high risk of blindness and mortality.With advances in machine learning and deep learning techniques,artificial intelligence(AI)has shown increasing promise for application in the diagnosis,management,and prognosis evaluation of uveal malignant melanoma.AI can thoroughly analyze the multi-modal data,such as clinical images,pathological images,and genetic data,and assist clinicians in diagnosis and treatment planning.AI analyzes ophthalmic photography and radiological image to assist in differential diagnosis,and predicts side effects and outcomes of radiotherapy to optimize treatments.AI constructs the models for accurate prognosis based on clinical features and digital pathology images,and its accuracy is comparable to that of gene expression profiling tests.The clinical application of AI in uveal malignant melanoma faces the challenges of data availability,technology limitations,and effective human-machine collaboration.However,with ongoing research in both uveal malignant melanoma and AI,AI is expected to improve the accuracy and efficiency of diagnosis,management,and prognosis assessment,ultimately improving patient outcomes.

关 键 词：葡萄膜 黑色素瘤 人工智能 机器学习 诊断 治疗 预后评估 

分 类 号：R739.7[医药卫生—肿瘤]