中文摘要：

恶性肿瘤严重危害人类健康，早期诊断和有效治疗可降低致死率。借助在体和无创等优势，现代生物医学影像技术能为肿瘤诊疗提供大量有价值的结构和功能信息。由于肿瘤是一种多因素的发展型疾病，并且每种成像模态也都存在自身缺陷，单一时间点、单一模态的医学影像检查，无法揭示其本质、发生和发展过程。不同模态医学图像信息融合，可以实现信息互补和交叉验证，是揭示肿瘤机理、实现早期准确诊断和有效治疗的必然选择。本文在分析各模态成像技术特点的基础上，从配准、多模态医学图像融合方法和肿瘤诊疗中融合应用等3个方面分别进行综述，并指出了未来需要研究的问题。

英文摘要：

Malignant tumors are threatening human health seriously. Early diagnosis and effective treatment can reduce the mortality rate. Taking advantage of in-vivo and non-invasive measurements, modern biomedical imaging is able to provide tumor diagnosis and treatment with massive valuable information on both structure and function. However, cancer is a multi-factorial and development disorder, meanwhile each modality of imaging owns its drawback and limitations. Therefore, it is hardly to elucidate the tumor＇s intrinsic mechanism and developing process via a single-modality image acquisition. Multi-modal medical image fusion, which can realize the information complement and cross-validation, is considered to be the promising choice. In this review, after analyzing the characteristics of each imaging technique, the state-of-the-art of registration, image fusion methods and their clinic applications are reviewed, and then the available problems and potential future direction are discussed.