对肝硬化相关结节的准确定性分类及在治疗前预测肝细胞癌（hepatocellular carcinoma, HCC）的肿瘤侵袭性、预后与治疗反应对于实现HCC的早筛、早诊和个体化治疗至关重要。目前基于西方国家肝脏疾病数据制定、全球应用较广泛的肝脏影像报告及数据系统（Liver Reporting And Data System, LI-RADS）在我国以及亚洲地区的HCC临床诊疗实践中存在评价指标缺乏量化、诊断效能不足、难以揭示肿瘤生物学行为和评估预后等突出问题。因此本项目拟利用定量功能磁共振成像技术，对照组织病理与分子病理结果，对2018版LI-RADS进行细化分级和预后风险分析，并采用深度学习方法，以期建立适应国情的、以准确、高效、稳定为特征的量化LI-RADS（quantitative LI-RADS, qLI-RADS）智能分类系统，从而为HCC的个性化精准诊疗提供重要影像学支撑。

Accurate categorization and characterization of cirrhotic liver nodules and pretreatment prediction of tumor aggressiveness, prognosis and treatment response is critical for early detection, diagnosis and individualized treatment of hepatocellular carcinoma (HCC). Based on liver disease data of Western countries, Liver Reporting and Data System (LI-RADS) has gain worldwide attention in recent years. However, the application of LI-RADS in China and other Asian countries is restricted by its unquantified parameters, unsatisfactory diagnostic accuracy and limited capacity in revealing tumor biology and prognosis. Therefore, based on quantitative functional magnetic resonance imaging techniques, this study aims to sub-classify the LI-RADS v2018 categories according to refined diagnostic and prognostic risk stratification with histopathologic and molecular pathology analyses as the reference standard. Moreover, we also aim to develop an accurate, efficient and robust quantitative LI-RADS (qLI-RADS) intelligent categorization system with deep learning techniques to provide a concrete imaging evaluation platform for individualized HCC diagnosis and management in China.