基于多组学融合的智能诊断在肝细胞癌微血管侵犯的应用研究

微血管侵犯（MVI）是影响肝细胞癌（HCC）术后肝外转移的独立危险因素，但目前其诊断主要依靠手术标本的病理检查，对于手术方式的选择和治疗方案具有明显滞后性。本项目拟运用单细胞测序技术筛选出一组与MVI相关的候选基因，研究其作为药物靶点的可能性，继而基于影像信号-病理特征-药靶分子三者的关联性研究HCC患者术前多模态影像与差异分子表达的相关性，通过计算机辅助CAD智能系统建立HCC患者术前预测MVI的影像学模型，实现药物靶点活性状态的可视化。项目拟进一步采用多中心研究在1000余例HCC患者的影像学资料、临床病理随访资料、基因表达等三个层面联合验证该模型的正确性和可行性，从而为HCC患者术前病情评估、手术方式选择、监测预后及抗HCC药物选择提供更方便和安全的途径。

Microvascular invasion (MVI) is a common independent risk factor for the treatment and prognosis of hepatocellular carcinoma (HCC). However, its diagnosis relies mainly on the pathological examination of surgical specimens, which has a significant lag for the choice of surgical methods and treatment options. This project intends to screen out a group of candidate genes related to MVI by using single cell sequencing technology , to explore its possibility as a drug target, and then based on the correlation of image signal-pathological features-drug target molecules, to study the correlation between HCC patients and pre-multimodal images. Finally, a preoperative predictive MVI imaging model will be established for HCC patients to visualize the active state of drug targets by computer-aided CAD intelligence system. The project aims to further use the multi-center study to verify the correctness and feasibility of the model in the imaging data, clinical pathological follow-up data and gene expression of more than 1000 patients with HCC, so as to evaluate the preoperative condition of HCC patients and provide a more convenient and safe route for surgical options, prognosis, and anti-HCC drug selection.