肾透明细胞癌的转移导致其治疗困难。影像组学为精准预测肾透明细胞癌转移提供了新方法。然而，现有影像组学研究尚未深入到生物学含义层面，医学影像背后的生物学信息还有待挖掘。本项目以解决肾透明细胞癌转移预测临床问题作为切入，继而开展了影像与基因表达特异性之间关联性的研究。首先通过选择具有潜在生物语义的影像特征，建立肾透明细胞癌转移影像预测模型，发现具有转移预测功能的影像标志物；在影像基因组学分析框架下构建影像标志物的加权基因共表达网络，采用通路分析的方法探究肾透明细胞癌转移特异性从微观基因到宏观影像的信号通路。旨在为肾透明细胞癌转移诊断和影像基因组学分析方法研究提供新方法和新思路。

Metastasis of clear cell renal carcinoma (ccRCC) makes treatment difficult. Radiomics provides a new method to accurately predict the metastasis of ccRCC. However, the current research has not yet penetrated into the biological implications of medical imaging. The biological information of medical image still needs to be excavated. The proposed project started from solving the clinical problem of predicting metastasis of ccRCC. Then, the correlation between image and gene specific expression was studied. Firstly, imaging features with potential biological semantics were extracted for establishing the prediction model of ccRCC metastasis. Then, image markers with predictive function were found. Under the framework of image and gene association analysis, an image marker weighted gene co-expression network was constructed. The signal pathway analysis method was used to explore the transition path of metastasis specificity from gene to image of ccRCC. The aims of the project are to provide new way for diagnosing the metastasis of ccRCC and new idea for radiogenomics analysis method.