胰腺癌是常见的致死性恶性肿瘤，而淋巴结转移是胰腺癌的独立预后因素。术前准确评估胰腺癌的淋巴结转移情况对于预测预后及制定合适的治疗方案具有重要临床意义。既往用于预测胰腺癌淋巴结转移的血清学标志物准确度不高，而常规影像学技术主要依靠淋巴结大小来评估转移情况，均具有一定局限性。血清胶原蛋白片段作为胰腺癌间质生成和细胞外基质重构的潜在标志物，对其进行定量分析有望用于预后评价；而基于MRI扩散加权图像的影像组学分析可提取出反映肿瘤生物学行为特点的图像信息。本项目拟应用影像组学方法对胰腺癌患者的术前MRI扩散加权图像进行分析，以期找出能精准预测胰腺癌淋巴结转移的影像组学特征，并结合临床信息、血清胶原蛋白片段等血清学标志物、常规影像特点等构建出评估淋巴结转移情况的联合预测模型。本项目有望为临床医生制定恰当的治疗策略提供参考，从而提高患者生存率、改善患者预后，具有重要的临床实用价值。

Pancreatic cancer is a common malignancy with high morbidity and mortality. Lymph node metastasis is an independent prognostic factor. Accurate preoperative evaluation of lymph node metastasis in pancreatic cancer is of great clinical significance in predicting prognosis and making appropriate treatment plan. Previous serological markers and conventional imaging methods have drawbacks in predicting lymph node metastasis of pancreatic cancer. As a potential marker of desmoplasia and extracellular matrix remodeling in pancreatic cancer, serum collagen fragments is expected to be used for prognosis evaluation. Radiomics based on diffusion weighted images can extract the image information that reflect the biological behavior of tumor. In this project, the preoperative diffusion-weighted MRI images of pancreatic cancer patients will be analyzed by radiomics process to find out the radiomics features that can accurately predict the lymph node metastasis of pancreatic cancer. Then these radiomics features will be integrated with clinical data, serological markers such as collagen fragments and conventional image characteristics for creating a combined prediction model. This project can be helpful for clinicians to design appropriate treatment strategies, so as to improve the survival rate and prognosis.