区域淋巴结转移是影响直肠癌临床治疗决策和预后的主要因素之一。目前临床上区域淋巴结转移评价主要依靠影像医生根据淋巴结大小、形状、内部信号等征象进行主观判断，诊断准确率不高、诊断时间较长并受限于评价者自身经验。新辅助治疗引起的纤维化反应进一步增加了评估区域淋巴结状态的难度。MRI功能成像技术及定量参数虽然可以提高区域淋巴结转移诊断的敏感性，但并不能提高总体诊断准确率。深度学习方法给直肠癌区域淋巴结诊断的医学影像智能分析带来了新的思路。目前尚缺乏深度学习在直肠癌区域淋巴结区域分割或 转移诊断的应用研究报道。本研究拟应用深度学习方法和多模态磁共振参数构建直肠癌区域淋巴结转移的辅助诊断模型，研究将开展前瞻性研究实现MRI图像标记淋巴结与病理检出淋巴结的一一对应，提取与淋巴结转移直接相关的特征，构建高灵敏度、高特异度的智能化、可视化计算机辅助工具，协助影像医生提高直肠癌区域淋巴结诊断准确率。

Regional lymph node metastasis is one of the key factors that affect the clinical treatment decision and prognosis of rectal cancer. At present, the clinical evaluation of regional lymph node metastasis mainly depends on the subjective judgment of radiologists according to the size, shape and internal signal of lymph nodes. The existing problems were low diagnostic accuracy, long diagnostic time and limited experience of the radiologists. Fibrosis induced by neoadjuvant therapy further increases the difficulty of assessing regional lymph node status. Functional MRI and its quantitative parameters can improve the sensitivity of assessing regional lymph node metastasis, but can not improve the overall diagnostic accuracy. Deep learning method brings a new idea to the intelligent analysis of medical imaging for regional lymph node diagnosis of rectal cancer. At present, there is no report on the application of deep learning in regional lymph node segmentation or metastasis diagnosis of rectal cancer. In this study, we propose to construct an auxiliary diagnostic model for regional lymph node metastasis of rectal cancer using deep learning method and multimodal magnetic resonance parameters. We will carry out a prospective study to achieve a one-to-one correspondence between MRI-labeled lymph nodes and pathologically detected lymph nodes, extract features directly related to lymph node metastasis, and construct a highly sensitive and high-quality model. Intelligent and visualized computer aided tools with both high sensitivity and specificity will help radiologists to improve the accuracy of regional lymph node diagnosis of rectal cancer.