食管癌转移性淋巴结的综合影像特征分析及计量诊断模型的建立

The incidence of esophageal cancer in China accounted for more than 50% of the world, and with a high mortality rate. Accurate staging is an important prerequisite for a reasonable choice of treatment options, improve the efficacy and control prognosis. Imaging examinations are the most effective means to diagnosis and staging of esophageal cancer. However, Diagnosis of lymph node metastasis based on the lymph nodes length of shorter diameter is often used currently. Large inflammatory lymph nodes can not be discerned, and metastasis lymph nodes with a smaller diameter is difficult to determine has been plagued our major clinical problem. This study will collect lymph node size, shape, edge, density, enhancement, signal intensity, acoustic image, FDG uptake value, location, number of imaging characteristics. Analysis and understanding of the connotation of the image, to be valid recognition and utilization. As well as, combined with the patient's gender, age, symptoms, tumor location, length, depth and scope of invasion, pathological type, degree of differentiation of esophageal cancer related factors. Use Bayes's discriminant methods, synthetic data of the large sample size to analysis of various indicators to determine the role and value of the lymph node metastasis. Establish a quantitative diagnosis model for clinical application, forecast the probability of lymph node metastasis. To achieve a more scientific and accurate tumor staging, choose the best treatment options and improve the therapeutic effect.

我国是食管癌大国，发病率占世界总数的50％以上，死亡率高居不下。疗前准确分期、选择合理治疗方案是提高疗效、控制预后的重要前提。影像学检查是食管癌诊断和分期最为有效的手段。但目前对淋巴结转移的诊断多以淋巴结径线大小为依据，而炎性肿大淋巴结不能鉴别，径线较小的转移淋巴结难以判定，一直是困扰临床的主要问题。本研究将提取食管癌淋巴结的大小、形态、边缘、密度、强化、信号、回声、放射性摄取值、部位、数量等多种影像学检查的可视信息，分析各种图像特征，科学地理解其内容含义，加以有效的识别利用；同时，结合患者性别、年龄、症状、肿瘤部位、长度、侵犯深度、范围、病理类型、分化程度等肿瘤相关因素，综合较大样本量的数据，利用Bayes方法判别各项指标的作用及价值，建立一个适合临床应用的诊断模型。通过多项指标的综合判断，预报食管癌淋巴结发生转移的概率，达到更加科学，准确地进行肿瘤分期的目的。

食管癌是我国高发恶性肿瘤，死亡率居高不下，而准确分期是提高疗效的重要前提，影像学检查则是诊断和分期最为有效的手段。但目前对食管癌淋巴结转移的诊断多以淋巴结径线大小为依据，而炎性肿大淋巴结不能鉴别、径线较小的转移淋巴结难以判定，一直是困扰着临床的主要问题。. 本研究经提取食管癌淋巴结的大小、形态、边缘、密度、强化、信号、回声、放射性摄取值、部位、数量等多种影像学检查的可视信息，分析和理解各种图像特征的内在含义，加以有效的识别利用；同时，结合肿瘤部位、侵犯深度、病理类型、等临床相关因素，利用Bayes方法判别各项指标的作用及价值，建立一个适合临床应用的诊断模型。通过多项指标的综合判断，预报食管癌淋巴结发生转移的概率，达到更加科学，准确地进行肿瘤分期的目的。. 经过4年的努力，我们按计划完成了食管癌CT、MRI、EUS、PET-CT各种影像学检查的多项可视性参数的采集、测量工作，结合相关临床信息，分别建立了数据库。通过对各自图像特征的分类、筛选，相关指标的整理、汇总，综合利用这些数据的统计、分析，我们初步建立了食管癌淋巴结转移CT判别诊断模型；探讨了适合临床应用的食管癌淋巴结转移CT诊断的新标准；尝试了MRI对食管癌淋巴结转移的量化诊断。同时，对EUS的诊断参数进行了筛选和优化，改进了传统内镜淋巴结鉴别诊断的多因素组合指标；验证了食管癌淋巴结PET-CT摄取值的各项相关参数。结果表明：多项指标组合应用是判断食管癌淋巴结转移的有效方法，较单一指标更有助于提高诊断的准确性。. 截止目前为止，共撰写论文6篇，发表3篇（国内2篇、国外1篇）；培养研究生5名，毕业3名（硕士2名、博士1名）；编写完成了“食管癌淋巴结转移判别分析软件”（内部试用中）；正在申请专利1项、登记软件著作权1项。