糖尿病肾病无创诊断探索研究

项目名称： 糖尿病肾病无创诊断探索研究

项目编号： No.61471399

项目类型： 面上项目

立项/批准年度： 2015

项目学科： 无线电电子学、电信技术

项目作者： 朱晗玉

作者单位： 中国人民解放军总医院

项目金额： 90万元

中文摘要： 与糖尿病合并非糖尿病性肾脏疾病相比，糖尿病肾病的病程进展快、治疗原则保守、预后差。肾功能不全患者一般会在5-10年发生尿毒症。肾活检是诊断糖尿病肾病的金标准，但由于其费用高、有创且技术要求高而难以常规运用于临床。因此，我们必须研发一种可靠、易行且无创的临床鉴别诊断方法。本研究拟利用解放军总医院拥有大样本及多中心糖尿病肾病患者研究队列的优势，大规模有计划地采集糖尿病合并肾脏损害随访队列的临床数据；在前期研究工作的基础上，构建无创性诊断模型和临床鉴别诊断模型，并建立基于多吸引子元胞自动机的糖尿病肾病无创辅助诊断原型系统；结合肾活检，对糖尿病肾病辅助诊断系统的准确性进行检验和评估，并进一步挖掘糖尿病肾病进展的危险因素。本研究旨在为糖尿病肾病的临床判断提供可量化的指标, 这将有利于提高疾病诊断的科学性和减少医疗风险，具有重要的临床价值。

中文关键词： 生物信息处理；医学信息检测；元胞自动机；糖尿病肾病；无创诊断

英文摘要： In China, diabetic nephropathy is the primary cause of uremic dialysis patients. In comparison with non-diabetic renal disease, diabetic nephropathy has more rapid progression, more conservative treatment and poorer prognosis. Uremia generally occurs within 5-10 years, depending on how serious of renal insufficiency of the diabetic nephropathy patients. Renal biopsy is the gold standard for diagnosis of diabetic nephropathy. However, it is poorly applied in clinic because it is an invasive method with high cost and technical requirements. Therefore, we have to develop a reliable, easy and noninvasive method to distinguish NDRD from DN. In our previous pilot study, we established a preliminary probability regression equation to identify NDRD and DN, but the results were unsatisfactory. More scientific and effective modeling is needed for clinical applications. In this study, we will take advantage of our multi-center and large diabetic nephropathy patient cohorts, and collect the clinical data in a large-scare for the follow-up cohorts and the diabetes patients who are also associated with proteinuria. Through this multi-center study, combined with our previously well-established database, we will analyze and find out the characteristic parameters for diabetic nephropathy diagnosis and set up a non-invasive diagnostic model and a multiple attractor cellular automata assisted system, whose accuracy will be evaluated with renal biopsy inspection; we will investigate the risk factors associated with diabetic nephropathy progress so as to further demonstrate the necessity of developing a non-invasive diagnostic model. Our study is aimed to provide a quantifiable indicator for clinical judgment, which is critical for us to improve the disease diagnosis and reduce the medical risk in clinic.

英文关键词： Biological information processing;Medical information detection;Cellular automata;Diabetic nephropathy;Noninvasive diagnosis