糖尿病肾病无创诊断探索研究

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.

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。

糖尿病肾病目前是尿毒症的主要原发病之一，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，建立临床随访数据库；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，并结合肾活检，对糖尿病肾病辅助诊断系统的准确性进行检验和评估，该模型预测准确率为96.2%，并进一步挖掘糖尿病肾病进展的危险因素，发现血糖变异性，血脂异常，促甲状腺激素(TSH) 升高，三碘甲状腺氨酸（T3）和游离三碘甲状腺原氨酸(FT3)降低，肾小球毛细血管IgM和C1q沉积，是糖尿病肾病进展的危险因素，为糖尿病肾病预警和早期诊断提供支持。本研究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。在此课题期间，共发表SCI论文12篇，medline收录论文1篇，中文论文12篇；主编/副主编专著4部，参编专著2部；获得授权发明专利1项及软件著作权4项；作为主要完成人获中国中西医结合学会科学技术一等奖和河南省医学科学技术进步一等奖各1项；培养研究生3名；晋升副主任医师/副主任技师各1人；项目负责人被评为“首都十大杰出青年医生”，课题组成员被评为“解放军总医院“3+1”新秀人才扶持对象”。