【摘要】 目的 研究肾穿刺活检病理切片远程诊断的正确率和影响因素以探讨其实施标准。方法 选择40例已有病理诊断的肾穿刺活检标本的病理切片，肾脏病医师和病理医师分别对20例进行选样。通过3种方法获取数字化病理图像，随后将其从浙江青田传输至上海，由另一位病理医师根据这些图像和提供的病史进行诊断(采用盲法)并对画面质量进行评估和对影响诊断正确率的因素进行分析。结果 40 例中28例诊断完全正确，2例基本正确，4例有差异正确，4例诊断错误，2例延迟诊断。当场诊断的38例病例的严格诊断正确率为73.7%，而临床重要性诊断正确率为89.5%。由不同医师选样的两组的诊断正确率相等(严格诊断正确率为73.7%，临床重要诊断正确率为89.5%)。不常见疾病的诊断正确率明显低于常见疾病(前后两者的严格诊断正确率分别为14.3%、87.1%，临床重要诊断正确率分别为42.9%、100%，P<0.001)。病例图像中包含的肾小球数目与诊断结果(完全正确诊断3分、基本正确诊断2分、差异性正确诊断1分、错误诊断0分)呈正相关(P<0.05)。结论 本研究的远程病理诊断正确率基本符合临床要求，肾穿刺活检病理切片远程病理诊断方法易行，普通方法采集图像、压缩、传输、显示和诊断方式不影响诊断正确率，而通过提高肾脏病理的肾小球采样数可以提高诊断正确率，据此提出其实施标准。

     【关键词】 远程病理 ； 肾穿刺活检

    Investigation of inplementing standard: telepathology of renal needle biopsy

    JIANG Gui-hua， JIANG Jin-gen， MAO Jing-fang， et al.Shanghai Sixth People’s Hospital,Shanghai 200233， China

    【Abstract】 Objective Study the accuracy and its relative facts to evaluate the feasibility of telepathology (TP) diagnosis for renal needle biopsy. Methods 40 cases of renal needle biopsy are included based on the microscopy results. A nephrology doctor and a pathology doctor selected 20 cases each. Digital images are obtained from 3 paths. Then digital images are transmitted from Qingtian, Zhejiang to Shanghai. Another pathology doctor makes diagnosis blindly for each case based on the images and the clinical histories, also evaluating the quality of the pathological images, and the relative facts are discussed on technologic and clinical facets. Results 28 diagnoses of those of the 40 cases are absolutely correct, 2 essentially correct, 4 partial correct, 4 wrong and 2 delayed. Of the 38 cases diagnosed on the spot, the strict accuracy and clinically important one are 73.7% and 89.5% respectively. The diagnostic accuracy of the 2 groups whose images are selected by different doctors is identical (strict accuracy: 73.7% and clinically important one: 89.5%). The accuracy of uncommon diseases is obviously lower than that of common diseases (strict one is 14.3%、87.1% respectively and clinically important one is 42.9%、100% respectively, P<0.001). The numbers of glomeruli selected is positively correlative with the digitized results of diagnosis (if diagnosis is absolutely correct it gets 3, essentially correct gets 2, partial correct gets 1 and wrong gets 0, P<0.05). Conclusion Methods of TP of renal needle biopsy is feasible and easy. The accuracy of TP diagnosis in our trial can essentially meet the clinical demand. General methods to capture, compress, transmit, display images and to diagnose won’t influence the diagnostic accuracy. Selecting more renal glomeruli may improve the diagnostic accuracy. ......