应用数学优选法对组织切片中细胞核DNA定量分析  

Application of mathematical optimal method in DNA quantification on tissue slices

作　　者：朱荣[1] 唐峰 胡锡琪[1] 徐元鼎[1] ZHU Rong;TANG Feng;HU Xi-qi;XU Yuan-ding(Department of Pathology,Shanghai Medical College,Fudan University,Shanghai 200032,China;Department of Surgery Pathology,Huashan Hospital,Fudan University,Shanghai 200040,China)

机构地区：[1]复旦大学上海医学院病理学系,上海200032 [2]复旦大学附属华山医院病理科,上海200040

出　　处：《复旦学报（医学版）》2019年第6期793-796,813,共5页Fudan University Journal of Medical Sciences

摘　　要：目的以数学优选法为基础,建立一种全新的组织切片细胞核DNA的定量分析方法,应用于良恶性肿瘤的组织学鉴别诊断。方法以42例恶性肿瘤和7例非肿瘤性良性病变为研究材料,两位资深病理医师光镜观察并做出病理诊断,通过Feulgen染色进行DNA定量分析。在组织切片上通过大数据统计分析处理技术,即数学中的优选原则结合一定的筛选法则,对切片中的细胞先做处理和优选,再进行分析。结果采用优选法的DNA定量分析系统诊断灵敏度为97.62%,特异度为100.00%,阳性预测值为100.00%,阴性预测值为87.50%,正确指数(Youden指数)为0.98,假阳性率为0,假阴性率为2.38%,准确度为97.96%。结论优选法不但获得理论上的成功,通过临床验证同样获得了满意的效果,在人工智能应用于传统组织病理诊断方面具有很大潜力。Objective Based on mathematical optimal method,we tried to set up an innovative DNA quantification analysis on tissue slices,which can be used in histopathological differential diagnoses.Methods Two senior pathologists microscopically made the diagnoses for 42 cases of malignant tumors and 7 cases of non-tumor benign lesions.Meanwhile,the Feulgen stain was performed for DNA quantification.We treated the data by mathematical optimal method combined with some filter rules before the analysis to optimize the cells in the tissue.Results In our DNA quantification system based on mathematical optimization,the diagnostic sensitivity was 97.62%,and specificity was 100.00%.The positive predictive value was 100.00%,and negative predictive value was 87.50%.The false positive rate was 0,and false negative rate was 2.38%.Corrective index(Youden index)was 0.98,and the accuracy was 97.96%.Conclusions Our optimal method is not only successful in theory,but also satisfied in clinic test.This method would contribute the traditional histopathological diagnosis by the application of artificial intelligence.

关 键 词：DNA定量分析 优选法 细胞核 组织切片 

分 类 号：R331[医药卫生—人体生理学]