基于深度学习技术的日本血吸虫抗体检测结果智能判读模型的建立和效能评价

doi:10.20199/j.issn.1672-2302.2024.06.007

摘要/Abstract

摘要：

目的建立一种基于深度学习技术的日本血吸虫抗体间接血凝试验（indirect hemagglutination assay, IHA）检测结果的智能判读模型，并评价其判读效能，实现血吸虫病IHA检测结果判读自动化和智能化。方法制备、收集日本血吸虫阳性及阴性家兔（人）血清，制作不同凝集程度的IHA检测结果图片，并对凝集结果进行评定;采集IHA反应板上无反应孔的图片，对图片进行图像增强处理，建立图片数据集。基于卷积神经网络PP-LCNet建立智能判读模型，对模型进行训练及测试。IHA检测经不同比例（1∶5~1∶100）稀释的日本血吸虫“金标准”阳性家兔血清及“金标准”阴性家兔血清，由专业技术人员和智能判读模型分别判定结果，计算二者的特异性、敏感性、准确率、F1值、约登指数和诊断一致性（Kappa检验），绘制受试者工作特征（receiver operation characteristic, ROC）曲线，比较二者判读结果的准确性。结果共获得IHA反应图片15 956张，其中阴性5 878张，弱阳性2 164张，阳性2 271张，无反应孔图片5 643张。经过图像增强处理后，图片集合计31 856张，其中训练集25 487张，测试集6 369张。效能评价结果显示，除特异性和专业技术人员一致外（均为100.00%），模型的准确率、敏感性、F1值、约登指数、ROC曲线下面积分别为99.43%、99.09%（97.84%，100.34%）、99.54%、0.990 9、0.995±0.009，均略高于专业技术人员。智能判读模型的敏感性和专业技术人员A[95.45%（92.70%，98.20%）]和B[93.18%（89.85%, 96.51%）]之间的差异有统计学意义（χ²=6.125、11.077，P均<0.05），和专业技术人员C[98.64%（95.31%, 97.17%）]之间的差异无统计学意义（χ²=0.000，P>0.05）。智能模型的判定结果和“金标准”保持较高的一致性（Kappa=0.988，P<0.05），略高于3位专业技术人员（Kappa=0.940、0.910、0.982，P均<0.05）。结论基于深度学习技术建立的智能判读模型识别日本血吸虫抗体IHA检测结果的准确性很高，实际工作中可用于日本血吸虫病IHA检测结果的智能判读。

关键词: 日本血吸虫, 间接血凝试验, 深度学习, 图像识别, 智能判读

Abstract:

Objective To develop an intelligent interpretation model for the results of indirect hemagglutination assay（IHA）of Schistosoma japonicum antibody in sera based on deep learning algorithms, and to preliminarily evaluate its efficacy so as to realize automatically and intelligently translating the results of schistosomiasis by diagnostic IHA. Methods Prepare and collect serum samples from rabbits (humans) with positive and negative results of Schistosoma japonicum, create IHA test result images with different levels of agglutination, and evaluate the agglutination results. The images with microtiter tray absent of IHA reaction were collected, and treated by enhancement, on which basis the image data sets were constituted. Then the intelligent interpretation model was created based on convolutional neural network PP-LCNet algorithms, and trained and tested respectively by training set and test set. The S. japonicum antibody gold-standard negative and positive rabbit sera diluted in different ratio (1∶5-1∶100) were detected by IHA, and the results were interpreted by three well-experienced professionals and the intelligent translating model independently. The performances of the model and professionals were evaluated as the specificity, sensitivity, accuracy, F1 score, diagnostic consistency (Kappa value). The receiver operation characteristic (ROC) curves were plotted to compare the coherence by the model and the professionals. Results In total, 15 956 images quantified by IHA were obtained, in which 2 271 were positive, 2 164 weak positive and 5 878 were negative reactions. Reaction was absent in 5 643 images of the microtiter trays. After the image enhancement, a total of 31 856 image sets were obtained, and divided training set (n=25 487) and test set (n=6 369). Efficacy evaluation showed that, except for the consistency in specificity (The agreement was 100.00% by the model and experts), the accuracy, sensitivity, F1 scores, Youden index and areas under ROC curves were 99.43%, 99.09%(97.84%, 100.34%), 99.54%, 0.990 9 and 0.995±0.009, respectively by the model results, which were slightly higher than the findings by professionals. The sensitivity of the intelligent interpretation model showed a statistically significant difference (χ²=6.125, 11.077, both P<0.05) between professional technicians A [95.45% (92.70%, 98.20%)] and B [93.18% (89.85%, 96.51%)], while there was no statistically significant difference (χ²=0.000, P>0.05) between professional technicians C [98.64% (95.31%, 97.17%)]. The results generated by the intelligent interpretation model highly agreed with gold standard. The consistency was somewhat better than the three professionals (Kappa= 0.988 vs. Kappa= 0.940, 0.910, 0.982, respectively, all P<0.05). Conclusion The intelligent interpretation model based on deep learning technology established in this study can accurately identify the results of S. japonicum antibody detection in sera by IHA, suggesting that it can be used for intelligently translating the IHA results for schistosomiasis japonica.

Key words: Schistosoma japonicum, Indirect Hemagglutination Assay, Deep learning, Image Recognition, Intelligent interpretation

中图分类号:

R383.2+4

马晓荷, 章乐生, 汪峰峰, 路标, 孙成松, 李清越, 王旗, 操治国, 汪天平. 基于深度学习技术的日本血吸虫抗体检测结果智能判读模型的建立和效能评价[J]. 热带病与寄生虫学, 2024, 22(6): 358-363.

MA Xiaohe, ZHANG Lesheng, WANG Fengfeng, LU Biao, SUN Chengsong, LI Qingyue, WANG Qi, CAO Zhiguo, WANG Tianping. Establishment and efficacy evaluation of a deep learning-based intelligent interpretation model for IHA detection results of Schistosoma japonicum antibody[J]. Journal of Tropical Diseases and Parasitology, 2024, 22(6): 358-363.

图/表 5

图1

图2

图3

表1

图4

参考文献 34

[1]	He JX, Baxter SL, Xu J, et al. The practical implementation of artificial intelligence technologies in medicine[J]. Nat Med, 2019, 25(1):30-36. doi: 10.1038/s41591-018-0307-0 pmid: 30617336
[2]	徐自远. 面向人工智能算法下图像识别技术分析[J]. 数字技术与应用, 2021, 39(10):4-6.
[3]	Tomašev N, Glorot X, Rae JW, et al. A clinically applicable approach to continuous prediction of future acute kidney injury[J]. Nature, 2019, 572(7767):116-119.
[4]	Topol EJ. High-performance medicine: the convergence of human and artificial intelligence[J]. Nat Med, 2019, 2(1):44-56.
[5]	Yasaka K, Akai H, Abe O, et al. Deep learning with convolutional neural network for differentiation of liver masses at dynamic contrast-enhanced CT: a preliminary study[J]. Radiology, 2018, 286(3):887-896. doi: 10.1148/radiol.2017170706 pmid: 29059036
[6]	Wei JY, Song XY, Wei XE, et al. Knowledge-augmented deep learning for segmenting and detecting cerebral aneurysms with CT angiography: a multicenter study[J]. Radiology, 2024, 312(2):e233197.
[7]	Xi HB, Wang WJ. Deep learning based uterine fibroid detection in ultrasound images[J]. BMC Med Imaging, 2024, 24(1):218. doi: 10.1186/s12880-024-01389-z pmid: 39160500
[8]	Hong Z, Li L, Zhang LJ, et al. Elimination of schistosomiasis Japonica in China: from the one health perspective[J]. China CDC Wkly, 2022, 4(7):130-134.
[9]	马晓荷, 汪敏, 朱磊, 等. 安徽省不同地区日本血吸虫群体线粒体基因遗传变异研究[J]. 热带病与寄生虫学, 2021, 19(5):254-258.
[10]	操治国. 我国血吸虫病防治的进展、挑战与对策[J]. 热带病与寄生虫学, 2022, 20(3):130-135.
[11]	张利娟, 何君逸, 杨帆, 等. 2023年全国血吸虫病防治进展[J]. 中国血吸虫病防治杂志, 2024, 36(3):221-227.
[12]	Ross AG, Bartley PB, Sleigh AC, et al. Schistosomiasis[J]. N Engl J Med, 2002, 346(16):1212-1220.
[13]	周帅锋, 余路新, 汪世平. 血吸虫病的诊断检测技术及研究进展[J]. 热带医学杂志, 2009, 9(3):335-340.
[14]	张世清, 章乐生, 汪峰峰, 等. 《日本血吸虫抗体检测间接红细胞凝集试验》(WS/T 630—2018)标准解读[J]. 热带病与寄生虫学, 2020, 18(1):1-4.
[15]	国家卫生标准委员会寄生虫病标准专业委员会.日本血吸虫抗体检测间接红细胞凝集试验:WS/T 630—2018[S/OL]. (2018-09-26)[2024-05-27]. http://www.nhc.gov.cn/wjw/s9499/201810/6b5bb415063b48c1b54fbdc36b01c406/files/14136386fda24e428f54b8f1b41fd104.pdf.
[16]	Cui C, Gao TQ, Wei SY, et al. PP-LCNet: a lightweight CPU convolutional neural network[EB/OL]. [2024-05-28]. https://arxiv.org/abs/2109.15099v1.
[17]	张利娟, 何君逸, 杨帆, 等. 2023年全国血吸虫病防治进展[J]. 中国血吸虫病防治杂志, 2024, 36(3):221-227.
[18]	代波, 汪天平, 许晓娟, 等. 2020年洪涝灾害后安徽省钉螺扩散情况调查[J]. 热带病与寄生虫学, 2022, 20(4):191-196.
[19]	谢婧姿, 李宜锋, 袁敏, 等. 2020年洪涝灾害后江西省钉螺扩散情况调查[J]. 热带病与寄生虫学, 2022, 20(4):197-201.
[20]	王晖, 单晓伟, 钟晨晖, 等. 2020年洪涝灾害后湖北省钉螺扩散情况调查[J]. 热带病与寄生虫学, 2022, 20(4):202-205,209.
[21]	汤凌, 程湘晖, 连花, 等. 2020年洪涝灾害后湖南省钉螺扩散情况调查[J]. 热带病与寄生虫学, 2022, 20(4):206-209.
[22]	Bergquist R, Zhou XN, Rollinson D, et al. Elimination of schistosomiasis: the tools required[J]. Infect Dis Poverty, 2017, 6(1):158. doi: 10.1186/s40249-017-0370-7 pmid: 29151362
[23]	汪天平, 吕山, 秦志强, 等. 共享WHO指南努力实现我国消除血吸虫病目标[J]. 中国血吸虫病防治杂志, 2022, 34(3):235-240.
[24]	王恩木, 章乐生, 张世清, 等. 日本血吸虫抗体检测试剂盒(间接血凝法)应用效果评价[J]. 热带病与寄生虫学, 2009, 7(4):203-205.
[25]	孙成松, 汪峰峰, 王玥, 等. 日本血吸虫抗体检测试剂盒(IHA法)现场筛查效果的评估[J]. 中国病原生物学杂志, 2013, 8(11):982-985.
[26]	何伟, 朱荫昌, 华万全, 等. 血吸虫病快速免疫诊断:胶体染料试纸条法的研究[J]. 中国血吸虫病防治杂志, 2000, 12(1):18-20.
[27]	丁建祖, 干小仙, 沈慧英, 等. 快速检测抗日本血吸虫抗体的金标免疫渗滤法的建立及应用[J]. 中国寄生虫病防治杂志, 1998(4):308-310.
[28]	周杰, 官威, 危芙蓉, 等. 间接血凝试验在日本血吸虫病诊断中的价值研究[J]. 中国血吸虫病防治杂志, 2016, 28(4):375-380.
[29]	刘阳, 吴子松, 张奕, 等. 2016—2017年四川省血吸虫病消除工作考核评估[J]. 预防医学情报杂志, 2018, 34(11):1355-1360.
[30]	Ni QQ, Sun ZY, Qi L, et al. A deep learning approach to characterize 2019 coronavirus disease (COVID-19) pneumonia in chest CT images[J]. Eur Radiol, 2020, 30(12):6517-6527.
[31]	Yang F, Poostchi M, Yu H, et al. Deep learning for smartphone-based malaria parasite detection in thick blood smears[J]. IEEE J Biomed Health Inform, 2020, 24(5):1427-1438.
[32]	施亮, 熊春蓉, 刘毛毛, 等. 基于深度学习技术的湖北钉螺视觉智能识别模型的建立[J]. 中国血吸虫病防治杂志, 2021, 33(5):445-451.
[33]	薛靖波, 夏尚, 李召军, 等. 基于无人机影像深度学习算法的血吸虫病家畜传染源智能识别研究[J]. 中国血吸虫病防治杂志, 2023, 35(2):121-127.
[34]	杨维中, 兰亚佳, 吕炜, 等. 建立我国传染病智慧化预警多点触发机制和多渠道监测预警机制[J]. 中华流行病学杂志, 2020, 41(11):1753-1757.

	敏感性（95%CI）（%）	特异性（%）	准确率（%）	F1值	约登指数	AUC（95%CI）值
专业技术人员A	95.45（92.70, 98.20）	100.00	97.14	96.47	0.954 5	0.977（0.956, 0.998）
专业技术人员B	93.18（89.85, 96.51）	100.00	95.71	95.24	0.931 8	0.966（0.941, 0.991）
专业技术人员C	98.64（95.31, 97.17）	100.00	99.14	99.32	0.986 4	0.993（0.982, 1.005）
智能判读模型	99.09（97.84, 100.34）	100.00	99.43	99.54	0.990 9	0.995（0.986, 1.005）