人工神经网络预测阿德福韦酯单用治疗拉米夫定耐药慢性乙型肝炎患者的疗效  

Prediction of viral resistance in Lamivudine-resistant chronic hepatitis B patients who receive Adefovir monotherapy by artificial neural network

作　　者：赵攀[1] 王浩[2] 王春亚[3] 李文淑[1] 王琦[4] 

机构地区：[1]解放军302医院临床试验中心肝衰竭诊疗与研究中心,北京100039 [2]解放军302医院医学信息中心,北京100039 [3]首都医科大学附属北京安贞医院急诊科 [4]军事医学科学院统计学教研室

出　　处：《胃肠病学和肝病学杂志》2015年第5期542-543,共2页Chinese Journal of Gastroenterology and Hepatology

摘　　要：目的研究影响拉米夫定(LAM)耐药慢性乙型肝炎患者换用阿德福韦酯(ADV)疗效的因素。方法纳入2007年9月-2012年1月于解放军302医院采用LAM首治并且因出现耐药而换用ADV继续治疗的HBe Ag(+)慢性乙型肝炎患者,采用人工神经网络(ANN)方法对资料进行分析。结果预测模型显示,年龄、LAM用药时间和HBe Ag定量可用于预测病毒耐药的发生,模型的敏感度和特异度分别为90.91%和90.74%。结论 ANN预测模型有较好的敏感度和特异度,有助于病毒耐药的监测和管理。Objective To study the impact factors of Adefovir （ADV） monotherapy in Lamivudine （LAM）-resistant patients with chronic hepatitis B. Methods Clinical data of chronic HBV-infected HBeAg （ ＋ ） patients who had de- veloped antiviral drug resistance under de novo LAM monotherapy and subsequently took ADV monotherapy as rescue strategy were collected and the data were analyzed by the method of artificial neural network （ ANN ）. Results Age, duration of LAM and HBeAg quantity were associated with the development of drug resistance in the prediction model, which had sensitivity of 90.91% and specificity of 90.74%. Conclusion The ANN prediction model has superior sen- sitivity and specificity, which is helpful in the management of viral resistance.

关 键 词：拉米夫定 阿德福韦酯 慢性乙型肝炎 人工神经网络 

分 类 号：R512.62[医药卫生—内科学]