| 本文已被：浏览 110次 下载 137次 |
码上扫一扫！ |
| 采用遗传-反向传播人工神经网络法构建新疆地区癫痫患儿拉考沙胺血药浓度预测模型 |
| 赵婷1,孙岩2,李红健1,张惠兰1,于静2,冯杰1,王婷婷1,于鲁海1 |
|
|
| ((1. 新疆维吾尔自治区人民医院,新疆维吾尔自治区临床药学研究所,乌鲁木齐 830001;2. 北京儿童医院新疆医院,新疆维吾尔自治区儿童医院,乌鲁木齐 830001)) |
|
| 摘要: |
| 目的:利用遗传-反向传播(GA-BP)人工神经网络法构建新疆地区癫痫患儿拉考沙胺( LCM) 血药浓度的预测模型。方
法:采用超高效液相色谱法测定400 例癫痫患儿的LCM 稳态血药浓度,收集患儿临床资料,提取相关数据,采用GA-BP 人工神
经网络法构建LCM 血药浓度的预测模型。结果:模型验证结果显示,80 例预测浓度的平均预测误差(MPE)绝对值均<10%,预
测误差(PE)绝对值<20%的比例是100%,PE 绝对值<10%的比例是92. 50%,平均预测绝对误差(MAE)为2. 28%,提示GA-BP
模型预测的准确度和精密度均较好,预测浓度和实测浓度的相关系数为0. 998,预测结果较理想。结论:应用GA-BP 人工神经
网络法预测LCM 血药浓度是可行的,可应用于LCM 个体化给药研究,促进临床合理用药。 |
| 关键词: 癫痫 拉考沙胺 血药浓度 遗传-反向传播人工神经网络 |
| DOI：doi:10.13407/j.cnki.jpp.1672-108X.2024.04.002 |
|
| 基金项目:新疆维吾尔自治区自然科学基金面上项目,编号2016D01C097。 |
|
| Prediction Model of Lacosamide Plasma Concentration in Children with Epilepsy in Xinjiang by Genetic Algorithm-Back Propagation Artificial Neural Network Method |
| Zhao Ting1, Sun Yan2, Li Hongjian1, Zhang Huilan1, Yu Jing2, Feng Jie1, Wang Tingting1, Yu Luhai1 |
| ((1. People’s
Hospital of Xinjiang Uygur Autonomous Region, Institute of Clinical Pharmacy of Xinjiang Uygur Autonomous Region, Urumqi 830001, China;
2. Xinjiang Hospital of Beijing Children’s Hospital, Children’s Hospital of Xinjiang Uygur Autonomous Region, Urumqi 830001, China)) |
| Abstract: |
| Objective: To establish the prediction model of lacosamide (LCM) plasma concentration in children with epilepsy in
Xinjiang by genetic algorithm-back propagation (GA-BP) artificial neural network method. Methods: Steady-state plasma concentration
of LCM in 400 children with epilepsy was determined by ultra-high performance liquid chromatography. Clinical data of children were
collected and extracted. GA-BP artificial neural network method was used to construct the prediction model of LCM plasma
concentration. Results: Model validation results showed that the mean prediction error (MPE) absolute value of predicted concentration
in 80 cases was <10%, the proportion of the absolute value of prediction error (PE) <20% was 100%, the proportion of the absolute
value of PE<10% was 92. 50%, and the mean absolute error (MAE) of prediction was 2. 28%, suggesting that the GA-BP model
predicted with good accuracy and precision. The correlation coefficient between the predicted concentration and actual measured
concentration was 0. 998, and the predicted results were satisfactory. Conclusion: GA-BP artificial neural network method is feasible to
predict the plasma concentration of LCM, and can be applied to the study of individualized drug administration of LCM to promote
rational drug use in clinic. |
| Key words: epilepsy lacosamide plasma concentration genetic algorithm-back propagation artficial neural network method |
