| 2 | 0 | 23 |
| 下载次数 | 被引频次 | 阅读次数 |
目的 构建一种针对男性膀胱出口梗阻的机器学习诊断模型。方法 采用回顾性分层抽样收集华中科技大学同济医学院附属同济医院2019—2024年580例男性患者尿动力学检查数据（梗阻290例/非梗阻290例），提取15项特征值，建立数据集。选取不同的特征组合构建模型并利用数据集对不同模型进行训练并验证。以精确度、召回率、F1值、准确度和特异性为评价指标评估模型的性能。结果 采用五折交叉验证，当特征数量为11项时模型性能最优，其曲线下面积值达0.95±0.02。该模型性能显著优于传统方法 （尿道阻力线性图与膀胱排出梗阻指数）。结论 基于反向传播神经网络的机器学习模型，对男性膀胱出口梗阻具有良好的诊断价值。
Abstract：Objective To develop a machine learning diagnostic model for male bladder outlet obstruction. Methods Retrospective urodynamic studydata from 580 male patients(290 obstructive cases and 290 nonobstructive cases) at Tongji Hospital Affiliated To Tongji Medical College HUST(2019-2024) were collected using stratified sampling. Fifteen characteristic parameters were extracted to establish the dataset. Models with different feature combinations were constructed, trained and validated. Performance was evaluated using precision, recall, F1-score, accuracy and specificity. Results The model achieved optimal performance(area under the curve: 0.95±0.02) with 11 selected features via five-fold cross-validation, significantly outperforming traditional methods(Linearized Passive Urethral Resistance Relation and bladder outlet obstruction index). Conclusion The backpropagation neural network-based machine learning model demonstrates strong diagnostic performance for male bladder outlet obstruction.
[1] SCH?FER W,ABRAMS P,LIAO L M,et al. Good urodynamic practices:uroflowmetry,filling cystometry,and pressure-flow studies[J]. Neurourology and urodynamics,2002,21(3):261-274.
[2] DE GROAT W C,GRIFFITHS D,YOSHIMURA N. Neural control of the lower urinary tract[J]. Comprehensive Physiology,2015,5:327-396.
[3] SCH?FER W. Principles and clinical application of advanced urodynamic analysis of voiding function[J]. Urologic clinics of North America,1990,17(3):553-566.
[4] ROSIER P F W M,SCHAEFER W,LOSE G,et al. International Continence Society Good Urodynamic Practices and Terms 2016:urodynamics,uroflowmetry,cystometry,and pressure-flow study[J]. Neurourology and urodynamics,2017,36(5):1243-1260.
[5] LIU X,ZHONG P,GAO Y,et al. Applications of machine learning in urodynamics:a narrative review[J].Neurourology and urodynamics,2024,43(7):1617-1625.
[6] WANG H S,CAHILL D,PANAGIDES J,et al. Pattern recognition algorithm to identify detrusor overactivity on urodynamics[J]. Neurourology and urodynamics,2021,40(1):428-434.
[7] SZMICKI D,BURZY?SKI B,SO?TYSIAK-GIBA?A Z,et al. Prediction of detrusor underactivity based on noninvasive functional tests and clinical data in patients with symptoms of bladder outlet obstruction[J]. European review for medical and pharmacological sciences,2020,24(21):10992-10998.
[8] ZHOU Q,LI G,CUI K,et al. Using machine learning to construct the diagnosis model of female bladder outlet obstruction based on urodynamic study data[J]. Investigative and clinical urology,2024,65(6):559-566.
[9]张义.基于知识图谱的男性下尿路症状临床大数据平台研究[D].上海：中国人民解放军海军军医大学，2019.
[10]MUKAKA M M. Statistics corner:a guide to appropriate use of correlation coefficient in medical research[J].Malawi medical journal:the journal of Medical Association of Malawi,2012,24(3):69-71.
[11] EDELMANN D,MóRI T F,SZéKELY G J. On relationships between the Pearson and the distance correlation coefficients[J]. Statistics&probability letters,2021,169:108960.
[12] TELLO G,AL-JARRAH O Y,YOO P D,et al. Deepstructured machine learning model for the recognition of mixed-defect patterns in semiconductor fabrication processes[J]. IEEE transactions on semiconductor manufacturing,2018,31(2):315-322.
[13] WEI M,WANG Z L,WANG X Y,et al. Prediction of TBM penetration rate based on Monte Carlo-BP neural network[J]. Neural computing and applications,2021,33(2):603-611.
[14] LI J,CHENG J H,SHI J Y,et al. Brief introduction of back propagation(BP)neural network algorithm and its improvement[M]//Advances in Computer Science and Information Engineering. Berlin,Heidelberg:Springer,2012:553-558.
[15] LIU L. Research on water environment monitoring based on the Internet of Things combined with neural network[J]. Optical memory and neural networks,2021,30(3):206-213.
[16] FUSHIKI T. Estimation of prediction error by using Kfold cross-validation[J]. Statistics and computing,2011,21(2):137-146.
[17] JANSSENS A C J W,MARTENS F K. Reflection on modern methods:revisiting the area under the ROC Curve[J]. International journal of epidemiology,2020,49(4):1397-1403.
[18] Hastie T,Tibshirani R,Friedman J H. The elements of statistical learning:Data mining,inference,and prediction[M]. 2nd ed. New York:Springer,2009.
[19] YEN C C,MA C-Y,TSAI Y C. Interpretable machine learning models for predicting critical outcomes in patients with suspected urinary tract infection with positive urine culture[J]. Diagnostics,2024,14(17):1974.
[20]方慧婷，李雨杰，刘世超，等.男性膀胱出口梗阻诊断方法的研究进展[J].泌尿外科杂志（电子版），2023,15(3):69-75.
[21] International continence society. ICS standards 2024:The2024 compilation of international continence society standardisations, consensus statements, educational modules,terminology and fundamentals documents,with the international consultation on incontinence algorithms[S]. Hoboken:Wiley,2024.
[22] JIN L-H,PARK C S,KIM D,et al. Flow starting point and voiding mechanisms measured by simultaneous registrations of intravesical,intra-abdominal,and intraurethral pressures in awake rats[J]. International neurourology journal,2014,18(2):68-76.
[23] PILSETNIECE Z,VJATERS E. Urodynamic values role for females with different types of urinary incontinence[J]. Russian open medical journal,2021,10(3):e0316.
[24]张艳，张瑞莉，孟令峰，等.老年男性逼尿肌活动低下的尿动力学和临床特点分析[J].中华老年医学杂志，2021,40(7):886-889.
基本信息:
DOI：10.13885/j.issn.2097-681X.2025.07.002
中图分类号:R694;TP181
引用信息:
[1]杨正龙,胡友民,陈忠等.基于机器学习的男性膀胱出口梗阻诊断模型[J].兰州大学学报(医学版),2025,51(07):9-14.DOI:10.13885/j.issn.2097-681X.2025.07.002.
基金信息:
湖北省自然科学联合基金项目（2023AFD072）; 华中科技大学创新与转化孵化项目（2022ZHFY11）