首页|基于决策树算法的3.0T MRI及DWI参数对局部晚期宫颈癌腔内后装放疗疗效的预测价值研究

基于决策树算法的3.0T MRI及DWI参数对局部晚期宫颈癌腔内后装放疗疗效的预测价值研究

扫码查看
目的:研究基于决策树算法的3.0T MRI及DWI参数对局部晚期宫颈癌腔内后装放疗疗效的预测价值.方法:对2019年4月至2024年4月在我院接受腔内后装放疗治疗的局部晚期宫颈癌患者进行回顾性研究,并根据放疗效果分为有效组(71例)和无效组(79例).两组患者均接受3.0T MRI扫描、磁共振动态增强(DCE-MRI)扫描及DWI扫描,获取表观弥散系数(ADC)、容量转移常数(Ktrans)、速率常数(Kep)、血管外细胞外容积分数(Ve),并基于决策树算法分析影响疗效的相关因素,建立ROC曲线,探讨各因素对疗效的诊断价值.结果:有效组的ADC高于无效组,Kep、Ktrans、Ve低于无效组,差异具有统计学意义(t=23.634、17.044、16.266、18.788,P<0.001).决策树模型结果显示,ADC、Kep、Ktrans、Ve是影响晚期宫颈癌患者放疗疗效的影响因素(P<0.05),其准确度为90.70%,ROC曲线的曲线下面积为0.932(95%CI:0.866~0.977),敏感度为87.30%,特异性为94.40%.结论:本研究基于3.0T MRI扫描及DWI扫描参数构建了决策树模型,该模型对局部晚期宫颈癌腔内后装放疗疗效的诊断价值较高,能有效为临床治疗措施提供参考依据.
Based on the Decision Tree Algorithm,the Correlation between 3.0T Magnetic Resonance Imaging Scan and Diffusion-weighted Imaging Parameters and the Efficacy of Intracavitary Brachytherapy for Locally Advanced Cervical Cancer was Analyzed
Objective:To investigate the predictive value of 3.0T MRI and DWI parameters,utilizing decision tree algorithms,for assessing the efficacy of intracavitary brachytherapy in patients with locally advanced cervical cancer.Methods:A retrospective study was conducted on patients with locally advanced cervical cancer who underwent intracavitary brachytherapy in our hospital from April 2019 to April 2024.According to the effect of radiotherapy,they were divided into effective group(71 cases)and ineffective group(79 cases).Both groups of patients underwent 3.0T MRI scan,dynamic contrast-enhanced MRI(DCE-MRI)scan and DWI scan to obtain apparent diffusion coefficient(ADC),volume transfer constant(Ktrans),rate constant(Kep),and extravascular extracellular volume fraction(Ve).Based on the decision tree algorithm,the ROC curve was established to explore the diagnostic value of each factor.Results:the effective group has a higher ADC than the ineffective group,and a lower Kep,Ktrans and Ve,the differences were statistically significant(t=23.634,17.044,16.266,18.788;P<0.001).The results of the decision tree model showed that ADC,Kep,Ktransand Ve were the influencing factors affecting the efficacy of radiotherapy in patients with advanced cervical cancer(P<0.05).The accuracy was 90.70%,the area under the ROC curve was 0.932(95%CI:0.866-0.977),the sensitivity was 87.30%,and the specificity was 94.40%.Conclusion:In this study,a decision tree model was constructed based on 3.0T MRI scan and DWI scan parameters.The model has high diagnostic value for the efficacy of intracavitary brachytherapy for locally advanced cervical cancer,and can effectively provide reference for clinical treatment measures.

decision tree algorithm3.0T magnetic resonance imagingdiffusion-weighted imaginglocally advanced cervical cancerintracavitary brachytherapy

吴彬彬、胡良先、杨珊珊、黄俊

展开 >

中国人民解放军海军安庆医院(安庆一一六医院),安徽 246003

决策树算法 3.0T磁共振成像 弥散加权成像 局部晚期宫颈癌 腔内后装放疗

2025

影像科学与光化学
中国科学院理化技术研究所 中国感光学会

影像科学与光化学

影响因子:0.287
ISSN:1674-0475
年,卷(期):2025.43(1)