基于机器学习的列线图预测初诊前列腺癌临床分期:一项多中心回顾性研究
本研究利用生物标志物和临床病理特征构建并外部验证基于机器学习的列线图,用于预测初诊前列腺癌的进展分期。训练集与测试集来自同一中心362例患者(按3:7随机分配),外部验证集来自另一中心136例患者。逻辑回归模型表现最优,构建的进展期前列腺癌风险列线图和转移性前列腺癌风险列线图在训练集、测试集和外部验证集中的AUC分别为0.848、0.814、0.810和0.940、0.913、0.910。
要点
- 逻辑回归模型在机器学习模型中表现最优,用于构建进展期和转移性前列腺癌风险列线图。
- 进展期前列腺癌风险列线图在训练集、测试集和外部验证集的AUC分别为0.848、0.814、0.810。
- 转移性前列腺癌风险列线图在训练集、测试集和外部验证集的AUC分别为0.940、0.913、0.910。
核心结论
- 基于机器学习的进展期前列腺癌风险列线图和转移性前列腺癌风险列线图对前列腺癌进展分期表现出出色的预测性能,可协助临床医生对新诊断前列腺癌患者进行精细分类,促进个性化治疗方案和预后评估。
研究设计与样本
- 多中心回顾性研究,训练集和测试集来自第一附属医院362例确诊前列腺癌的住院患者,按3:7比例随机分配,外部验证集来自人民医院136例前列腺癌患者。
- 收集影像学和临床病理信息,通过逻辑回归识别区分进展期前列腺癌和转移性前列腺癌的最佳特征,采用机器学习算法构建并比较模型,表现最佳的算法用于建立前列腺癌进展分期模型,在训练集、测试集和外部验证集中使用指标、ROC曲线、校准和决策曲线分析评估模型性能。
主要发现
- 逻辑回归分析显示,PSA(P=0.001)、最大肿瘤直径(P=0.026)、Gleason评分(P<0.001)和RNF41(P<0.001)是预测进展期前列腺癌的最佳特征;ALP(P<0.001)、PSA(P<0.001)和GS评分(P=0.024)是预测转移性前列腺癌的最佳特征。
- 在机器学习模型中,逻辑回归模型表现出更优性能,因此用于构建进展期前列腺癌风险列线图和转移性前列腺癌风险列线图,其在训练集、测试集和外部验证集中的AUC分别为0.848、0.814、0.810和0.940、0.913、0.910。
- 校准曲线和决策曲线分析证实列线图的一致性和对临床决策的净获益。
局限性
- 原文未报告局限性。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
原文
OBJECTIVE: To construct and externally validate machine learning-based nomograms for predicting progression stages of initial prostate cancer (PCa) using biomarkers and clinicopathologic features. METHODS: Three hundred sixty-two inpatients diagnosed with PCa at the First Affiliated Hospital were randomly assigned to training and testing sets in a 3:7 ratio, while 136 PCa patients from People's Hospital formed the external validation set. Imaging and clinicopathologic information were collected. Optimal features distinguishing advanced prostate cancer (APC) and metastatic PCa (mPCa) were identified through logistic regression (LR). ML algorithms were employed to build and compare ML models. The best-performing algorithm established models for PCa progression stage. Models performance was evaluated using metrics, ROC curves, calibration, and decision curve analysis (DCA) in training, testing, and external validation sets. RESULTS: Following LR analyses, PSA (P = .001), maximum tumor diameter (P = .026), Gleason score (P <.001), and RNF41 (P <.001) were optimal features for predicting APC, while ALP (P <.001), PSA (P <.001), and GS score (P = .024) were for mPCa. Among ML models, the LR models exhibited superior performance. Consequently, the LR algorithm was used for the APC-risk-nomogram and mPCa-risk-nomogram construction, with AUC values of 0.848, 0.814, 0.810, and 0.940, 0.913, 0.910, in the training, testing, and external validation sets, respectively. Calibration and DCA curves affirmed nomograms' consistency and net benefits for clinical decision-making. CONCLUSION: In summary, ML-based APC-risk-nomogram and mPCa-risk-nomogram exhibit outstanding predictive performance for PCa progression stages. These nomograms can assist clinicians in finely categorizing newly diagnosed PCa patients, facilitating personalized treatment plans and prognosis assessment.
引用
- Chen L, Fu Z, Dong Q, et al. Machine Learning-based Nomograms for Predicting Clinical Stages of Initial Prostate Cancer: A Multicenter Retrospective Study. Urology. 2024. PMID: 39153604. doi:10.1016/j.urology.2024.08.011
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。