赣南医科大学学报2025，Vol.45Issue(9)：873-880,8.DOI:10.3969/j.issn.1001-5779.2025.09.010
联合临床、影像及bpMRI影像组学特征构建列线图模型预测高级别前列腺癌
A nomogram model based on clinical,imaging and bpMRI radiomics features to predict high grade of prostate cancer
摘要
Abstract
Objective:To explore the value of a combined nomogram model based on clinical data,imaging,and biparametric magnetic resonance imaging(bpMRI)radiomics to predict high grade of prostate cancer.Methods:A total of 137 prostate cancer patients admitted to our hospital from January 2022 to December 2023 were included in this study.They were divided into a low-grade group(Gleason≤3+4)and a high-grade group(Gleason≥4+3)according to pathological results.Also they were randomly divided into a training group(n=102)and a testing group(n=35)in ratio of 3∶1.The radiomics features of T2WI+ADC images were extracted and screened.A radiomics model was established and radiomics scores(Rad-Score)values were calculated.The training group samples were used to analyze the clinical,imaging,and radiomics parameters by univariate and multivariate analysis to screen the independent risk factors for predicting high-grade prostate cancer.The logistic models of clinical imaging,radiomics,and combined parameters were established and the nomogram was drawn.Delong test and receiver operating characteristic curve(ROC)were used to compare the differences and performance among models,and the decision curve analysis was used to evaluate the clinical benefits of each model.Results:Six radiomics features related to the Gleason grade of prostate cancer were ultimately selected,a Rad-Score prediction model was established,and Rad-Score were calculated.Through univariate and multivariate analyses,MRI-defined prostate specific antigen density(mPSAD),PI-RADS score,and Rad-Score were incorporated into a nomogram model.The AUCs of this model in the training and test sets were 0.881 and 0.853,respectively,exceeding those of the clinical imaging model(AUC:0.793,0.755)and the Rad-Score model(AUC:0.808,0.788),with statistically significant differences(P＜0.05).Decision curve analysis demonstrated that the nomogram model provided greater clinical benefit compared to the other models.Conclusion:The nomogram model based on clinical,imaging and T2WI+ADC imaging demonstrates favorable predictive performance for high-grade prostate cancer and holds potential for clinical application.关键词
前列腺肿瘤/双参数磁共振/影像组学/Gleason分级/列线图Key words
Prostatic cancer/Biparametric magnetic resonance imaging/Radiomics/Gleason grade/Nomogram分类
临床医学引用本文复制引用
周牧野,苏蕾,陈艾琪,乔佳业,谢宗玉,马宜传,陈刘成..联合临床、影像及bpMRI影像组学特征构建列线图模型预测高级别前列腺癌[J].赣南医科大学学报,2025,45(9):873-880,8.基金项目
蚌埠医学院自然科学重点项目(2022byzd063) （2022byzd063）
