通讯作者: 汪国余, hswangguoyu@163.com

DOI：10.12201/bmr.202406.00029

Application value of biparametric magnetic resonance radiomics combined with prostate-specific antigen density in Gleason grade group of prostate cancer

• 摘要：目的:探讨双参数MRI (biparametric MRI，bpMRI)影像组学联合前列腺特异性抗原密度（prostate-specific antigen density，PSAD）在低、高级别前列腺癌（prostate cancer，PCa）中的诊断价值。方法:回顾性分析179例PCa患者，按照7:3比例，将患者随机分为训练组和测试组。根据Gleason分级分组（gleason grade group，GGG），将GGG组≤2定义为低级别PCa，GGG＞2组定义为高级别PCa。基于T2WI和ADC序列提取影像组学特征，采用最大相关最小冗余、最小绝对收缩和选择算子进行特征选择和降维，并进行5倍交叉验证，保留最佳特征组合，并构建单因素和多因素logistic回归模型，分别为3个单因素模型(ADC、T2WI和PSAD模型)和2个联合模型(T2WI-ADC和T2WI-ADC-PSAD模型)。通过ROC曲线和Delong检验评估各模型的诊断性能。最后，采用决策曲线分析(DCA)评价模型的临床效用。结果:单序列模型中，ADC序列模型诊断性能最好（AUC=0.713，95%CI：0.582～0.833），T2WI-ADC联合模型诊断性能较单序列模型没有提升（AUC=0.709 95%CI：0.572～0.830）。所有模型中，T2WI-ADC-PSAD联合模型的诊断性能最好（AUC=0.772，95%CI：0.661～0.874），Delong检验显示测试组中联合模型与其他模型的AUC比较无明显统计学差异（P＞0.05）。DCA结果显示，当阈值概率在相关范围内，T2WI-ADC-PSAD联合模型为临床决策提供了更高的净获益。结论:BpMRI影像组学联合PSAD可提高对低、高级别PCa的诊断效能，并指导患者的治疗决策。

  关键词： 双参数磁共振; 前列腺特异性抗原密度; 影像组学; 预测模型; Gleason分级分组

   

  Abstract: Objective: To investigate the diagnostic value of biparametric MRI (bpMRI) radiomics and prostate-specific antigen density (PSAD) in predicting low-grade and high-grade prostate cancerMethods: A total of 179 PCa patients were retrospectively analyzed and randomly assigned into training (n?=?125) and test (n = 54) cohorts. According to gleason grade group (GGG), GGG≤2 was defined as low-grade PCa, and GGG > 2 was defined as high-grade PCa. Radiomics features were extracted based on T2WI and ADC sequences, and maximum relevance and minimum redundancy and least absolute shrinkage and selection operator were used for feature selection and dimensionality reduction, and 5-fold cross validation was performed to retain the best radiomics features. Based on the best feature combination, univariate and multivariate logistic models were constructed, including three univariate models (ADC, T2WI and PSAD model) and two combined models (T2WI-ADC and T2WI-ADC-PSAD model). Receiver operating characteristic(ROC) curve and Delong’s test were used to evaluate the performance of each model. Finally, decision curve analysis (DCA) was used to evaluate the clinical utility of the model.Results: Among the single sequence models, the ADC model had the best diagnostic performance (AUC=0.713, 95%CI: 0.582-0.833), however, the T2WI-ADC combined model did not improve the diagnostic performance compared with ADC model (AUC=0.709, 95%CI: 0.572-0.830). Among all the models, T2WI-ADC-PSAD combined model had the best diagnostic performance (AUC=0.772, 95%CI: 0.661-0.874). Delong’s test showed that there was no significant difference in AUC between the combined model and the other models in the test cohort (P > 0.05). The DCA showed that the T2WI-ADC-PSAD model provided a higher net benefit for clinical decision-making when the threshold probability within the relevant range.Conclusion: BpMRI radiomics combined with PSAD can improve the detection of high-grade PCa, and guide patient treatment decisions.

  Key words: Biparametric MRI; Prostate-Specific Antigen Density; Radiomics; Prediction model; Gleason grade group

  提交时间：2024-06-19

  版权声明：作者本人独立拥有该论文的版权，预印本系统仅拥有论文的永久保存权利。任何人未经允许不得重复使用。
• html

• 图表

• 思小霞, 赵青, 景江新. 超声影像组学在预测乳腺癌分子分型中的应用进展. 2024. doi: 10.12201/bmr.202410.00040

  徐婧, 柯清仙, 黄丽敏, 刘晓莺, 林丽珍. 基于LEARNS模式的健康教育对腹腔镜前列腺癌根治术后尿失禁患者的效果研究. 2024. doi: 10.12201/bmr.202409.00038

  魏雪瑶, 郑辉, 陈莹, 刘震, 倪如暘, 刘艳亭. 基于特异性指标与实施效果评价的临床路径系统的设计与实现. 2020. doi: 10.12201/bmr.202010.00839

  阮旭凌, 刘琦, 郭志恒, 晏峻峰. 基于LDA和XGBoost算法的乳腺癌预测模型构建研究. 2022. doi: 10.12201/bmr.202106.00007

  周林. 血清GP73联合预后营养指数对HCC患者TACE术后预后的评估价值. 2024. doi: 10.12201/bmr.202409.00043

  杜晓菲, 陈杰, 仇丽霞, 柳雅立. BOPPPS联合案例教学法在传染病教学中的应用. 2024. doi: 10.12201/bmr.202406.00015

  施明杰, 叶洪舟, 袁琛. 大环内酯类耐药肺炎支原体肺炎患儿发展为重症肺炎的危险因素分析及预测模型构建. 2024. doi: 10.12201/bmr.202409.00021

  黄昱诚, 杨旭明, 乔琼. 基于数据挖掘的乳腺癌复发因素预测模型构建. 2020. doi: 10.12201/bmr.202009.00011

  叶懿萱. 伊伐布雷定对COPD合并HFpEF患者心肺运动试验参数、心率变异性及hs-CRP、NT-proBNP的影响. 2024. doi: 10.12201/bmr.202408.00002

• 序号	提交日期	编号	操作
  1	2024-06-16	bmr.202406.00029V1	下载

• 公开评论  匿名评论  仅发给作者 提交评论