| 摘要: |
| ［摘要］ 目的 基于多参数磁共振成像（mpMRI）影像组学与临床病理特征构建鉴别前列腺癌（PCa）Gleason分级的列线图模型，并对该模型的诊断效能和应用价值进行评估。方法 招募2024年1月至2024年12月睢宁县人民医院收治的PCa患者180例。根据病理检查和Gleason分级结果将其分为低级别PCa组（88例，Gleason分级为1～2级），高级别PCa组（92例，Gleason分级为3～5级）。所有患者均接受MRI检查，提取mpMRI影像组学特征，最终从T2加权成像（T2WI）及扩散加权成像（DWI）序列中各筛选出2个最优特征，以此4个指标构建mpMRI影像组学评分体系。通过多因素logistic回归分析筛选有助于鉴别高级别PCa和低级别PCa的指标，并基于筛得指标构建列线图模型。通过受试者工作特征（ROC）曲线、校准曲线和决策曲线分析列线图模型的诊断效能和应用价值。结果 高级别PCa组mpMRI影像组学评分、肿瘤最大径、前列腺特异性抗原（PSA）、前列腺特异性抗原密度（PSAD）水平以及T分期为T3～4期、N分期为N1期的占比显著高于低级别PCa组（P<0.05）。多因素logistic回归分析结果显示，mpMRI影像组学评分、肿瘤最大径、PSA、PSAD、N分期是区分高级别PCa和低级别PCa的有意义指标（P<0.05），基于这5个指标构建列线图模型。ROC曲线分析结果显示，列线图模型具有较好的鉴别诊断效能［AUC（95%CI）=0.996（0.972～0.999）］。校准曲线接近理想状态，提示列线图模型的诊断准确性好。决策曲线分析结果显示，在2%至100%的预测概率范围内，列线图模型能提供较好的临床净获益。结论 基于mpMRI影像组学与临床病理特征构建鉴别诊断高级别PCa和低级别PCa的列线图模型具有良好的诊断效能和临床应用价值。 |
| 关键词: 多参数磁共振成像影像组学 临床特征 病理特征 前列腺癌 Gleason分级 列线图 |
| DOI：10.3969/j.issn.1674-3806.2026.07.12 |
| 分类号:R 737.25 |
| 基金项目:徐州市卫生健康委员会科技项目（编号：XWKYSL20220243） |
|
| Construction of a nomogram model for differentiating Gleason grades in prostate cancer based on mpMRI radiomics and clinicopathological characteristics and evaluation of its diagnostic efficacy |
|
Song Tengteng1, Xie Lixiang2, Zhuo Shuo2, Zhang Zehua1, Li Shaodong2
|
|
1.Department of Radiology, Suining County People′s Hospital, Xuzhou 221300, China; 2.Department of Radiology, Affiliated Hospital of Xuzhou Medical University, Xuzhou 221000, China
|
| Abstract: |
| ［Abstract］ Objective To construct a nomogram model for differentiating Gleason grades in prostate cancer(PCa) based on multiparametric magnetic resonance imaging(mpMRI) radiomics and clinicopathological characteristics, and to evaluate the diagnostic efficacy and application value of the model. Methods A total of 180 patients with PCa who were admitted to Suining County People′s Hospital from January 2024 to December 2024 were recruited. According to the results of pathological examination and Gleason grades of the patients, they were divided into low-grade PCa group(88 patients, with Gleason grades of 1-2) and high-grade PCa group(92 patients, with Gleason grades of 3-5). All the patients underwent magnetic resonance imaging(MRI) examination, and their mpMRI radiomics features were extracted. Ultimately, 2 optimal features were selected from each of the T2-weighted imaging(T2WI) and diffusion-weighted imaging(DWI) sequences, and these 4 indicators were used to construct an mpMRI radiomics scoring system. Multivariate logistic regression analysis was used to screen the indicators that were helpful for differentiating high-grade PCa from low-grade PCa, and a nomogram model was constructed based on the screened indicators. The diagnostic efficacy and application value of the nomogram model were analyzed by using receiver operating characteristic(ROC) curve, calibration curve and decision curve. Results The mpMRI radiomics scores, maximum diameter of tumors, prostate specific antigen(PSA), prostate specific antigen density(PSAD), the proportion of T stages of T3-4 and the proportion of N stage of N1 in the high-grade PCa group were significantly larger than those in the low-grade PCa group(P<0.05). The results of multivariate logistic regression analysis showed that mpMRI radiomics scores, maximum diameter of tumors, PSA, PSAD and N stage were the meaningful indicators for distinguishing the high-grade PCa from the low-grade PCa(P<0.05), and a nomogram model was constructed based on the 5 indicators. The results of ROC curve analysis showed that the nomogram model had good efficacy of differential diagnosis［AUC(95%CI)= 0.996(0.972-0.999)］. The calibration curve was close to the ideal state, indicating that the diagnostic accuracy of the nomogram model was good. The results of decision curve analysis showed that within the predicted probability ranging from 2% to 100%, the nomogram model could provide better clinical net benefits. Conclusion The nomogram model constructed based on mpMRI radiomics and clinicopathological features has good diagnostic efficacy and clinical practical value for the differential diagnosis of high-grade PCa and low-grade PCa. |
| Key words: multiparametric magnetic resonance imaging(mpMRI) radiomics clinical characteristics pathological characteristics prostate cancer(PCa) Gleason grade nomogram |
