基于临床指标、mpMRI及PSMA表达异质性的侧特异性前列腺包膜外侵犯预测模型的构建与验证

Development and validation of a side-specific extraprostatic extension prediction model integrating clinical variables, mpMRI, and PSMA expression heterogeneity

摘要

目的 开发并验证一个整合临床指标、多参数磁共振成像（mpMRI）与穿刺样本前列腺特异性膜抗原（PSMA）免疫组化表达异质性特征的侧特异性前列腺包膜外侵犯（ssEPE）预测模型以优化根治性前列腺切除术（RP）前的神经保留决策。方法 回顾性收集2022年1月-2023年12月在两家三甲医院接受RP的294例前列腺癌患者的临床资料。以空军军医大学西京医院收治的203例406个前列腺侧叶为开发队列，以中国科学技术大学第一附属医院收治的91例182个前列腺侧叶为外部验证队列。以术后病理证实ssEPE为结局，通过单因素和多因素logistic回归分析筛选ssEPE的独立预测因子，并构建列线图模型。采用受试者工作特征曲线下面积（AUC）、校准曲线和决策曲线分析（DCA）评估列线图模型性能，并对该模型进行内部、外部验证以及与Soeterik non-MRI模型、Soeterik MRI模型及Martini模型进行“头对头”比较。结果 多因素logistic回归分析显示，前列腺特异性抗原密度（PSAD）、活检最高国际泌尿病理协会（ISUP）分级、阳性系统穿刺针数百分比、mpMRI评估ssEPE的情况以及穿刺样本前列腺特异性膜抗原香农多样性指数（PSMA-SDI）是ssEPE的独立预测因子（P＜0.05）。基于多因素分析结果构建列线图模型，该列线图模型在开发队列中的AUC为0.861（95%CI：0.826～0.896），校准良好，DCA显示开发队列在7%～92%风险阈值内具有临床净获益。该列线图模型的最佳截断值为0.439，在此阈值下，其诊断ssEPE的灵敏度为79.5%，特异度为78.3%。在外部验证队列中，该列线图模型的AUC为0.852，校准良好，在0%～78%风险阈值内具有临床净获益，内部验证平均AUC为0.840，模型整体性能稳定。本模型的AUC显著优于Soeterik non-MRI模型（AUC：0.825）、Soeterik MRI模型（AUC：0.827）及Martini模型（AUC：0.820）（DeLong检验，均P＜0.01），DCA显示本模型在15%～50%风险阈值范围内净获益高于其他三个模型。结论 成功开发并验证了一个整合临床指标、mpMRI及穿刺样本PSMA免疫组化表达异质性特征的ssEPE预测模型，其预测效能和临床实用性优于现有工具，可为前列腺癌患者RP术前个体化神经保留手术决策提供可靠依据。

Abstract

Objective To develop and validate a side-specific extraprostatic extension (ssEPE) prediction model that integrates clinical characteristics, multiparametric MRI (mpMRI), and heterogeneity in prostate-specific membrane antigen (PSMA) expression in biopsy specimens, so as to optimize nerve-sparing decision-making prior to radical prostatectomy (RP). Methods Clinical data of 294 prostate cancer patients who underwent RP at two Class A hospitals during Jan. 2022 and Dec. 2023 were retrospectively collected. A total of 406 prostate lobes from 203 patients treated in Xijing Hospital of Air Force Medical University served as the development cohort, and 182 prostate lobes from 91 patients treated in the First Affiliated Hospital of University of Science and Technology of China were assigned to the external validation cohort. Postoperative pathological confirmation of ssEPE was defined as the primary outcome. Independent predictors were identified using univariate and multivariate logistic regression analyses, and a nomogram model was constructed. The model was evaluated with area under the receiver operating characteristic curve (AUC), calibration curves, and decision curve analysis (DCA), with internal and external validation, and compared head-to-head with the Soeterik non-MRI model, Soeterik MRI model, and Martini model. Results Multivariate analyses revealed that prostate-specific antigen density (PSAD), highest International Society of Urological Pathology (ISUP) grade on biopsy, percentage of positive systematic biopsy cores, mpMRI-assessed ssEPE, and PSMA shannon diversity index (PSMA-SDI) from biopsy were independent predictors of ssEPE (all P＜0.05). In the development cohort, the model achieved an AUC of 0.861 (95%CI: 0.826-0.896) with good calibration. DCA demonstrated clinical net benefits for the development cohort across a risk threshold range of 7%-92%. The optimal cutoff value was 0.439, at which the sensitivity and specificity were 79.5% and 78.3%, respectively. In the external validation cohort, the model yielded an AUC of 0.852 with acceptable calibration and clinical net benefits within a 0%-78% risk threshold. Internal validation showed a mean AUC of 0.840, indicating stable overall performance. When compared with other models, the model achieved a significantly higher AUC than the Soeterik non-MRI model (0.825), Soeterik MRI model (0.827), and Martini model (0.820) (DeLong test, all P＜0.01). DCA showed that the model provided higher net benefit than the other three models within the risk threshold range of 15%-50%. Conclusion A ssEPE prediction model integrating clinical variables, mpMRI, and heterogeneity in PSMA expression from biopsy specimens was successfully developed and validated. Its predictive efficacy and clinical utility outperform existing tools, providing a reliable basis for individualized preoperative nerve-sparing decision-making in prostate cancer patients undergoing RP.

关键词

前列腺癌 / 侧特异性前列腺包膜外侵犯 / 前列腺特异性膜抗原 / 肿瘤异质性 / 根治性前列腺切除术 / 列线图模型

Key words

prostate cancer / side-specific extraprostatic extension / prostate-specific membrane antigen / tumor heterogeneity / radical prostatectomy / nomogram model

参考文献

[1] SALOMON L, SEBE P, DE LA TAILLE A, et al. Open versus laparoscopic radical prostatectomy: part I[J]. BJU Int, 2004, 94(2): 238-243.

[2] REBELLO RJ, OING C, KNUDSEN KE, et al. Prostate cancer[J]. Nat Rev Dis Primers, 2021, 7(1): 9.

[3] WALSH PC. The discovery of the cavernous nerves and development of nerve sparing radical retropubic prostatectomy[J]. J Urol, 2007, 177(5): 1632-1635.

[4] MOTTET N, BELLMUNT J, BOLLA M, et al. EAU-ESTRO-SIOG guidelines on prostate cancer. Part 1: screening, diagnosis, and local treatment with curative intent[J]. Eur Urol, 2017, 71(4): 618-629.

[5] NGUYEN LN, HEAD L, WITIUK K, et al. The risks and benefits of cavernous neurovascular bundle sparing during radical prostatectomy: a systematic review and meta-analysis[J]. J Urol, 2017, 198(4): 760-769.

[6] OHORI M, KATTAN MW, KOH H, et al. Predicting the presence and side of extracapsular extension: a nomogram for staging prostate cancer[J]. J Urol, 2004, 171(5): 1844-1849.

[7] TOSOIAN JJ, CHAPPIDI M, FENG Z, et al. Prediction of pathological stage based on clinical stage, serum prostate-specific antigen, and biopsy Gleason score: partin tables in the contemporary era[J]. BJU Int, 2017, 119(5): 676-683.

[8] Memorial Sloan Kettering Cancer Center. Nomogram for prediction of prostate cancer extent before radical prostatectomy[EB/OL]. (2024)[2025-11-07]. http://www.mskcc.org/momograms/prostate.

[9] BRAJTBORD JS, LEAPMAN MS, COOPERBERG MR. The CAPRA score at 10 years: contemporary perspectives and analysis of supporting studies[J]. Eur Urol, 2017, 71(5): 705-709.

[10] SOETERIK TFW, VAN MELICK HHE, DIJKSMAN LM, et al. Development and external validation of a novel nomogram to predict side-specific extraprostatic extension in patients with prostate cancer undergoing radical prostatectomy[J]. Eur Urol Oncol, 2022, 5(3): 328-337.

[11] HOFFMANN E, MASTHOFF M, KUNZ WG, et al. Multiparametric MRI for characterization of the tumour microenvironment[J]. Nat Rev Clin Oncol, 2024, 21(6): 428-448.

[12] CAROMILE LAO, DORTCHE K, RAHMAN MM, et al. PSMA redirects cell survival signaling from the MAPK to the PI3K-AKT pathways to promote the progression of prostate cancer[J]. Sci Signal, 2017, 10(470): eaag3326.

[13] KAITTANIS C, ANDREOU C, HIERONYMUS H, et al. Prostate-specific membrane antigen cleavage of vitamin B9 stimulates oncogenic signaling through metabotropic glutamate receptors[J]. J Exp Med, 2018, 215(1): 159-175.

[14] KASHYAP A, RAPSOMANIKI MA, BARROS V, et al. Quantification of tumor heterogeneity: from data acquisition to metric generation[J]. Trends Biotechnol, 2022, 40(6): 647-676.

[15] CORNFORD P, TILKI D, VAN DEN BERGH RCN, et al. EAU-EANM-ESTRO-ESUR-ISUP-SIOG guidelines on prostate cancer[M]. 2025 ed. Arnhem: EAU Guidelines Office, 2025: 1-251.

[16] SPRATT DE, SRINIVAS S, ADRA N. National Comprehensive Cancer Network. NCCN clinical practice guidelines in oncology: prostate cancer[M]. Version 2.2026. Plymouth Meeting, PA: National Comprehensive Cancer Network, 2025.

[17] ROCCO B, SIGHINOLFI MC, SANDRI M, et al. Is extraprostatic extension of cancer predictable? A review of predictive tools and an external validation based on a large and a single center cohort of prostate cancer patients[J]. Urology, 2019, 129: 8-20.

[18] ZHU M, GAO J, HAN F, et al. Diagnostic performance of prediction models for extraprostatic extension in prostate cancer: a systematic review and meta-analysis[J]. Insights Imaging, 2023, 14(1): 140.

[19] SAYAR E, PATEL RA, COLEMAN IM, et al. Reversible epigenetic alterations mediate PSMA expression heterogeneity in advanced metastatic prostate cancer[J]. JCI Insight, 2023, 8(7): e162907.

[20] KWONG JCC, KHONDKER A, MENG E, et al. Development, multi-institutional external validation, and algorithmic audit of an artificial intelligence-based side-specific extra-prostatic extension risk assessment tool (SEPERA) for patients undergoing radical prostatectomy: a retrospective cohort study[J]. Lancet Digit Health, 2023, 5(7): e435-e445.

[21] AMERICAN COLLEGE OF RADIOLOGY. PI-RADS prostate imaging-reporting and data system[M]. Version 2.1. Reston, VA: American College of Radiology, 2019.

[22] MEHRALIVAND S, SHIH JH, HARMON S, et al. A grading system for the assessment of risk of extraprostatic extension of prostate cancer at multiparametric MRI[J]. Radiology, 2019, 290(3): 709-719.

[23] RILEY RD, SNELL KI, ENSOR J, et al. Minimum sample size for developing a multivariable prediction model: PART Ⅱ-binary and time-to-event outcomes[J]. Stat Med, 2019, 38(7): 1276-1296.

[24] MARTINI A, GUPTA A, LEWIS SC, et al. Development and internal validation of a side-specific, multiparametric magnetic resonance imaging-based nomogram for the prediction of extracapsular extension of prostate cancer[J]. BJU Int, 2018, 122(6): 1025-1033.

基金资助

国家自然科学基金国际（地区）合作研究与交流项目（重点项目）(82220108004)

陕西省卫生健康科研创新能力提升计划(2024TD-08)