基于SEER数据库构建前列腺癌骨转移患者预测模型

Construction of a prediction model for patients with prostate cancer bone metastasis based on the SEER database

摘要

目的: 基于SEER数据库建立前列腺癌骨转移风险预测模型,为临床早期识别高危患者提供量化评估工具。方法: 从SEER数据库中筛选2018年1月1日至2021年12月31日期间确诊为前列腺癌骨转移的患者15 055例,采用分层随机抽样按照7∶3的比例划分为训练组(n=10 538)与验证组(n=4 517)。采用Cox比例风险回归模型筛选影响前列腺癌骨转移患者总生存期(overall survival,OS)、癌症特异性生存期(cancer-specific survival,CSS)的独立预后危险因素,基于上述独立影响因素分别构建OS、CSS列线图模型。采用一致性评价指标C指数、受试者工作特征(ROC)曲线、校准曲线及决策曲线分析法(DCA)从区分能力、拟合校准程度与临床净获益综合评价模型效能。结果: 多因素Cox回归分析显示,年龄≥80岁、种族、婚姻状态、格里森分级(Gleason,GS)评分7~9分、T4/TX分期、脑转移、肝转移、根治性手术、放疗、化疗是影响前列腺癌骨转移患者OS的独立危险因素(P均<0.05)。而50~59岁、年龄≥80岁、种族、婚姻状态、GS7~9分、T4/TX期、脑转移、肝转移及肺转移为影响前列腺癌骨转移患者CSS的独立危险因素(P均<0.05)。OS模型训练组C指数为0.710(95%CI:0.701~0.718),验证组为0.710(95%CI:0.697~0.723);CSS模型训练组C指数为0.705(95%CI:0.695~0.715),验证组为0.697(95%CI:0.682~0.712)。OS模型列线图ROC曲线显示,训练组1、2、3年的AUC为0.75、0.72、0.72;验证组1、2、3年的AUC为0.74、0.71、0.70。CSS模型列线图ROC曲线显示,训练组1、2、3年的AUC为0.74、0.71、0.71;验证组1、2、3年的AUC为0.72、0.69、0.69。OS模型校准曲线显示,列线图预测概率与实际生存概率吻合度高;CSS模型校准曲线结果显示,列线图预测概率与实际肿瘤特异性生存概率吻合度高。DCA分析显示OS及CSS预测模型具有良好的临床实用价值。结论: 构建的列线图模型用于预测前列腺癌骨转移患者1、2、3年的OS和CSS,该模型展现出良好的区分效能,具备较高的临床应用价值。

Abstract

Objective: To develop a predictive model for bone metastasis in patients with prostate cancer based on the SEER database, so as to provide a quantitative assessment tool for early identification of high-risk patients in clinical practice. Methods: A total of 15 055 patients diagnosed with prostate cancer complicated by bone metastasis between January 1, 2018 and December 31, 2021 were extracted from the SEER database. Stratified random sampling was performed to divide the cohort into a training set (n=10 538) and an internal validation set (n=4 517) at a ratio of 7∶3. Cox proportional hazards regression models were used to screen independent prognostic risk factors associated with overall survival (OS) and cancer-specific survival (CSS) of prostate cancer patients with bone metastasis. Nomogram models for OS and CSS were separately constructed based on the identified independent prognostic factors. The concordance C-index, receiver operating characteristic (ROC) curves, calibration curves and decision curve analysis (DCA) were adopted to comprehensively evaluate the model performance in terms of discrimination, calibration and clinical net benefit. Results: Multivariate Cox regression analysis revealed that age ≥80 years, race, marital status, Gleason score of 7-9, T4/TX stage, brain metastasis, liver metastasis, radical resection, radiotherapy and chemotherapy were independent risk factors for OS in patients with prostate cancer (all P<0.05). Age 50-59 years, age ≥80 years, race, marital status, Gleason score of 7-9, T4/TX stage, brain metastasis, liver metastasis and lung metastasis were independent risk factors for CSS (all P<0.05). The C-index of the OS nomogram was 0.710 (95%CI: 0.701-0.718) in the training cohort and 0.710 (95%CI: 0.697-0.723) in the validation cohort. For the CSS nomogram, the C-index was 0.705 (95%CI: 0.695-0.715) in the training cohort and 0.697 (95%CI: 0.682-0.712) in the validation cohort. ROC curves of the OS nomogram yielded AUC values of 0.75, 0.72 and 0.72 for survival at 1, 2 and 3 years in the training dataset, with corresponding values of 0.74, 0.71 and 0.70 in the validation cohort. For the CSS nomogram, the AUC at 1, 2 and 3 years were 0.74, 0.71 and 0.71 in the training cohort, and 0.72, 0.69 and 0.69 in the validation cohort. Calibration curves of the OS nomogram demonstrated favorable consistency between predicted survival probabilities and actual observed survival probabilities. Similarly, calibration curves of the CSS nomogram showed strong agreement between predicted cancer-specific survival probabilities and real-world observations. DCA indicated that both the OS and CSS nomograms achieved higher clinical benefit. Conclusion: The constructed nomogram model was used to predict the 1, 2 and 3 year OS and CSS in patients with bone metastatic prostate cancer. The model exhibited favorable discrimination capacity and possessed prominent clinical application value.

关键词

前列腺癌 / 骨转移 / SEER数据库 / 列线图 / 预后

Key words

prostate cancer / bone metastasis / SEER database / nomogram / prognosis

参考文献

[1] Siegel RL, Giaquinto AN, Jemal A . Cancer statistics, 2024[J]. CA Cancer J Clin, 2024, 74(1): 12-49.

[2] Han B, Zheng R, Zeng H, et al. Cancer incidence and mortality in China, 2022[J]. J Natl Cancer Cent, 2024, 4(1): 47-53.

[3] Yang C, Qi GS, Rong RM, et al. Clinical features and treatment strategies for older prostate cancer patients with bone metastasis[J]. Asian J Androl, 2013, 15(6): 759-763.

[4] Handforth C, D′Oronzo S, Coleman R, et al. Cancer treatment and bone health[J]. Calcif Tissue Int, 2018, 102(2): 251-264.

[5] Clézardin P, Coleman R, Puppo M, et al. Bone metastasis: mechanisms, therapies, and biomarkers[J]. Physiol Rev, 2021, 101(3): 797-855.

[6] Song XQ, Liu ZX, Kong QY, et al. Nomogram for prediction of peritoneal metastasis risk in colorectal cancer[J]. Front Oncol, 2022, 12: 928894.

[7] Callegaro D, Miceli R, Mariani L, et al. Soft tissue sarcoma nomograms and their incorporation into practice[J]. Cancer, 2017, 123(15): 2802-2820.

[8] 丁婷婷, 郝珊瑚, 王治国, 等 . ¹⁸F-PSMA-1007 PET/CT与¹⁸F-FDG PET/CT对前列腺癌骨转移的诊断价值比较 [J]. 江苏大学学报(医学版), 2023, 33(5): 369-373, 379.

[9] Shah S, Boucai L . Effect of age on response to therapy and mortality in patients with thyroid cancer at high risk of recurrence[J]. J Clin Endocrinol Metab, 2018, 103(2): 689-697.

[10] Dess RT, Hartman HE, Mahal BA, et al. Association of black race with prostate cancer-specific and other-cause mortality[J]. JAMA Oncol, 2019, 5(7): 975-983.

[11] Rendall MS, Weden MM, Favreault MM, et al. The protective effect of marriage for survival: a review and update[J]. Demography, 2011, 48(2): 481-506.

[12] Epstein JI, Amin MB, Reuter VE, et al. Contemporary gleason grading of prostatic carcinoma: an update with discussion on practical issues to implement the 2014 international society of urological pathology (ISUP) consensus conference on gleason grading of prostatic carcinoma[J]. Am J Surg Pathol, 2017, 41(4): e1-e7.

[13] Whittemore DE, Hick EJ, Carter MR, et al. Significance of tertiary gleason pattern 5 in gleason score 7 radical prostatectomy specimens[J]. J Urol, 2008, 179(2): 516-522.

[14] Ma B, Wells A, Wei L, et al. Prostate cancer liver metastasis: dormancy and resistance to therapy[J]. Semin Cancer Biol, 2021, 71: 2-9.

[15] Teo MY, Rathkopf DE, Kantoff P . Treatment of advanced prostate cancer[J]. Annu Rev Med, 2019, 70: 479-499.

[16] Parker C, Castro E, Fizazi K, et al. Prostate cancer: ESMO clinical practice guidelines for diagnosis, treatment and follow-up[J]. Ann Oncol, 2020, 31(9): 1119-1134.

[17] Wang CY, Yang J, Zi H, et al. Nomogram for predicting the survival of gastric adenocarcinoma patients who receive surgery and chemotherapy[J]. BMC Cancer, 2020, 20(1): 10.

[18] Zhang H, Jiang X, Jiao L, et al. Development and external validation of a novel prognostic nomogram for overall survival in prostate cancer patients with bone metastatic: a retrospective study of the SEER-based and a single Chinese center[J]. J Cancer Res Clin Oncol, 2023, 149(14): 12647-12658.