Clinical characteristics and prostate-cancer-specific mortality of competitive risk nomogram in the second primary prostate cancer

Chaojie Xu; Dongchen Pei; Yi Liu; Jianhua Guo; Nan Liu; Qian Wang; Yang Yu; Zhengjun Kang

doi:10.3389/fonc.2023.918324

Clinical characteristics and prostate-cancer-specific mortality of competitive risk nomogram in the second primary prostate cancer

Front Oncol. 2023 May 16:13:918324. doi: 10.3389/fonc.2023.918324. eCollection 2023.

Authors

Chaojie Xu¹, Dongchen Pei¹, Yi Liu¹, Jianhua Guo¹, Nan Liu¹, Qian Wang¹, Yang Yu¹, Zhengjun Kang¹

Affiliation

¹ Department of Urology, the Fifth Affiliated Hospital of Zhengzhou University, Zhengzhou University, Zhengzhou, Henan, China.

Abstract

Background: With the development of early diagnosis and treatment, the second primary malignancy (SPM) attracts increasing attention. The second primary prostate cancer (spPCa) is an important class of SPM, but remains poorly understood.

Methods: We retrospectively analyzed 3,322 patients with spPCa diagnosed between 2004 and 2015 in the Surveillance, Epidemiology, and End Results (SEER) database. Chi-square test was applied to compare demographic and clinical variables and analyze causes of death. Multivariate competitive risk regression model was used to identify risk factors associated with prostate-cancer-specific mortality (PCSM), and these factors were enrolled to build a nomogram of competitive risk. The C-index, calibration curve, and decision curve analysis (DCA) were employed to evaluate the discrimination ability of our nomogram.

Results: The median follow-up (interquartile range, IQR) time was 47 (24-75) months, and the median (IQR) diagnosis interval between the first primary cancer (FPC) and spPCa was 32 (16-57) months. We found that the three most common sites of SPM were the urinary system, digestive system, and skin. Through multivariate competitive risk analysis, we enrolled race (p < 0.05), tumor-node-metastasis (TNM) stage (p < 0.001), Gleason score (p < 0.05), surgery (p = 0.002), and radiotherapy (p = 0.032) to construct the model to predict the outcomes of spPCa. The C-index was 0.856 (95% CI, 0.813-0.899) and 0.905 (95% CI, 0.941-0.868) in the training and validation set, respectively. Moreover, both the calibration curve and DCA illustrated that our nomogram performed well in predicting PCSM.

Conclusion: In conclusion, we identified four risk factors associated with the prognosis of spPCa and construct a competing risk nomogram, which performed well in predicting the 3-, 5-, and 10-year PCSM.

Keywords: SEER database; competing risk regression analysis; nomogram; prostate cancer-specific mortality (PSCM); second primary malignancy.

Grants and funding

This study was funded by the Health Care Commission of Henan Province (201403125, LHGJ20190422) and the Department of Education of Henan Province (21A320037).