Retrospective validation of bone risk stratification criteria for men with de novo metastatic hormone-naive prostate cancer in China.

Background: Bone metastasis has been suggested to be a significant impactor on the prognosis of newly diagnosed de novo metastatic hormone-sensitive prostate cancer (mHSPC), and some risk stratification models have been proposed on the basis of this hypothesis. However, the effectiveness of these risk stratification criteria has not been fully evaluated in China. This study aimed to evaluate the effectiveness of the risk stratification models in China. Methods: A total of 140 patients who were newly diagnosed with metastatic prostate cancer followed by primary androgen deprivation-based therapy from January 2008 to June 2021 at our institution were enrolled in this study. The patients were divided into different groups on the basis of high- and low-volume disease (H/LVD) criteria, high-and low-risk disease (H/LRD) criteria, extremity bone metastasis criteria (EBM), and extent of disease (EOD) criteria. The area under the receiver operating characteristic (ROC) curve (AUC) and decision curve analysis (DCA) were used to compare the validity and net benefit of these models. Using the Cox proportional hazards model, we performed univariable and multivariable analyses of the factors influencing overall survival (OS) and the time of progression to metastatic castration-resistant prostate cancer (CRPC). Results: The median patient age was 72 years. Most patients had a Gleason score ≥8 (102 cases, 72.9%) and clinical T stage >2 (75 cases, 53.6%). The median follow-up time was 25 months (range, 2-95 months). Ninety-two patients progressed to CRPC and fifty-seven patients died during the follow-up. The AUC of OS and CRPC showed that the EOD model had higher validity than the other risk stratification models. DCA shows that the net benefit of the EOD model on OS was better than that of the other risk stratification models. As for CRPC, the net benefit of the EOD model was second only to that of the H/LRD model when the threshold was <0.5; however, when the threshold was >0.5, the EOD model outperformed the other models. The effectiveness of EOD as an independent prognostic variable was verified through univariable and multivariable analyses. Conclusion: The EOD model yields reasonable risk stratification for use in Chinese mHSPC patients, providing further evidence supporting its role in clinical decision-making.