Early prediction via PET/CT for skeletal-related events in osteoporotic men with metastatic prostate cancer undergoing androgen deprivation therapy.

OBJECTIVE

Androgen deprivation therapy (ADT) is the cornerstone of treatment for metastatic prostate cancer (mPCa), but it increases the risk of skeletal-related events(SREs), especially for patients with osteoporosis. This study aimed to develop a predictive model for post-ADT SREs progression based on pre-treatment 18F-FDG PET/CT imaging.

METHODS

A total of 118 mPCa patients with osteoporosis who underwent 18F-FDG PET/CT pre-ADT were retrospectively collected from two centers. A comparative analysis of clinical and radiological characteristics, both quantitative and qualitative, was conducted between patients with and without baseline SREs, which were defined as skeletal-related events documented pre-ADT and bone-targeting agents. Subsequently, logistic regression analysis was employed to identify clinical predictors distinguishing patients with SRE progression from those without. Support Vector Machine (SVM) model was then constructed to predict post-ADT SRE progression. Model performance was assessed using receiver operating characteristic (ROC) curves and decision curve analysis (DCA).

RESULTS

In Center 1, 100 patients were stratified into progression (n = 45) and non-progression (n = 55) groups for model training and internal validation. An external validation cohort (n = 18) was obtained from Center 2. The target-to-background ratio of the primary tumor (PCa_TBRmax) was significantly higher in the SRE progression group (p = 0.036). In internal validation, the SVM model achieved the highest AUC of 0.84, while the combined model (incorporating PCa_TBRmax and PSA) reached the AUC of 0.93. In external validation, the combined model achieved an AUC of 0.86.

CONCLUSIONS

The combined clinical-radiological model based on PET/CT demonstrated predictive ability for SRE progression following ADT, offering a valuable tool for early risk stratification and personalized treatment planning in mPCa patients with osteoporosis.