Estimating Bone Metastasis Risk in Prostate Cancer: A Three-Parameter Model Using Bone Sialoprotein, ISUP Grading,...

Introduction: Osseous metastasis is the most common site of distant spread in prostate cancer. Several factors contribute to predicting bone metastasis, including elevated PSA levels, short PSA doubling time, advanced ISUP grading, local tumor progression, and novel biomarkers. However, no clinical scoring system currently exists to assess bone metastasis risk at the time of prostate cancer diagnosis. Furthermore, no study has investigated the correlation between predictive factors and bone sialoprotein (BSP) expression in the primary tumor.

Immunohistochemistry was used to evaluate BSP expression in transrectal ultrasound-guided biopsies from prostate cancer patients. Data from 673 patients were analyzed over a 7-9 year follow-up period to assess the development of bone metastases. BSP expression was also evaluated in patients with benign prostatic hyperplasia (BPH). Additionally, BSP expression was analyzed alongside established risk factors using multivariate logistic regression to determine their combined predictive value for bone metastasis.

Bone metastases developed in 12.5% (84/673) of patients. BSP expression was negative (0-5%) in 23.8% of cases, while 22.2% exhibited high expression (>40%). Patients with bone metastases had significantly higher BSP expression than those without (55.5 ± 19.7% vs. 25.7 ± 24.9%; p < 0.001). In contrast, 97% of patients without prostate carcinoma had BSP values below 5%. Among metastatic patients: 82.9% had BSP expression of at least 40%, and none had values below 20%. As a single predictive parameter, BSP showed a sensitivity of 50% and a specificity of 81.6%. However, using multivariate analysis, a three-parameter scoring model integrating BSP expression, ISUP grading, and the number of affected core needle biopsies achieved 88.6% sensitivity and 81.1% specificity for predicting bone metastases.

BSP expression serves as a potential indicator for bone metastasis development but lacks sufficient sensitivity as a standalone clinical marker. Similarly, local tumor progression and histopathologic grading (ISUP) fail as single predictors. However, integrating BSP expression with established risk factors significantly enhances predictive accuracy. Given that all three parameters are derived from routine histopathological analysis, BSP immunohistochemistry should be considered for integration into clinical practice for early risk stratification in prostate cancer patients.