Abstract
OBJECTIVES: Approximately, 90% of men with advanced prostate cancer will develop bone metastasis. However, there have been few reports about noninvasive biomarker to detect and predict clinical outcome of bone metastasis (BM) in prostate cancer patients. METHODS: We examined 1127 patients who underwent prostate biopsy from August 2012 to June 2017. We also investigated bone turnover markers such as bone-specific alkaline phosphatase, type I collagen cross-linked N-terminal telopeptide, C-terminal pyridinoline cross-linked telopeptide of type I collagen, and tartrate-resistant acid phosphatase type 5b (TRACP 5b). RESULTS: A total of 282 patients were diagnosed as prostate cancer with complete clinical data, and 34 patients with bone metastasis. Multivariate analysis revealed C-terminal pyridinoline cross-linked telopeptide of type I collagen, tartrate-resistant acid phosphatase type 5b, and prostate-specific antigen (PSA) were independent biomarkers in detection of BM (p < 0.05, respectively). Furthermore, we developed predictive model formula based on tartrate-resistant acid phosphatase type 5b and PSA, for which the area under the curve was 0.95. In patients with bone metastasis, multivariate cox proportional hazards analysis revealed that this model was significantly associated with poor clinical outcome of cancer-specific survival (p < 0.05). In validation cohort with 137 patients, we also confirmed the utility of this model for diagnosis of BM (the area under the curve = 0.95). CONCLUSIONS: Our developed formula of tartrate-resistant acid phosphatase type 5b in accordance with PSA may serve as the useful tool in diagnosis and prediction of clinical outcome for prostate cancer with bone metastasis.