Login to Access Video or Poster Abstract: MP14-20
Sources of Funding: None

Introduction

Two recent randomized trials have suggested a survival benefit for patients with high volume metastatic prostate cancer (PCa) that initially receive chemotherapy. However, there is a paucity of data at the population-level characterizing which men presenting with bone metastatic disease are at highest risk for PCa-specific mortality (PCSM). The purpose of this study was to assess the demographic and clinicopathologic factors among patients with PCa bone metastasis at diagnosis and identify predictors of PCSM.

Methods

Patients with prostate adenocarcinoma were identified between 2010-2013 from the SEER database (n=200,616). Among this group, 8,040 men presented with bone metastasis, forming the study cohort. Descriptive statistics were used to compare demographic and clinicopathologic variables between patients experiencing PCSM and those that were alive/died of other causes. A Fine and Gray&[prime]s sub-distribution competing risks model was performed to generate hazards ratios (HR) for the identification of predictors of PCSM.

Results

There were 2,497 men (31.1%) experiencing PCSM and 5,543 men (68.9%) without PCSM (n=643 dead of other causes; n=4,900 alive) over a median follow-up of 35 months (IQR: 34-37). On univariate analysis, patients suffering PCSM were older (median 72 vs 70 years of age, p<0.0001), unmarried (40.5% vs 36.7%, p=0.001), more likely to live in the Southeast US (24.6% vs 20.8%, p=0.0006), have biopsy Gleason Group (bGG) 5 disease (40.7% vs 38.6%) or have no prostate biopsy (28.1% vs 19.5%, p<0.0001), and have concomitant PCa brain (2.1% vs 0.8%, p<0.0001), liver (6.1% vs 2.4%, p<0.0001) and lung metastases at diagnosis (9.0% vs 5.7%, p<0.0001) compared to patients without PCSM. Multivariable competing risks modelling identified older age (HR 1.023, 95CI 1.019-1.027), non-black/white race (vs black HR 0.77, 95CI 0.62-0.95), unmarried status (vs married HR 1.10, 95CI 1.01-1.20), living in the Southeast US (vs Northeast US HR 1.24, 95CI 1.07-1.44), PSA (HR 1.005, 95CI 1.003-1.008), bGG 4 (vs 1 HR 1.53, 95CI 1.04-2.26), bGG 5 (vs 1 HR 2.18, 95CI 1.50-3.19), no prostate biopsy (vs bGG 1 HR 2.97, 95CI 2.02-4.37), and brain (vs no HR1.48, 95CI 1.05-2.10), liver (vs no HR 2.18, 95CI 1.79-2.65) and lung metastases at diagnosis (vs no HR 1.33, 95CI 1.13-1.56) as predictive of PCSM.

Conclusions

Men presenting with PCa-bone metastatic disease have aggressive tumor biology and are at risk of PCSM in <3 years. Patients with aggressive prostate bGG disease presenting with bone and concomitant visceral metastasis (particularly liver metastasis) should be considered for early, aggressive systemic therapy and/or clinical trials.

Funding

None