Abstract
Abstract only
193
Background: Outcomes of patients (pts) undergoing CHT for mHSPC are heterogeneous, with some rapidly developing castration-resistance (CRPC). While AVPC ("anaplastic") features are described in CRPC, less is known in the mHSPC setting. In this multi-institutional cohort, we explored pre-treatment factors associated with poor outcome. Methods: De-novo mHSPC pts treated with CHT from June 2014 to July 2017 at The Ottawa Hospital Cancer Centre (TOHCC) and London Regional Cancer Centre (LRCC) were retrospectively identified. AVPC features (defined below) were collected and cumulatively scored (0, 1, or 2+), along with baseline, treatment and outcome data. Statistical comparisons utilized Cox regression analysis and Kaplan-Meier method for association with CRPC and survival. Results: 92 pts (58 TOHCC, 34 LRCC) met inclusion for study; 83 (90%) had "high-volume" disease (≥4 bone lesions; or ≥1 visceral metastasis), 69/73 (95%) prostate biopsies scored Gleason 8-10, and 55 (60%) had AVPC features: >5 cm nodal/pelvic mass (28), visceral metastases (21), lytic bone metastases (16), elevated LDH (12), low PSA (4), or neuroendocrine differentiation (2). Pre-docetaxel PSA fall of <50, 50-75, 75-90, and ≥90% baseline occurred in 12, 13, 19, and 56% respectively. Docetaxel was initiated a median of 66 days after androgen-deprivation, with 82% completing 5 or 6 cycles. 9 pts (10%) progressed during docetaxel, and 31 pts (34%) developed CRPC < 12 months. Median time to CRPC was 18.4, 14.0, and 11.0 months for 0, 1, and 2+ AVPC features (log rank p=0.009). CRPC was also associated with pre-docetaxel PSA fall <75% (p<0.001) and high alkaline phosphatase (p=0.016). At 18 months median follow-up, 16/55 (29%) with AVPC features have died versus 1/37 (3%) without (log rank p=0.001). In multivariable analysis, AVPC features and pre-docetaxel PSA fall were independently (p<0.05) associated with survival. Conclusions: In our study, mHSPC pts with AVPC features and suboptimal (<75%) pre-docetaxel PSA decline had poor prognosis. These features should be validated in larger cohorts to potentially identify mHSPC pts suitable for further study in clinical trials.