Quality of life outcomes for patients with metastatic castration-resistant prostate cancer and pretreatment prognostic score.

BACKGROUND: A prognostic risk score (Halabi score) in metastatic castration-resistant prostate cancer (mCRPC) accurately predicts overall survival, but its association with quality of life (QOL) has not been defined. We hypothesize that a higher pretreatment Halabi score is associated with worse QOL outcomes over time in mCRPC patients. METHODS: Patient-level data from the docetaxel plus prednisone control arm of Mainsail, a Phase 3 clinical trial in mCRPC were accessed via ProjectDataSphere. Pretreatment Halabi score included disease-related factors: metastatic site, opioid use, Eastern Cooperative Oncology Group performance status (ECOG-PS), alkaline phosphatase, albumin, hemoglobin, lactic acid dehydrogenase, and PSA, with higher score indicating worse survival. Three QOL scales were created: Functional Assessment of Cancer Therapy-Prostate (FACT-P, higher score = better QOL), Brief Pain Inventory-Short Form Severity score (BPI-SFSS, higher score = higher pain severity), and BPI-SF Interference score (BPI-SFIS, higher score = greater pain interference). Mixed linear model was used to estimate the associations between Halabi score and QOL scores assessed at different time points (baseline, 2 months, and 6 months). RESULTS: This analysis included 412 mCRPC patients (median age = 68 years, 82% white, 5% Black, median log PSA = 4.4 ng/mL). After multivariable adjustment, Halabi score was significantly associated with QOL scores at all time points. At 6 months, multivariable adjusted FACT-P decreased by 10.0 points (worsening), BPI-SFSS increased by 0.8 points (worsening), and BPI-SFIS increased by 0.9 points (worsening) for each unit increase in Halabi risk score. In multivariable analysis of individual components, ECOG-PS, site of metastasis, and opioid use were significantly associated with worse QOL scores at baseline. CONCLUSIONS: Chemotherapy-naïve mCRPC patients with poorer Halabi prognostic risk scores have poorer QOL and greater pain intensity and interference at baseline and during follow-up.

Duke Scholars

Author Susan Halabi Biostatistics & Bioinformatics, Division of Biostatistics