O Abdel-Rahman1. 1. Clinical Oncology Department, Faculty of Medicine, Ain Shams University, Cairo, Egypt. Electronic address: omar.abdelrhman@med.asu.edu.e.
Abstract
AIMS: To develop and internally validate a new simplified model 'prostascore' to help predict the outcomes of treatment-naive patients with advanced prostate cancer. PATIENTS AND METHODS: Through the SEER*Stat program, the Surveillance, Epidemiology and End Results (SEER) database was queried for eligible records spanning the period from 2010 to 2013. The resulting group of patients was equally divided into two sets: training group (to guide model development) and validation group (to validate the model prediction). Multivariate analysis for the candidate prognostic factors (extent of extra-prostatic disease, prostate-specific antigen [PSA] level and grade) was conducted through a Cox proportional model. Prostascore was then calculated for each patient. Cancer-specific and overall survival analyses according to prostascore were conducted through Kaplan-Meier analysis/Log-rank testing. RESULTS: In total, 8727 patients with treatment-naive advanced prostate cancer and complete baseline data were identified in the period from 2010 to 2013. The following factors were associated with better cancer-specific survival in the training set (isolated regional nodal disease, lower PSA level and lower grade group; P < 0.0001). After assignment of a prostascore for each patient, cancer-specific survival was compared according to the score. Pairwise comparisons between all different scores were conducted. For cancer-specific survival evaluation according to the prostascore model, P values for pairwise comparisons among different score points were significant (P < 0.05) in the validation set. CONCLUSION: Prostascore is an easy and reliable tool for predicting the outcomes of patients with treatment-naive advanced prostate cancer. Further validation within the context of other treatment settings and population-based cohorts is recommended.
AIMS: To develop and internally validate a new simplified model 'prostascore' to help predict the outcomes of treatment-naive patients with advanced prostate cancer. PATIENTS AND METHODS: Through the SEER*Stat program, the Surveillance, Epidemiology and End Results (SEER) database was queried for eligible records spanning the period from 2010 to 2013. The resulting group of patients was equally divided into two sets: training group (to guide model development) and validation group (to validate the model prediction). Multivariate analysis for the candidate prognostic factors (extent of extra-prostatic disease, prostate-specific antigen [PSA] level and grade) was conducted through a Cox proportional model. Prostascore was then calculated for each patient. Cancer-specific and overall survival analyses according to prostascore were conducted through Kaplan-Meier analysis/Log-rank testing. RESULTS: In total, 8727 patients with treatment-naive advanced prostate cancer and complete baseline data were identified in the period from 2010 to 2013. The following factors were associated with better cancer-specific survival in the training set (isolated regional nodal disease, lower PSA level and lower grade group; P < 0.0001). After assignment of a prostascore for each patient, cancer-specific survival was compared according to the score. Pairwise comparisons between all different scores were conducted. For cancer-specific survival evaluation according to the prostascore model, P values for pairwise comparisons among different score points were significant (P < 0.05) in the validation set. CONCLUSION: Prostascore is an easy and reliable tool for predicting the outcomes of patients with treatment-naive advanced prostate cancer. Further validation within the context of other treatment settings and population-based cohorts is recommended.