Evaluation of Psa Patterns in Patients with Localized Prostate Cancer Treated with Radiation Therapy and Androgen Deprivation Therapy
T. Yang,X. Pei,Z. Zhang,M. Kollmeier,D. Spratt,Z. Zumsteg,W. Polkinghorn,S. McBride,M. Zelefsky
DOI: https://doi.org/10.1016/j.ijrobp.2014.05.1353
2014-01-01
Abstract:PSA kinetics following androgen deprivation therapy (ADT) and its relationship with testosterone re-normalization is complex and poorly understood. The purpose of this study is to characterize post-ADT PSA kinetics and its correlation with oncologic outcomes. Between 1991 and 2011, 795 patients with localized prostate cancer were treated with RT and neoadjuvant, concurrent ± adjuvant ADT, all with ≥3 PSA measurements within the first 2 years after ADT or before salvage therapy. The median follow-up was 101 months (13-204 months). Three distinct PSA patterns were identified: 1) stable trend (ST), PSA slope that corresponded to a double time (DT) >30 months; 2) slow rise trend (SR), 7-30 months; 3) rapid rise trend (RR), ≤6 months. Testosterone re-normalization slope (T-DT) was computed in 428 patients who had ≥2 post-ADT testosterone measurements within the first year. The relationship between PSA-DT and T-DT (dichotomized at 6 months) was assessed using Spearman rank correlation and Fisher's exact test. Chi-square test was performed to assess clinical factors and PSA trend association. Distant metastasis (DM) rates were compared using log-rank test. Significant clinical predictors for cause specific death (CSD) and DM were determined using competing risks regression model and Cox regression model. ST, SR, and RR kinetics were observed in 360 (45%), 318 (40%), and 117 (15%) patients, respectively. The median PSA level at 2-year for each PSA trend was, ST = 0.22 (range, 0-3.7), SR = 0.33 (range, 0-8.2), RR = 0.54 (range, 0.1-25.8). More NCCN high-risk group patients had RR PSA trend (44%) compared to intermediate risk (35%) and low risk (22%, p = 0.009) patients. Rate of testosterone re-normalization was significantly associated with PSA slopes (p<0.001, Rs = 0.30), indicating testosterone re-normalization was more rapid (T-DT ≤ 6 months) in RR patients compared to patients with SR or ST PSA trends. 21% RR patients developed DM, while only 8% of ST and 6% of SR trend patients developed DM (p<0.001). On univariate analysis, RR kinetics was associated significantly with CSD (HR = 4.8, p<0.001) and DM (HR = 3.2, p<0.001). After correcting for duration of ADT, RT modality, cT stage, pre-treatment PSA, and Gleason score, RR kinetics remained to be predictive of DM on multivariate analysis (HR = 3.6, p<0.001). For men with localized prostate cancer undergoing RT and ADT, we identified 3 distinct categories of PSA kinetics post-ADT. A rapid re-normalization of testosterone was associated with a rapid rise in PSA, which was associated with an increased likelihood for DM and prostate cancer mortality. Further evaluation of these kinetics patterns is necessary, but may serve as a method for early intervention with second-generation ADT for men with rapidly rising post-ADT PSA.