Serum testosterone is not a correlate of prostate cancer lymph node involvement, but does predict biochemical failure for lymph node positive disease

Previously we found that serum testosterone (serum-T) correlated with the development of distant metastasis in patients with clinically localized prostate cancer treated with radiotherapy. In this report, the relationship of serum-T to lymph node positivity and to patient outcome for patients with regional lymph node involvement treated with androgen ablation alone was investigated. Serum-T was available in 514 of 854 men with clinically localized prostate cancer who underwent pelvic lymphadenectomy at M.D. Anderson Cancer Center between 1984 and 1993. Pretreatment prostatic acid phosphatase (PAP) and prostate specific antigen (PSA) were assayed in 98% and 95% of patients, respectively. Androgen ablation was achieved via orchiectomy or a luteinizing hormone releasing hormone agonist. Median follow-up was 66 months for the node positive subgroup (n = 92). Serum-T did not correlate with palpable stage, Gleason score, pretreatment PSA, or lymph node involvement. Age ⩽ 60 years and pretreatment PAP > 0.8 mU/ml correlated significantly with higher serum-T. In lymph node positive patients treated with androgen ablation, higher serum-T levels corresponded to both pretreatment PSA > 10 ng/ml and PAP > 0.8 mU/ml. Serum-T predicted for biochemical failure, but not metastatic relapse or overall survival. Actuarial 5-year biochemical failure rate was 73% for serum T > 500 ng/dl and 57% for serum-T ≤ 500 (p = 0.009). Multivariate analysis showed serum-T to be an independent correlate of rising PSA, both as a continuous (p = 0.001) or categorical (p = 0.037) variable. Serum-T did not significantly correlate with lymph node positivity, and therefore is not a marker for regional disease spread. However, serum-T was significantly associated with biochemical failure in node-positive patients treated with androgen ablation alone.