Defining the appropriate radiation dose for pretreatment PSA <= 10 ng/mL prostate cancer

Purpose: To investigate whether a dose response exists for biochemical no e vidence of disease (bNED) control in prostate cancer patients with pretreat ment prostate-specific antigen (PSA) less than or equal to 10 ng/mL and to identify the patient subgroups affected. Methods and Materials: Between 5/89 and 10/97, 488 T1-T3 NX-0 M0 prostate c ancer patients with PSA less than or equal to 10 ng/mL were treated with th ree-dimensional conformal radiation therapy (3D-CRT) alone. Median and mean pretreatment PSA values were 6.3 and 6.2, respectively. Gleason scores of 2-6 and 7-10 were noted in 386 and 102 men, respectively. AJCC 1992 palpati on T1-T2AB tumors were noted in 415 patients. Perineural invasion (PNI) was noted in 60 men. Mean and median age was 67 and 68 years, respectively. Do se to the center of the prostate ranged from 6260 cGy to 8409 cGy with a me an and median of 7423 cGy and 7278 cGy, respectively. Patients were stratif ied into three groups according to dose: <7250 cGy, 7250-7599 cGy, and grea ter than or equal to 7600 cGy. Median dose in these three groups was 7067 c Gy, 7278 cGy, and 7734 cGy, respectively. Univariate analysis was performed to determine differences in bNED control (American Society for Therapeutic Radiology and Oncology [ASTRO] Consensus Guidelines definition of Failure) by dose group for the entire cohort, for 310 good prognosis patients (T1-T 2A, Gleason score 2-6, absence of PNI), and for 178 poor prognosis patients (T2B-T3 or Gleason score 7-10 or presence of PNI) (1). Multivariate analys is (MVA) was performed to determine if dose was an independent predictor of bNED control. Median follow-up was 36 months. Results: A dose response was not demonstrated for the entire group of patie nts with pretreatment PSA less than or equal to 10 ng/mL. Doses of <7250 cG y, 7250-7599 cGy, and greater than or equal to 7600 cGy were associated wit h 5-year bNED control rates of 73%, 86%, and 89%, respectively (p = 0.12). MVA demonstrated prognosis group (p = 0.038) to be the only independent pre dictor of bNED control. Good prognosis patients had a 5-year bNED of 85% an d no dose response was seen. The subgroup of poor prognosis patients demons trated a 5-year bNED control rate of 81% and a dose response was seen for t hose receiving greater than or equal to 7600 cGy, compared to the two lower dose groups (94% vs. 75% vs. 70%; p = 0.0062). MVA for the poor prognosis subset demonstrated dose (p = 0.01) to be the only independent predictor fo r improved bNED control. Conclusions: The poor prognosis subset of PSA 10 ng/mL prostate cancer pati ents benefit from dose escalation. A dose response Is not demonstrated for prostate cancer patients with pretreatment PSA less than or equal to 10 ng/ mL and other favorable features. (C) 2000 Elsevier Science Inc.