Purpose: We defined the efficacy of radiotherapy for the treatment of
high grade (Gleason scores 8 to 10) adenocarcinoma of the prostate. Ma
terials and Methods: A total of 50 patients underwent radiotherapy wit
h curative intent for clinically localized prostate cancer with Gleaso
n scores of 8 to 10 at 1 of 4 facilities affiliated with the Universit
y of California San Francisco. Patients were considered to have bioche
mical failure if they had a significant increase in prostate specific
antigen (PSA) of 0.5 ng./ml. per year, an increase in PSA to greater t
han 1.0 ng./ml. or a positive biopsy. Results: Among the 50 patients m
edian PSA was 22.7 ng./ml. (range 1.3 to 93.4). Tumors were clinical s
tage T1 or T2 in 46% of the cases and stage T3 or T4 in 54%. The overa
ll actuarial probability of freedom from biochemical failure at 4 year
s was 23%. In a multivariate analysis including all patients pretreatm
ent PSA was the only predictor of PSA failure, with 64% free of progre
ssion if the pretreatment PSA was 10 ng./ml. or less compared to only
16% at 3 years if PSA was more than 10 ng./ml. (p = 0.01). In a multiv
ariate analysis restricted to patients with PSA less than 20 ng./ml. 8
3% of those treated to more than 71 Gy. were free of progression compa
red to 0% for those treated to less than 71 Gy. (p = 0.03). In a multi
variate analysis PSA 10 ng./ml. or less (related risk 11.4, p = 0.02),
T stage 1 or 2 (relative risk 3.8, p = 0.05) and radiation dose more
than 71 Gy. (relative risk 4.0, p = 0.06) were associated with a favor
able outcome. Conclusions: At 4 years the freedom from PSA failure fol
lowing radiotherapy for high grade prostate cancer was comparable to p
reviously reported surgical series. The high failure rate among patien
ts with PSA greater than 20 ng./ml. suggests that these patients shoul
d be considered for investigational approaches. The apparent improveme
nt in freedom from progression with the use of higher doses provides r
eason for optimism.