Al. Hanlon et Ge. Hanks, Failure pattern implications following external beam irradiation of prostate cancer: Long-term follow-up and indications of cure, CANCER J, 6, 2000, pp. S193-S197
The purpose of this study was to present patterns and risk of biochemical f
ailure following external beam irradiation of prostate cancer and to make c
omparisons to a published modern radical prostatectomy series. Between Janu
ary 1987 and December 1994, 328 men were treated definitively at Fox Chase
Cancer Center for localized prostate cancer using conventional or three-dim
ensional conformal radiotherapy. The median biochemical follow-up was 6.4 y
ears, with all patients having at least 5 years follow-up. Two prognostic p
atient groups were established on the basis of proportional hazards modelin
g that considered treatment and presenting tumor characteristics. For each
of the two prognostic groups, biochemical failure and hazard functions were
estimated using the ASTRO consensus definition of failure and life table m
ethodology. Failure risk comparisons were made to modern published radical
prostatectomy series.
Multivariate analysis demonstrated the independent predictive power of pret
reatment PSA level, palpation stage, Gleason score, and dose. Thus, the fav
orable prognosis group, Group I, consisted of 83 patients who were treated
with a dose level greater than or equal to 74 Gy and who presented with PSA
levels < 20 ng/ml, T1/T2A tumors, and Gleason score 2-6. Group II consiste
d of 245 patients with at least one of the following: pretreatment PSA leve
l greater than or equal to 20 ng/ml, T2B/T3 tumor, Gleason score 7-10, dose
< 74 Gy. The 5- and 8-year bNED estimates were 76% and 76% for Group I, an
d 51% and 49% for Group II. Only three failures occurred after 5 years, all
from Group II, representing 2% of the total failures observed. Hazard func
tion estimates indicate maximum risk of failure at 24 to 36 months, taperin
g to a low rate at 4 years with no failures observed after 4 years. Differe
nces in patterns of failure by prognostic group show maximum risk of failur
e at 24 months (median, 31 months) for Group I, and 12 to 36 months (median
, 22 months) for Group II. Group II reaches low levels of risk at 6 years,
in contrast to 4 years for the patients with a more favorable prognosis. We
concluded that patients treated with external beam radiation alone show li
ttle risk of failure after 4 to 6 years. This result suggests that the 5-ye
ar bNED control rate approximates the eventual cure rate of prostate cancer
.