ABSTRACT
RISK STRATIFICATION MODELS FOR PROSTATE CANCER
The American Joint Cancer Center (AJCC) TNM staging
system (Table 1A) has limited utility in predicting out-
come and directing therapy, particularly in men with
intermediate risk disease (2). Clinical staging is insuffi-
ciently accurate to rely on as the sole determinant of
treatment approach; clinical understaging is reported in up
to 60% of cases. Several independent prognostic factors
other than AJCC stage including patient age, Gleason
score, PSA at diagnosis, PSA velocity during two years
before diagnosis, the number of positive biopsy cores, and
the percentage of tumor within the positive biopsy cores
have been demonstrated to influence outcome. The
Gleason score appears to correlate with disease extent
and prognosis and is the single best predictor of biologic
activity and tumor stage (3). Higher Gleason scores are
associated with increased probability of disease extension
to the prostate capsule, seminal vesicles, lymph nodes, and
distant sites. Partin and others developed series of nomo-
grams, mostly using Gleason score, PSA level, and T
stage to predict the probability of extracapsular extension
(ECE), seminal vesicle invasion (SVI), and lymph node
involvement (LNþ) as well as to predict treatment outcomes, including PSA failure and survival. Several risk
classification schemes have been proposed based on
grouping of patients with different prognostic features
and similar clinical outcomes (4-10). Many physicians
have adopted these simplified approaches to make
treatment recommendations.