Concato J (West Haven VAMC)
Jain D (Yale University)
Li WW (Angiogenesis Foundation)
Wells CK (West Haven VAMC)
Prostate cancer exhibits a wide spectrum for the risk of death at the time of diagnosis; some patients have indolent disease (with normal life expectancy), whereas other patients have aggressive disease (dying within 1-2 years). Importantly, clinicians currently have very limited ability to predict survival in individual patients, and treatment decisions are therefore often difficult. The objective of this research was to determine whether new laboratory-based (e.g., angiogenic) factors provide additional prognostic information—beyond traditional clinical (e.g., anatomic and histologic) factors—regarding mortality among men with prostate cancer.
Using a community-based population (N=71,661) of veterans, 1386 men with incident prostate cancer during 1991-1995 were identified. The current (interim) analysis randomly sampled 125 men who lived, and 125 men (matched by age, date of diagnosis, and site) who died, as of 12/31/99. Among 228 patients (91%) with complete data, we determined the independent impact of candidate prognostic factors (re: the intact patient or the tumor biopsy) on overall mortality during 8-13 years of follow-up. Baseline characteristics included demographic and clinical factors (e.g., comorbidity); conventional measures of anatomic spread, histologic grade, and serum prostate-specific antigen (PSA) level; as well as novel angiogenic (CD61), tumor suppressor (p53), and apoptosis-related (bcl-2) factors, along with microvessel density. Clinical data were obtained via medical-record review; laboratory data were obtained using standard immunohistochemical techniques. Proportional hazards analysis generated multivariable hazard ratios (HRs) and 95% confidence intervals (CIs).
The median age was 72 years; 6% were African-American; 66% had moderate-severe comorbidity. Most cancers (85%) were localized; the median Gleason score was 7; and the median PSA value was 10.4 ng/mL. Mortality as of 12/31/03 was 74% (168/228). In multivariable analysis, severe comorbidity (HR=2.1; 95% CI 1.8-2.5), non-localized tumor (HR=1.5; 95% CI 0.96-2.4), and poor histology (HR=1.8; 95% CI 1.1-3.1), were associated (as expected) with mortality. In addition, increased bcl-2 (HR=2.2; 95% CI 1.3-3.7) and higher microvessel density (HR=1.5; 95% CI 1.0-2.2) were independently predictive of death.
Among men with prostate cancer, laboratory-based factors obatined at biopsy provide independent prognostic information.
Our results advance understanding of the wide spectrum of aggressiveness of prostate cancer, and can be used in clinical care to improve treatment decisions and in research to stratify patients in clinical trials.