Prognostic features in men who died of prostate cancer

Since the advent of widespread prostate specific antigen (PSA) based screening in the United States, the risk of over diagnosis as well as delayed diagnosis due to existing PSA thresholds has become a concern. Treatment decision planning is generally linked to prognostic variables, most notably PSA, clinical stage and Gleason grade. We examined these and other prognostic variables in a cohort of men who ultimately died of prostate cancer.

Of 413 men with prostate cancer in a cohort in San Antonio, Texas between 1993 and 2003 who died of disease we identified 211 who died as a direct result of prostate cancer. In these cases we assessed presenting symptoms, PSA history, tumor stage and Gleason score at diagnosis.

Of the 211 men 141 (67%) underwent screening for prostate cancer prior to diagnosis. Of 190 men with PSA data at diagnosis available 7 (3.7%) had PSA less than 4.0 ng/ml and 27 (14%) had PSA 4.0 to 10.0 ng/ml. Clinical stage distribution was cT1 in 21.1% of cases, cT2 in 50.7% and cT3 in 26.8%. Of 167 men for whom biopsy Gleason grade was available 16.8%, 16.2%, 24% and 43.1% had Gleason sum 5 or less, 6, 7 and 8 or higher, respectively.

While most men who ultimately die of prostate cancer have poor prognostic features, a substantial number have features associated with a potentially good prognosis, including low PSA and low Gleason grade. Many men who died of prostate cancer had undergone prior screening. These data demonstrate the need for improved markers of prognosis and continued assessment of the efficacy of screening with PSA.