cProgPisansky97

Cancer 79:2154-61, 1997

An Enhanced Prognostic System for Clinically
Localized Carcinoma of the Prostate

Thomas M. Pisansky, M.D.1, Michael J. Kahn, Ph.D.2, David G. Bostwick, M.D.3

1 Division of Radiation Oncology, Mayo Clinic and Mayo Foundation, Rochester,
Minnesota. 2 Cancer Center Statistics Unit, Mayo Clinic and Mayo Foundation,
Rochester, Minnesota. 3 Department of Laboratory Medicine and Pathology, Mayo
Clinic and Mayo Foundation, Rochester, Minnesota.

Received September 16, 1997; revision received February 10, 1997; accepted February27, 1997.

ABSTRACT

BACKGROUND. This investigation was conducted to develop an enhanced prognostic system based on readily available and independently predictive tumor-related factors for patients with clinically localized prostate carcinoma.

METHODS. The outcome of 500 patients treated solely with irradiation for clinical TNM classifications T1-4, N0 or NX, M0 prostate carcinoma was used to identify factors independently associated with disease relapse. Logistic regression constructed a risk score equation, and optimized cutoff points to characterize patient groups with low, intermediate or high risks for relapse were established with receiver operating characteristic curve analysis.

RESULTS. Clinical tumor stage (P < 0.00001), Gleason score (P = 0.0002), and pretherapy serum prostate specific antigen (P < 0.00001) were independently associated with clinical or biochemical relapse. These factors were included in a risk score equation that defined patient groups with a distinctly different outcome. For the low, intermediate, and high risk groups, the relapse-free probabilities at 5 years after irradiation were 92%, 67%, and 24%, respectively (P < 0.0001).

CONCLUSIONS. Readily available, pretherapy disease-related characteristics formed thebasis of an enhanced prognostic system for patients with clinically localized prostatecarcinoma. A multivariate prognostic system of this nature estimated patient prognosis in amore exacting fashion than a system exclusively based on anatomic factors.

Cancer 79:2154-61, 1997