Source
Department of Biostatistics and Applied Mathematics, The University of Texas M.D. Anderson Cancer Center, Houston, TX 77030, and Department of Radiation Oncology, William Beaumont Hospital, Deroit, MI, USA. hdt@odin.mdacc.tmc.edu
Abstract
PURPOSE:
The aim of this study was to investigate effect of increasing dose on risk groups for clinical failure (CF: local failure or distant failure or hormone ablation or PSA>or=25 ng/ml) in patients with T1-T2 prostate cancer treated with external beam radiotherapy.
METHODS AND MATERIALS:
Patients (n=4,537) were partitioned into nonoverlapping dose ranges, each narrow enough that dose was not a predictor of CF, and risk groups for CF were determined using recursive partitioning analysis (RPA). The same technique was applied to the highest of these dose ranges (70-76 Gy, 1,136 patients) to compare risk groups for CF in this dose range with the conventional risk-group classification.
RESULTS:
Cutpoints defining low-risk groups in each dose range shifted to higher initial PSA levels and Gleason scores with increasing dose. Risk groups for CF in the dose range 70-76 Gy were not consistent with conventional risk groups.
CONCLUSIONS:
The conventional classification of risk groups was derived in the early PSA era, when total doses<70 Gy were common, and it is inconsistent with risk groups for patients treated to doses>70 Gy. Risk-group classifications must be continuously re-examined whenever the trend is toward increasing total dose.