Abstract
PURPOSE: To determine whether computed tomography/magnetic resonance imaging-based day 0 (d0) dosimetry is a meaningful predictor of day 21 (d21) dosimetry in low-dose-rate brachytherapy for localized prostate cancer. METHODS AND MATERIALS: The study population consisted of 277 men with localized (T1-2 N0 M0), low-/intermediate-risk prostate cancer treated with low-dose-rate brachytherapy. Computed tomography/magnetic resonance imaging fusion was used for postimplant dosimetry at d0 and d21. Logistic regression was used to construct receiver operating characteristic curves for achieving each constraint at d21, based on d0 D90 and V100, and Youden's index was used to evaluate cutpoints. Freedom from biochemical failure (FBCF) was estimated with the Kaplan-Meier method. RESULTS: The median d0 D90 increased from 133 to 150 Gy at d21, and median d0 V100 increased from 87% to 91%. For achieving the D90 constraint at d21, the optimal cut-point for d0 D90 was 135 Gy, with 84% of these patients maintaining a d21 D90 > 145 Gy. For achieving the D90 constraint at d21, the optimal cut-point for d0 V100 was 87%, with 83% of these patients maintained a d21 V100 > 90%. There was no improvement in FBCF in patients with a d0 D90 > 135 Gy or D90 > 145 Gy. Similarly, there was no improvement in FBCF in patients with a d0 V100 > 87% or V100 > 90%. CONCLUSIONS: Meeting dosimetric constraints on d0 does not obviate d21 dosimetric analysis. Constraints used for dose prescriptions on d0 are not the ideal predictors of d21 dosimetry.