Composite biologically effective dose volume histograms. (a) 11 biochemically failed patients and a matched control group of 22 patients. The total BEDs for the failed group were and for the controls . (b) 44 failed patients and a matched control group of 88 patients. The total BEDs for the failed group were and for the controls .
Histograms as a percentage of patients in the biochemical failure and nonfailure [biological NED (bNED)] groups. The distribution is bimodal with peaks near 110 and , which correspond to the mean BEDs for mono-and combined modality therapy patients, respectively.
Correlation between implant BED and , both expressed in gray. The 55 biochemical failures are marked with filled diamonds, while the 110 matched control nonfailures are marked with filled circles. The linear regression line has an coefficient of 0.753.
Correlation between implant BED and the generalized , both expressed in gray and stratified by radionuclide. The 33 patients implanted with are denoted with filled circles, while the 132 patients implanted with are denoted with filled diamonds. The linear regression lines have an coefficient of 0.991 and 0.986 for and , respectively.
Radiobiological parameters used in calculations from the AAPM Task Group 137 recommendations based on the results of Wang et al. (Refs. 6 and 8).
Demographic, clinical, and treatment parameters for the study population stratified by biochemical failures and nonfailures matched 2:1 on the basis of risk group, radionuclide, prescribed brachytherapy dose, and closest proximity in time to the biochemically failed patient.
Dosimetric and radiobiological parameters from a full DVH recalculation for biochemically failed patients compared to nonfailures matched 2:1 and stratified by monotherapy and combined modality therapy.
Predictors for biochemical failure in univariate and multivariate analyses in the detailed dosimetry cohort of biochemical failures and patients matched 2 to 1 on the basis of risk group, radionuclide, prescription dose, and date of implant. Conditional logistic regression analysis was used to factor out the matching variables and stratify the cohort into each failed patient and his controls.
Significant Pearson bivariate correlation coefficients for updated dosimetric and radiobiological parameters. Significant were defined as those with values . Correlation coefficients are highlighted in bold italic.
Article metrics loading...
Full text loading...