Login to Access Video or Poster Abstract: MP53-05
Sources of Funding: None

Introduction

Chemical castration (Total Testosterone, TT,<0,50 ng/mL) is common treatment in intermediate/high risk prostate cancer (PCa) adjuvant to radiotherapy, and in advanced/metastatic PCa._x000D_ After androgen deprivation therapy (ADT) we assume variability and delay until recovery over castration and/or eugonadic state (TT>3.5 ng/mL)._x000D_ We evaluate variability, associated factors, and the design of nomograms for TT recovery after ADT withdrawal._x000D_

Methods

Ambispective study on 205 patients after ADT cessation._x000D_ Predictive variables: age at initiation/cessation of ADT, biopsy/surgical specimen (in case) Gleason score, duration of ADT, primary therapy of PCa, and LHRH agonist. Result variables: Recovery of TT over castrate level and eugonadic level. Univariate analysis: Kaplan-Meier curves (log rank test). Multivariate models are built by Cox proportional hazards model._x000D_ The calibration and the discrimination ability of the model, and probability density functions and clinical utility curves are evaluated._x000D_

Results

The median biochemical and clinical follow-up are 27 (P25-75:15-39.5? Range: 1-98) and 39 months (P25-75:_x000D_ 28-51? Range: 1-107 months). We find high individualised variabilty in TT recovery._x000D_ - Recovery over castration levels: 25% of patients do not recover. Our intervals of recovery for 25, 50, 75 and 100% of recoverer patients are 4, 7, 10 and 42 months._x000D_ - Recovery over eugonadic levels: 84% of patients do not normalize TT. After 12 months of ADT withdrawal only 8% of patients recovered normal levels, the maximum rate of recovery is at 32 months, no one patient recovered later on._x000D_ In multivariate analysis "duration of ADT" and "age at ADT withdrawal" are significant predictors (H.R:0.69,p=0.0002 and HR:0.44,p<0.0001, for 0.5-Recovery? and HR:0.45,p=0.0006 and HR:0.39,p=0.0002, for 3.5-Recovery)._x000D_ We build two nomograms of recovery at 1, 2 and 3 years, with light overestimation for intermediate values in calibration analysis and a discrimination capacity ('c index') of 0.709 and 0.723,_x000D_ with AUC of 0.778, 0.813 and 0.805, and 0.707, 0.788 and 0.811 at 1, 2, and 3 years, respectively._x000D_

Conclusions

We confirmed a high variability in TT recovery after ADT withdrawal._x000D_ We found as the most relevant predictors "duration of ADT" and "age at ADT withdrawal"._x000D_ We obtained accurate calibrated and discriminative predictive nomograms of castration and eugonadic recovery._x000D_ Those models would allow us to estimate and counsell functional recovery after adjuvant therapy, and to keep castration after ADT withdrawal or intermittency in a more cost-effectiveness practice._x000D_

Funding

None