Login to Access Video or Poster Abstract: MP70-08
Sources of Funding: none

Introduction

The purpose of this study is to create a pre-treatment and post-treatment model for prediction of local recurrence in men who underwent primary full gland cryoablation of the prostate.

Methods

We studied 961 patients from the Cryo On-Line Data (COLD) registry who underwent primary full gland cryoablation for prostate cancer and subsequently received post-treatment prostate biopsy. Pre- and post-treatment variables of interest were identified and fit in a Cox regression model to predict time to positive biopsy using backward stepwise elimination of variables, knots and interactions of p>0.1 to accept the model with maximum AIC. The pre-treatment and post-treatment regression models were each internally validated and calibrated using 500 bootstraps and then constructed into nomograms.

Results

Mean age was 69.2 (± 7.9) years old. The median pre-treatment PSA was 6.8 ng/ml (IQR 4.6-9.8). Biopsy prognostic grade group (PGG) ranged from 1 to 5 in 70%, 12%, 6%, 8% and 3% of men respectively and 14% had a clinical stage ≥cT3. Post-treatment biopsy was positive in 28.2% of patients. The pre-operative nomogram included: PSA, PGG, ≥cT3 and use of neoadjuvant androgen deprivation therapy (nADT) (Figure 1). The post-operative nomogram included: PGG, ≥cT3, nADT, PSA nadir, biochemical recurrence (BCR, Phoenix), time till undetectable PSA, time on undetectable PSA and various interaction terms (Figure 2). The pre-operative nomogram achieved a corrected C-index of 61% while the post-operative nomogram achieved a corrected C-index of 71%.

Conclusions

We developed the predictive models for time to local cancer recurrence before and after primary full gland cryoablation of the prostate. Addition of post-treatment factors appears to improve predictive ability. These models can guide the urologist in selecting the appropriate candidate for cryoablation and determining the intensity of post-treatment follow-up._x000D_ _x000D_

Funding

none