External validity of the Stephenson nomogram predicting the outcomes of prostate cancer patients treated salvage radiotherapy after radical prostatectomy: The importance of genomic data

Document Type

Conference Proceeding

Publication Date


Publication Title

European Urology Open Science


Introduction & Objectives: Salvage radiation therapy (SRT) is commonly offered to patients with prostate cancer (PCa) who have biochemical recurrence (BCR) after radical prostatectomy (RP). One of the most commonly used tool to predict the oncological outcomes of such individuals is the updated Stephenson nomogram. However, the latter lack external validity data, and doesn't account for the genomic profile of PCa. The goal of our study was to externally validate the updated Stephenson nomogram across different Decipher risk profiles.

Materials & Methods: Patients undergoing SRT for biochemically recurrent pN0 PCa post-RP (defined as PSA >=0.2ng/ml) at six academic centers across the US between 1990-2010 were included. Data for all available demographic, pathologic and oncologic outcomes were compared across different Decipher risk profile (low; Decipher score <0.45, intermediate 0.45-0.6, and high >0.6). The Stephenson nomogram was externally validated using c-index and calibration curves, using 5-year BCR-free survival as end-point.

Results: Median (interquartile range [IQR]) age, pre-sRT PSA and Decipher score was 60 (57-64) years, 0.43 (0.2-1.4)ng/mL and 0.43 (0.28-0.57) respectively; 86% had pathological Gleason >7, 78.3% had >pT3a disease and 78.2% had positive margins. Androgen deprivation therapy was given to 45 (36.3%) patients. The proportion of patients with low, intermediate and high risk Decipher scores was 65 (52.4%), 35 (28.2%) and 24 (19.4%); the median (IQR) pre-sRT PSA in these groups was 0.4 (0.2-0.7), 0.5 (0.2-1.9) and 0.8 (0.4-2.5) ng/mL respectively (p=0.03). Overall and stratified BCR-free survival is shown in Figure 1a/1b (median (IQR) follow-up of 71.4 [35.6-133.7] months). The c-index of Stephenson nomogram was 0.59 for the overall cohort, and 0.46, 0.53 and 0.73 for low, intermediate and high-risk Decipher respectively. Likewise, the calibration curves improved with higher Decipher risk groups.

Conclusions: The performance of the widely used Stephenson nomogram to predict outcomes of sRT was dependent on Decipher risk score, with superior accuracy in patients with higher genomic risk scores. This information can aid both patients and physicians in shared-decision making for sRT for biochemically recurrent PCa post-RP.



First Page


Last Page


This document is currently not available here.