Title: Development and internal validation of nomograms predicting prostate cancer specific survival in patients with high risk disease. results of a multi-institutional analysis
Authors: Briganti, A
Spahn, M
Joniau, Steven
Marchioro, G
Tombal, B
Hsu, H
Bader, P
Graefen, M
Tizzani, A
Van Cangh, P
Gontero, P
Van Poppel, Hendrik #
Issue Date: Mar-2011
Publisher: Elsevier science bv
Host Document: European urology supplements vol:10 issue:2 pages:324-324
Conference: Annual EAU Congress edition:26 location:Vienna date:18-22 March 2011
Article number: 1048
Abstract: Introduction & Objectives: Several studies have shown that the outcome of
high risk prostate cancer (PCa is not invariably poor. However, no tool predicting
cancer specific survival (CSS) is currently available for this patient category. The
aim of this study was to develop and validate pre-and post-operative nomograms
predicting cancer specific survival (CSS) in patients with high risk PCa treated with
radical prostatectomy (RP) and pelvic lymph node dissection (PLND)
Materials & Methods: The study included 1645 patients with pre-operative high
risk PCa (defined as either PSA>20 ng/ml or cT3/cT4 or biopsy Gleason 8-10)
treated with RP and PLND at 8 European tertiary referral centers between 1987
and 2009. All patients had pre-operative data (including PSA at surgery, biopsy
Gleason sum, clinical stage) as well as post-operative pathological information
(including stage and Gleason sum, surgical and lymph node status, number of
nodes removed).Moreover, data regarding use of adjuvant treatments were
available for all patients. Univariable and multivariable Cox regression analyses
tested the association between predictors and CSS. Two different models were
fitted, including either pre-operative (PSA, biopsy Gleason sum, clinical stage) or
post-operative data (pathological stage and Gleason sum, surgical margin and
lymph node status, use of adjuvant therapies).Cox regression coefficient-based
pre- and post-operative nomograms were thus developed and internally validated
with 200 bootstrap resamples. The Harrell’s concordance index quantified the
accuracy of each nomogram
Results: Overall, 5,8 and 10-year CSS rates were 95.4, 92 and 89.8%,
respectively. At univariable Cox regression models all pre-operative variables were
significant predictor of CSS (all p≤0.007). At multivariable analyses clinical stage
and biopsy Gleason sum maintained a significant association with CSS (all p≤0.01)
while PSA did not (p=0.5). A nomogram including all the mentioned pre-operative
variables demonstrated a 71.5% accuracy. When using post-operative variables, a significant association was found between all covariates (pathological stages,
lymph node and surgical margin status, hormonal therapy; all p≤0.001), except
for adjuvant radiation therapy (p=0.5). At multivariable analyses, presence of
extracapsular extension, pathological Gleason sum and lymph node invasion were
independent predictors of CSS (all p<0.001). A nomogram based on the mentioned
post-operative variables was 79.6% accurate.
Conclusions: Long-term outcome of high risk patients is not invariably poor. We
developed and internally validated two accurate nomograms predicting CSS in
these patients based on either pre-operative or post-operative variables (AUC:
71.5 and 79.6%, respectively). Our predictive tools might be useful in selecting
those patients who might be suitable for a surgical, extensive approach
ISSN: 1569-9056
Publication status: published
KU Leuven publication type: IMa
Appears in Collections:Urology Section (-)
# (joint) last author

Files in This Item:

There are no files associated with this item.


All items in Lirias are protected by copyright, with all rights reserved.