Validation of postoperative nomograms in prostate cancer patients with long-term follow-up

Gregory P. Swanson, Changhong Yu, Michael W. Kattan, Michael R. Hermans

Research output: Contribution to journalArticlepeer-review

19 Scopus citations

Abstract

Objective: To validate established postoperative nomograms in a patient population with long-term follow-up. There are well-documented risk factors that can help predict for failure after radical prostatectomy for prostate cancer. When combined together, the predictive power is greater than any single factor. It now appears that the most powerful tool in this regard is the nomogram because it considers the various factors as continuous variables. Materials and Methods: The original postoperative nomogram was developed in 1999 and modified and updated in 2005. A slightly modified version was published in 2009. We evaluated these nomograms against a large (n = 715) postradical prostatectomy patient population with long-term follow-up (median 9.4 years). Results: The concordance index for the 1999, 2005, and 2009 nomograms was 0.7831, 0.7680, and 0.7859. From plots of the calibration curves in each nomogram, the nomogram underestimated the actual 10-year biochemical recurrence risk in our patient population. Conclusions: Nomograms dynamically incorporate various risk factors and account for their interrelationships. Our findings confirm that the nomograms are predictive in an external patient population with an accuracy of >70%. The predictive ability of nomograms will be improved with the development of other predictive factors, which will most likely occur through the development of molecular markers.

Original languageEnglish (US)
Pages (from-to)105-109
Number of pages5
JournalUrology
Volume78
Issue number1
DOIs
StatePublished - Jul 1 2011

ASJC Scopus subject areas

  • Urology

Fingerprint

Dive into the research topics of 'Validation of postoperative nomograms in prostate cancer patients with long-term follow-up'. Together they form a unique fingerprint.

Cite this