PURPOSE: To investigate the incremental value of magnetic resonance (MR) imaging in addition to clinical variables for the prediction of pathological outcomes and disease recurrence in clinically high-risk prostate cancer.
MATERIALS AND METHODS: In all, 922 consecutive patients with clinically high-risk prostate cancer underwent MR imaging prior to radical prostatectomy (RP). Multivariate logistic regression and Cox proportional hazards models with clinical variables only, or in combination with MR imaging data, were created to predict pathological outcomes and biochemical recurrence. Models were compared using receiver operating characteristic (ROC) curves and Harrell's concordance index (c-index).
RESULTS: The proportion of patients with pathological extracapsular extension (ECE), seminal vesicle invasion (SVI), and lymph node metastasis was 57.5%, 12.7% and 6.3%, respectively. The sensitivity and specificity of ECE, SVI and lymph node metastasis detection were 43% and 84.2%, 34.9% and 93.8%, and 14.0% and 96.9%, respectively. For the prediction of ECE and SVI, the area under the ROC curve of the model with clinical variables and MR imaging data was larger than that of the model with clinical variables alone (Ece: 0.734 vs. 0.697, p=0.001; SVI: 0.750 vs. 0.698, p< 0.001, respectively). The 5 year biochemical recurrence-free survival rate was 56.1%. For the prediction of biochemical recurrence, the c-index of the multivariate model with clinical variables only or clinical variables with MR imaging data was 0.563 and 0.599, respectively (p= 0.003).
CONCLUSIONS: MR imaging findings have incremental value in addition to clinical variables in the prediction of pathological outcomes and disease recurrence.
Written by:
Jeong IG, Lim JH, You D, Kim MH, Choi HJ, Kim JK, Cho KS, Hong JH, Ahn H, Kim CS. Are you the author?
Department of Urology, Asan Medical Center, University of Ulsan College of Medicine, Seoul, Korea.
Reference: J Urol. 2013 Jun 18. pii: S0022-5347(13)04632-6.
doi: 10.1016/j.juro.2013.06.035
PubMed Abstract
PMID: 23791890
UroToday.com Prostate Cancer Section
