Perineural versus lymphovascular invasion: Which is a better marker for unfavorable biochemical outcomes following prostatectomy? Results from the Duke Prostate Center Database.

Published

Journal Article

34 Background: Despite a substantial effort to study the clinical significance of perineural invasion in prostate biopsies (PNIb) its prognostic significance remains controversial. The limited data on perineural invasion in radical prostatectomy (RP) specimens (PNIp) suggests that biopsy specimens inadequately represent whole gland pathology thus rendering investigation of PNIp even more clinically relevant. While lymphovascular invasion in pathology specimens (LVIp) has been more rigorously investigated, studies have shown conflicting results. More importantly, there has been minimal comparison of PNIp and LVIp in the same cohort to determine which marker is the superior prognostic factor. METHODS: We retrospectively analyzed data from 1611 men who underwent RP from 1999 to 2010 from the Duke Prostate Center database. We evaluated PNIp and LVIp as predictors of time to BCR by comparing hazard ratios (HR) and 95% confidence intervals (CI) using crude and adjusted proportional hazards regression models that included both variables and controlled for demographic and clinico-pathological covariates. RESULTS: A total of 1304 (81%) men had PNIp while only 82 (5%) men had LVIp. On crude regression, both PNIp (HR=3.39; 95% CI=1.94-5.84; p<0.001) and LVIp (HR=2.33; 95% CI=1.49-3.64; p<0.001) were significant predictors of adverse BCR risk. After adjusting for clinico-pathological covariates, PNIp remained significantly associated with increased BCR risk (HR=1.85; 95%CI =1.04-3.31; p=0.04). Specifically, men with PNIp were 85% more likely to experience BCR relative to PNIp (-) men. In contrast, LVIp was not independently associated with BCR risk (p=0.23). CONCLUSIONS: In a cohort of men who underwent RP, PNIp is predictive of adverse BCR outcomes independent of clinicopathological parameters that include LVIp. Consequently, LVIp is a poor predictor of BCR risk. PNIp may thus provide additional prognostic information for men treated with RP and its inclusion in predictive nomograms requires study. Further analyses to determine if PNIp is likewise associated with metastasis and cancer-specific survival are warranted.

Full Text

Duke Authors

Cited Authors

  • Singh, AA; Banez, LL; Gerber, L; Freedland, SJ; Robertson, CN; Ferrandino, M; Polascik, T; Walther, PJ; Moul, JW

Published Date

  • February 10, 2012

Published In

Volume / Issue

  • 30 / 5_suppl

Start / End Page

  • 34 -

PubMed ID

  • 28142904

Pubmed Central ID

  • 28142904

Electronic International Standard Serial Number (EISSN)

  • 1527-7755

Digital Object Identifier (DOI)

  • 10.1200/jco.2012.30.5_suppl.34

Language

  • eng

Conference Location

  • United States