Downloaded from http://jcp.bmj.com/ on April 3, 2015 - Published by group.bmj.com

JCP Online First, published on March 11, 2015 as 10.1136/jclinpath-2014-202767 Original article

Prostate biopsy concordance in a large populationbased sample: a Surveillance, Epidemiology and End Results study David Schreiber,1,2 Andrew T Wong,1,2 Justin Rineer,3 Jeremy Weedon,2 David Schwartz1,2 1

Department of Veterans Affairs, New York Harbor Healthcare System, Brooklyn, New York, USA 2 SUNY Downstate Medical Center, Brooklyn, New York, USA 3 UF Orlando Health, Orlando, Florida, USA Correspondence to Dr David Schreiber, Department of Veterans Affairs, New York Harbor Healthcare System, 800 Poly Place, Suite 114A, Brooklyn, NY 11209, USA; David. [email protected] Received 14 November 2014 Revised 13 February 2015 Accepted 22 February 2015

ABSTRACT Aims To use the Surveillance, Epidemiology and End Results database in order to evaluate prostate biopsy concordance in a large population-based sample. Methods We identified 34 195 men who were diagnosed with prostate cancer and underwent a radical prostatectomy from 2010 to 2011. All patients also had to have both clinical and pathological Gleason scores available for analysis. The concordance of the biopsy Gleason score to the pathological Gleason score was analysed using the coefficient of agreement (κ). Univariate and multivariate logistic regression analyses were performed to determine potential factors that may impact concordance of Gleason score. Results Overall, the clinical and pathological Gleason scores matched in 55.4% of patients. The concordance rates were 55.3% for Gleason 6, 66.9% for Gleason 3 +4, 42.9% for Gleason 4+3 and 24.8% for Gleason 8, with frequent downgrading to Gleason 7. The κ for Gleason score concordance was 0.36 (95% CI 0.35 to 0.37), indicating fair agreement. The weighted κ for Gleason score concordance was 0.51 (95% CI 0.50 to 0.52), indicating moderate agreement. Additionally, the Bowker tests of symmetry were highly significant ( p0.8 are consistent with excellent agreement.13 Additionally, in order to assess for differences in agreement based on the biopsy score, we analysed the weighted κ and performed the Bowker test of symmetry. Subgroup analysis performed for men was coded by SEER as having a ≥12 core biopsy. Univariate and multivariate logistic regression analyses were performed to determine potential factors that may impact biopsy undergrading and biopsy overgrading of the Gleason score. The variables entered into the univariate analysis were also included in the multivariate model. The factors included in the analysis were age at diagnosis, race, T-stage, prostate-specific antigen (PSA) grouping, and biopsy Gleason score. All analyses were performed using SPSS V.21 (IBM, Armonk, New York, USA) or SAS release V.9.2 (SAS Institute, Cary, North Carolina, USA). All probability tests were two sided with a significant level set at 70 Race White Black Other Unknown Clinical T classification T1c T2a-c T3–4 PSA grouping 0.1–10 ng/mL 10.1–20 ng/mL >20 ng/mL unknown Biopsy Gleason score Gleason ≤6 Gleason 3+4 Gleason 4+3 Gleason 8 Gleason 9–10

Undergrading p Value

OR (95%CI)

p Value

1 1.10 (1.02 to 1.18) 1.27 (1.12 to 1.43)

0.02