Interobserver variability in the pathological assessment of radical prostatectomy specimens: findings of the...

The aim of this study was to strengthen the validity of future findings in the Laparoscopic Prostatectomy Robot Open (LAPPRO) study by investigating the extent of interobserver variability between local pathologists and re-evaluating reference pathologists.

LAPPRO is a Swedish prospective study comparing robot-assisted laparoscopic prostatectomy to open retropubic radical prostatectomy. Patients were recruited from 2008 to 2011. A random selection of 289 prostatectomy specimens was re-evaluated, in a blind fashion, by two reference pathologists from a University Hospital in Denmark and compared with original reports from local pathologists.

The exact concordance rate of Gleason score (GS) between local and reference pathologists was 56% (Spearman correlation coefficient 0.54). Exact concordance rates (κ value) for pathological tumour stage (pT), extraprostatic extension (EPE), surgical margin status (SMS) and seminal vesicle invasion (SVI) were 87% (0.63), 86% (0.59), 92% (0.76) and 98% (0.82), respectively. In subanalyses for surgical technique, exact concordance rates of GS, pT, EPE, SMS and SVI were 58%, 83%, 84%, 90% and 97%, respectively, for surgical technique 1 (ST1), compared to 55%, 88%, 87%, 93% and 98%, for surgical technique 2 (ST2). In ST1 specimens undergrading of GS by the local pathologists compared to central review was more common than overgrading (26% vs 16%). The inverse relationship was seen in ST2 specimens (14% vs 32%).

Re-evaluation of randomly selected prostatectomy specimens in the LAPPRO cohort showed comparable results compared to previous studies of this kind. A systematic variation in the assessment of GS exists, attributable to individual differences in judgement between pathologists. Dichotomising GS (≤ 7 vs ≥ 8) overcomes the systematic variation.