Prediction of pathologic stage and postprostatectomy disease recurrence by DNA ploidy analysis of initial needle biopsy specimens of prostate cancer.

BACKGROUND DNA ploidy determination of carcinomas in radical prostatectomy specimens has shown significant correlation with patient outcome, but the predictive value of ploidy status of cancers obtained by transrectal ultrasound-guided needle biopsies has not been studied extensively. METHODS Eighty-nine paired needle biopsy specimens (NBX) and radical prostatectomy (RPX) specimens from patients with early clinical stage (A2-B2) prostate cancer were evaluated for DNA content by image analysis of Feulgen stained tissue sections. Findings were compared with Gleason grading on the same specimens by univariate and multivariate analyses for prediction of local tumor invasion, metastasis, disease recurrence, and serum prostate specific antigen concentration during a 0.9-6.0 year clinical follow-up period. RESULTS There was excellent correlation of ploidy status between NBX and RPX specimens (P < 0.0001); NBX and RPX grades did not correlate. On RPX specimens, aneuploid status correlated with high tumor grade (P < 0.0005). Aneuploidy in NBX specimens was associated with a twofold higher rate of extracapsular spread (ECS) (P = 0.04). Aneuploid NBX tumors featured a tenfold greater frequency of metastasis than did diploid NBX tumors (P < 0.005). Radical prostatectomy grade correlated with ECS (P < 0.001) and presence of metastatic disease (P = 0.04). On multivariate logistic regression analysis, aneuploidy in both NBX and RPX specimens was the most significant variable and independently predicted the presence of metastasis (P = 0.006 for NBX; P = 0.028 for RPX). Tumor grade of NBX and RPX specimens did not independently predict metastatic disease or disease recurrence, but RPX grade was associated independently with ECS (P = 0.005). Aneuploid NBX tumors recurred after RPX three times more often than did diploid cases, which was significant on univariate (P < 0.001) and multivariate (P = 0.018) analyses using the Cox proportional hazards model. There was no correlation with NBX or RPX Gleason score and disease recurrence. Preoperative serum PSA concentration did not correlate with tumor grade or ploidy status, but on multivariate analysis, when paired with ploidy status, independently contributed to the propensity for ECS, metastasis, and disease recurrence. CONCLUSIONS DNA content analysis of early clinical stage prostate carcinoma needle biopsy specimens by image analysis directly correlates with radical prostatectomy specimen ploidy status and is associated independently, with the presence of metastasis, postprostatectomy disease recurrence, and ECS. Needle biopsy tumor grading did not correlate with prostatectomy grade and did not predict disease outcome accurately.