Correlation between Gleason score of needle biopsy and radical prostatectomy specimen: Accuracy and clinical implications

Michael S. Cookson, Neil E. Fleshner, Scott M. Soloway, William R. Fair

Research output: Contribution to journalArticle

187 Scopus citations


Purpose: Despite the reliability of Gleason grading with respect to the same specimen, the correlation between the biopsy and prostatectomy specimen is less well defined. We compared the accuracy of Gleason grading of biopsies in predicting histological grading of radical retropubic prostatectomy specimens. Materials and Methods: Gleason scores of 18 gauge needle biopsies were compared to those of radical retropubic prostatectomy specimens in 226 consecutive patients. In addition to comparing numeric discrepancies, differences between biopsy and specimen Gleason scores of 2 or more and a change in group from Gleason scores 2 to 4, 5 to 7 or 8 to 10 were evaluated by kappa testing, as well as any change in group from Gleason scores 2 to 4, 5 and 6, 7 and 8 to 10. Results: The biopsy score was identical to the specimen score in 31% of cases, while 26% were discrepant by 2 or more Gleason scores. Overall, 54% of biopsies were under graded, while 15% were over graded. If only cases in which discrepancies of 2 or more Gleason scores and a change in group were considered, there was good overall agreement (kappa 0.468, accuracy 80%). Among the cases with any change in group, the accuracy was only 46% with poor agreement (kappa 0.153). Conclusions: Overall, the reliability of Gleason grading of needle biopsies in predicting final pathology was good. However, the limitations of Gleason grading based on biopsy should be considered when discussing treatment options and comparing results based on biopsy data.

Original languageEnglish (US)
Pages (from-to)559-562
Number of pages4
JournalJournal of Urology
Issue number2
StatePublished - Feb 1997



  • biopsy
  • neoplasm staging
  • prostatectomy
  • prostatic neoplasms

ASJC Scopus subject areas

  • Urology

Cite this