Patterns of prostate cancer biopsy grading: trends and clinical implications

The predictive value of Gleason grading from prostate needle biopsy was examined and the patterns of grade discordance with surgical specimens are discussed in terms of their clinical implications. Gleason scores from biopsy and matched radical prostatectomy specimens were compared in 428 consecutive patients. Patterns of concordance were examined with respect to numerical agreement as well as to whether differences result in a change in group assignment with respect to Gleason score group 2-4, 5-6, 7, and 8-10. The coefficient of agreement, kappa, and accuracy were used to measure predictive value. An exact Gleason score match was present in 41% of the cases, while 48% were undergraded and 17% overgraded. With respect to group assignment, 51% remained unchanged while 35% were undergraded and 14% overgraded. Kappa analysis yielded a value of 0.26, which represents a poor agreement beyond chance. A Gleason score of 5-6, 7, or 8 was reproduced in 52%, 53%, and 58% of cases, respectively. A systematic bias toward progressive undergrading of more well-differentiated cancers and overgrading of more poorly differentiated cancers on biopsy is suggested by the data. A pooled analysis with nine additional series (n = 2,687) confirms this conclusion. No correlation was found between the amount of tumor in the biopsy specimen and grade discordance. Biases in pathologic interpretation and sampling effects are suggested as responsible for grade discordance. Predictable differences exist between the histologic grade in prostate needle biopsies and the surgical specimen. Clinical staging of organ-confined prostate cancer should include the likelihood of histologic upgrading or downgrading when used to stratify patients for clinical trials, in comparing results among therapies based on biopsy grading and when recommending a radical therapy. Developing a methodology which reduces both sampling effects and pathologic interpretation bias would likely result in significantly improved accuracy of Gleason grading of prostate biopsies. Int. J. Cancer (Radiat. Oncol. Invest.) 90, 305-311 (2000).