Context: - The incidence of prostate cancer with Gleason scores 2 through 4 has been decreasing for decades, largely because of evolving criteria for Gleason scores, including the 2005 International Society of Urological Pathology recommendation that scores of 2 through 4 should rarely, if ever, be diagnosed based on needle biopsy. Whether trends in assigning Gleason scores 2 through 4 vary by facility type and patient characteristics is unknown.
Objective: - To assess trends in prostate cancer grading among various categories of treatment facilities.
Design: - Analyses of National Cancer Database records from 1990 through 2013 for 434 612 prostate cancers diagnosed by core needle biopsy, including multivariable regression for 106 331 patients with clinical T1c disease diagnosed from 2004 through 2013.
Results: - The proportion of prostate core needle biopsies with Gleason scores 2 through 4 declined from 11 476 of 53 850 (21.3%) (1990-1994) to 96 of 43 566 (0.2%) (2010-2013). The proportions of American Joint Committee on Cancer category T1c needle biopsies assigned Gleason scores 2 through 4 were 416 of 12 796 (3.3%) and 9 of 7194 (0.1%) during 2004 and 2013, respectively. Declines occurred earliest at National Cancer Institute-designated programs and latest at community programs. A multivariable logistic model adjusting for patient demographic and clinical variables and restricted to T1c cancers diagnosed in needle biopsies from 2004 through 2013 showed that facility type is independently associated with the likelihood of cancers in such specimens being assigned Gleason scores of 2 through 4, with community centers having a statistically significant odds ratio of 5.99 relative to National Cancer Institute-designated centers.
Conclusions: - These results strongly suggest differences in Gleason grading by pathologists practicing in different facility categories and variations in their promptness of adopting International Society of Urological Pathology recommendations.