Potential Impact of Gleason Grade Reproducibility on Active Surveillance Management in Prostate Cancer: A Multi-Institutional Study of Patients Enrolled in a Prospective Active Surveillance Study (PASS).
Jesse K McKenney, Jeff Simko, Michael Bonham, Lawrence D True, Sarah Hawley, Lisa Newcomb, PASS Pathology Group, Peter R Carroll, James D Brooks. Stanford, Stanford, CA; University of California, San Francisco; University of Washington, Seattle; Canary Foundation, Palo Alto, CA
Background: The impact of Gleason grade reproducibility on the management of patients in an active surveillance protocol has not been fully addressed.
Design: Three sets of digital images, each comprised of 17 separate prostatic carcinomas on core needle biopsy, were constructed [set 1 (classic Gleason patterns), set 2 (tangential Gleason pattern 3 vs Gleason pattern 4), set 3 (intraobserver test with same cases as set 2). Eleven pathologists from the PASS study sites assigned Gleason scores for each case. Interobserver and intraobserver reproducibility was assessed for assignment of the highest Gleason pattern (3 vs 4 or higher). In addition, a set of prostate needle core biopsies from of 97 consecutive patients enrolled in an active surveillance study (Canary PASS) were identified and those with available glass slides (n=82) were re-reviewed to determine the frequency of cases requiring the distinction of tangentially sectioned well-formed glands (Gleason pattern 3) from focally poorly formed glands (Gleason pattern 4).
Results: Interobserver reproducibility for classic Gleason patterns was “substantial” (Light's kappa: 0.76). Varying interpretation of carcinoma glands as tangentially sectioned Gleason pattern 3 or small poorly formed glands of Gleason pattern 4 led to the rare discrepancy in this test set. Interobserver reproducibility in test set 2 was only “fair” (Light's kappa: 0.26). Intraobserver reproducibility between sets 2 and 3 for each observer ranged from 65% to 100% (mean 81.5%; median 82%). In the set of biopsies retrospectively reviewed from 82 patients on active surveillance, 61 patients had carcinoma. 15 (24.5%) of these 61 patients had a set of biopsies with at least one focus in which the distinction between tangentially sectioned Gleason pattern 3 glands and focal poorly formed Gleason pattern 4 glands had to be carefully considered.
Conclusions: Reproducibility of grading classic Gleason patterns is high. However, variability in grading occurred in making the distinction between tangentially sectioned pattern 3 glands and the subset of pattern 4 glands that are poorly formed glands. Developing universally accepted histologic and/or molecular criteria for distinguishing these patterns and, subsequently, characterizing their natural history would be useful in the management of patients on active surveillance.
Category: Genitourinary (including renal tumors)
Wednesday, March 2, 2011 9:30 AM
Poster Session V # 69, Wednesday Morning