Letters Section Editor: Robert M. Golub,
MD, Senior Editor.
In Reply: Drs Osborn and Chodak consider Gleason
histopathologic grading to be a “more accurate predictor of prognosis”
than cytological assessment of differentiation. The implication of “accurate”
is far from clear in this context. It could vary from the statistical precision
of an effect estimate (such as 20-year survival) to the reliability of the
factor or its performance in distinguishing fatal from nonfatal cancerous
tumors (sensitivity, specificity, or predictive value). Prognostic factors
do not exist in a vacuum; their purpose is to guide clinical management. Although
the Gleason score conveys significant prognostic information, it is not obvious
that a dichotomous clinical decision is better guided by the Gleason score
with 10 discreet levels than by the 3 discreet categories of differentiation.
In our data, degree of differentiation distinguishes groups with an almost
50-fold difference in risk of death from prostate cancer.1
Johansson J, Andrén O, Andersson S, Magnusson A, Dickman PW, Holmberg L, Adami H. Natural History of Early Localized Prostate Cancer—Reply. JAMA. 2004;292(13):1549-1550. doi:10.1001/jama.292.13.1549-c