Prostate Cancer Grade Assignment: The Effect of Chronological, Interpretive and Translation Bias

Abstract

Surveillance, Epidemiology and End Results (SEER) data reveal an increasing incidence in the detection of moderately differentiated prostate cancer and a stable or decreasing incidence in well and poorly differentiated cancer. Plausible reasons for this phenomenon include the decrease in transurethral resections performed and an increase in the number of prostate specific antigen triggered ultrasound guided needle biopsies. We examined additional explanations for the grade shift with time and addresses the impact of this grade shift on clinical end points. Archived slides from 100 patients (82 needle biopsy specimens and 18 transurethral resection specimens) treated for prostate cancer between 1975 and 1985 were reexamined. Current grades were assigned using the WHO and the Gleason grading systems, and they were compared with the original grade assignments. Current Gleason score was translated to WHO grade by operational SEER criteria (2 to 4-well, 5 to 7-moderately and 8 to 10-poorly differentiated). Analysis of the 100 specimens by life table methodology calculated cancer specific survival according to the era of grade assignment. There was significant upward grade migration from the historic to the current WHO grade and from the translation by SEER methodology of the current Gleason score to WHO grade (p <0.0001). Interpretive and chronological bias caused expansion of the moderately differentiated category at the expense of well differentiated cancer and significant deviation in cancer specific survival curves (p <0.013). SEER translation bias resulted in expansion of the moderately differentiated grade category at the expense of poorly differentiated cancer and eliminated the significant difference in cancer specific survival between Gleason 5/6 and 7 (p <0.006). Contemporary understanding of Gleason grading has lowered the threshold for assignment to higher grade. While a change in tumor biology may be partially responsible for the trend toward higher grade assignment, grade migration as described is also a significant contributing factor in reported histological trends for newly diagnosed prostate cancer and it significantly compromises efforts at historical comparisons.