Diagnostic uncertainty expressed in prostate needle biopsies. A College of American Pathologists Q-probes Study of 15,753 prostate needle biopsies in 332 institutions.

OBJECTIVE: To determine the rate of diagnostic uncertainty in rendering diagnoses on prostate needle biopsies and to examine pathology practice variables that influence that rate.
DESIGN: Anatomic pathology departments participating in the College of American Pathologists Q-Probes laboratory quality improvement program retrospectively reviewed their last 50 consecutive prostate needle biopsy diagnoses. For each diagnosis, participants provided information concerning patients' prostate-specific antigen levels; number, locations, and laterality of biopsy specimens; number of tissue levels examined; performance of high-molecular-weight cytokeratin immunoperoxidase staining; and acquisition of consultations from general pathologists or experts in prostate pathology. Characteristics of pathology practices included yearly surgical and prostate needle biopsy caseloads, number of pathologists rendering biopsy diagnoses, use of standard descriptive checklists, access to patients' prostate-specific antigen and digital rectal examination results, percentages of prostate needle biopsies routinely submitted for internal consultations, and presence of departmental experts in prostate pathology. SETTING AND PARTICIPANTS: Three hundred thirty-two public and private institutions located in the United States (n = 318), Canada (n = 6), Australia (n = 5), United Kingdom (n = 2), and Guam (n = 1). MAIN OUTCOME MEASURE: The rate of diagnostic uncertainty in prostate needle biopsy diagnoses.
RESULTS: Participants submitted diagnoses on a total of 15 753 prostate needle biopsy cases, of which 33.4% were adenocarcinoma; 55.5% were benign; 3.9% were carcinoma in situ, prostatic intraepithelial neoplasia, or both; and 7.1% were diagnostically uncertain. The median rate of diagnostic uncertainty was 6%, ranging from 0 at the 10th percentile to 14% at the 90th percentile of all participating laboratories. Performing high-molecular-weight cytokeratin immunoperoxidase staining resolved diagnostic uncertainty in 68% of cases in which it was performed, and obtaining intradepartmental and extradepartmental consultations resolved diagnostic uncertainty in 70% to 87% of cases for which they were obtained. Knowledge of patients' prostate-specific antigen results and examining multiple biopsy cores had marginal effects on the rate of uncertainty. Thoroughness of prostate gland sampling and examination of multiple tissue block levels were not associated with the aggregate rate of diagnostic uncertainty. We found no particular pathology departmental practices or institutional demographic characteristics associated with institutional rates of diagnostic uncertainty.
CONCLUSIONS: Use of high-molecular-weight cytokeratin immunoperoxidase staining and obtaining intradepartmental and extradepartmental consultations may be effective in reducing diagnostic uncertainty in prostate biopsies.