Threshold for Diagnosing Prostate Cancer over Time

Paralleling the detection of earlier prostate cancer over the last several years, there have been numerous efforts to educate practicing pathologists on the diagnosis of limited prostate cancer on needle biopsy via journal articles, Web sites, books, and educational courses. The current study was undertaken to assess whether the threshold for the diagnosis of prostate cancer on needle biopsy has lowered over time. One thousand twelve prostate needle biopsy cases obtained in consultation by 1 of the authors because of diagnostic concerns over a period of 12 weeks (November 15, 2001 to February 15, 2002) were reviewed. Cases referred by either the patient or clinicians were excluded. The final diagnoses in this series were compared with a previously published series of needle biopsy cases of the prostate seen in consultation by the same author in 1993-1994. The percentage of cancer in the 2001-2002 series was 55.1%, compared with 69.6% for the 1993-1994 series. The mean and median numbers of malignant glands in the 2001-2002 series (mean, 21.9; median, 14; range, 2 to 296) were significantly smaller than in the 1993-1994 series (mean, 31.0; median, 20; range, 2 to 300; P < 0.00001). The incidence of atypical glands that were suspicious for but not diagnostic of carcinoma was 23.9% in the 2001-2002 series and was 10.7% in the 1993-1994 series. The mean and median numbers of atypical glands in the 2001-2002 series were 9.4 and 6, respectively (range, 1 to 70); these parameters were not available for the previous series. The percentage of high-grade prostatic intraepithelial neoplasia diagnoses was similar in the 2001-2002 and 1993-1994 series (5.1% and 4.6%, respectively), as was the overall frequency of benign cases (15.8% and 15.1%, respectively). The percentage of cases that were accompanied by immunohistochemical stains for 34βE12 in the 2001-2002 series was 44.4%, which was much more than the 2. 5% seen in the 1993-1994 series; if anything, this should have resulted in a lower atypical rate in the current series. In more recent years, cases sent for consultation have more limited cancer, with a correspondingly higher percentage of cases that are diagnosed as atypical, yet not diagnostic of cancer. The number of atypical glands in the recent series was very limited, such that a change over time in the threshold for diagnosing cancer by the consultant is an unlikely explanation. It appears that pathologists are becoming more skilled at diagnosing limited prostate cancer and are referring predominantly cases with fewer cancer glands and more difficult atypical cases with few glands.