Page 61 - An Evidence Review of Active Surveillance in Men With Localized Prostate Cancer

Tumor volume. The large epidemiologic datasets included in our review did not have
information pertaining to trends regarding tumor volume. We performed additional targeted
searches in MEDLINE using key words relevant to “tumor volume” and time trends, but did not
identify additional studies.

Tumor grade. Sixteen studies (7 SEER or its component registries, 6 CaPSURE, 2 NCDB, 1
POCS) covering 1973 to 2007 reported information on trends in tumor grade distribution at
disease presentation (Appendix Table C1.8). 4,22,30,37,46,52,56,65,67,68,71,73,75,77,86,90 Studies using SEER
data consistently demonstrated reductions in the proportion of patients diagnosed with well- or
poorly-differentiated tumors (including undifferentiated tumors) with concomitant increases in
the proportion of patients with moderately-differentiated disease. Within each study, this
temporal trend was found to be statistically significant in six of the seven studies that reported
the results of statistical tests assessing changes in the distribution of tumor grade over time. The
six studies using CaPSURE data and one study using POCS data reported temporal trends in the
distribution of Gleason scores. Generally, these studies showed a decrease over time in the
proportion of tumors with Gleason scores 2-4 and increases in tumors with Gleason scores 5-7.

Prostate specific antigen. Eight studies (7 CaPSURE, 1 POCS), covering 1989 to 2007,
reported information on trends in PSA levels at presentation (Appendix Table
C1.9). 4,64,65,67,68,71,77,90 Seven studies categorized PSA values (e.g., < 4, 4-10, >10 ng/mL) and
only one study reported the median PSA value by diagnosis year. Generally, studies found that
the PSA values at diagnosis have decreased over time (i.e., that a larger number of patients are
currently diagnosed with PSA concentrations below 10 ng/mL; this was true even for patients
90
with low risk localized prostate cancer ).
We did not identify studies reporting on trends in the proportion of screen-detected prostate
cancer cases among all cancer cases that met our inclusion criteria. One study demonstrated that
for all age groups above 65 years and both for blacks and whites, the proportion of men who
underwent PSA testing at least once and were diagnosed with prostate cancer within 90 days of
96
the test among all men undergoing PSA testing has decreased over time (1988-96).
Distinguishing between screen-detected prostate cases (i.e., cancer cases identified following
investigation triggered by a positive PSA test) and cases where use of the PSA test was used as a
confirmatory or add-on test (e.g., as part of the investigation of clinical symptoms or signs
suggestive of prostate cancer, or suspicious findings on prostate digital rectal examination) is
particularly challenging using administrative data and may be uncertain even after review of
96
complete medical records.

Diagnostic Strategies
Biopsy Frequency. Four studies (2 SEER component registries, 2 SEER-Medicare), covering
1982 to 2001, reported information on trends in the performance of prostate biopsies (Appendix
Table C1.10). 52,54,57,89 One study using data from the SEER-Detroit registry reported that the
proportion of prostate cancer patients diagnosed through biopsy (compared to those diagnosed
through other procedures, such as transurethral resection of the prostate) increased over time
52
(1982-95, P < 0.001). A similar trend was evident in a study using data from the SEER-New
54
Mexico registry. A SEER-Medicare study also demonstrated an increase in the age-adjusted
57
rate of biopsy procedures (from 685 to 2600 per 100,000 men) between 1986 and 1991. An
updated analysis from the same database, covering the period 1993-2001, reported that the age-

26