Prostate Cancer 2004: Insights From National Disease Registries

Prostate Cancer 2004: Insights From National Disease Registries

ABSTRACT: In 2004, the large majority of prostate cancers are detected via prostate- specific antigen (PSA) screening. Most are diagnosed at an early stage and are amenable to aggressive local treatment. However, the natural history of the disease may be prolonged, and all available active treatments exert a potential negative effect on patients’ HRQOL. Management options for localized prostate cancer have become increasingly complex in recent years, and rigorous trials are frequently difficult to perform due to the extended follow-up required to reach meaningful outcomes. In this context, the advent of the national prostate cancer disease registries—Prostate Cancer Outcomes Study (PCOS), Center for Prostate Disease Research (CPDR), Cancer of the Prostate Strategic Urologic Research Endeavor (CaPSURE), and Shared Equal Access Regional Cancer Hospital (SEARCH)—has greatly facilitated clinical research in prostate cancer. This review summarizes key findings from the registries in the areas of risk migration, practice patterns, outcome prediction, and quality-of-life outcomes. The availability of these large databases of patients will be a tremendous asset as prostate cancer management continues to evolve in the coming years.

Over the past decade, early detection
of prostate cancer cases
via prostate-specific antigen
(PSA) screening has produced a
downward stage migration at presentation.
Earlier diagnosis and therapeutic
advances have facilitated the
increased use of aggressive local treatment,
particularly radical prostatectomy,
and radiation therapy.[1]
However, despite growing evidence
supporting the finding of a reduction
in prostate cancer mortality with local
treatment of early-stage disease,[2] the
natural history of the disease-especially
among screen-detected cases-
may be protracted, and a majority of
men diagnosed with prostate cancer
do not actually die of their tumors.[3]
Moreover, all available treatments can
negatively affect patient health-related
quality of life (HRQOL)[4]; indeed,
given consistently excellent
survival rates after treatment for lowrisk
tumors, HRQOL outcomes have
assumed growing importance in the
recent literature regarding localized
prostate cancer.

Management practices for localized
prostate cancer, subject to myriad
scientific, clinical, economic, and
social influences, are becoming increasingly
complex. Treatment patterns
may vary from region to region,
and practices and outcomes reported
from most academic studies may not
be reflected consistently in community
practice and other contexts. The
advent of national disease registries,
which enroll large numbers of patients
consecutively outside of any particular
clinical trial, has greatly facilitated
the study of prostate cancer epidemiology
and practice patterns, and of
"real world" effectiveness of therapy
in terms of both oncologic and
HRQOL outcomes. These registries
are particularly useful for studying
trends among nonwhite patients who
tend to be underrepresented in most
published clinical studies. This review will summarize key findings in these
areas from the four primary prostate
cancer registries.

The National Prostate
Cancer Registries
The Prostate Cancer Outcomes
Study (PCOS) was initiated in 1994
by the National Cancer Institute (NCI)
to investigate patterns of treatment
and HRQOL outcomes in American
men diagnosed with prostate cancer.
Enrollment occurred during 12
months between 1994 and 1995, from
6 of the 10 Surveillance, Epidemiology
and End Results (SEER) cancer
registries. A total of 11,137 men who
were diagnosed with prostate cancer
during this time in this catchment area
were eligible for the PCOS.[5] The
study restricted eligibility to white,
African-American, and Hispanic men.
A prespecified random sampling design,
weighted to enhance accession
of African-American and Hispanic
men and those under 60 years of age,
was used to determine which of the
eligible men would be contacted for

PCOS HRQOL data were ascertained
via a survey questionnaire administered
at 6, 12, and 24 months
after initial diagnosis. A mixed-mode
survey approach was used to ensure
high response rates: Initial contact
comprised a mailed HRQOL questionnaire
with a cover letter. Nonresponders
were followed-up with a
telephone call and remailing of the
questionnaire if necessary. As a final
resort, a telephone interview was attempted.
The response rate differed
by individual SEER registry, ranging
from 54% to 76%. Nonresponders
were more likely to be older, nonwhite
residents of areas with lower
median household income and lower
prevalence of college education, and
to have undergone radical prostatectomy
(42% of responders vs 26% of

The mailed questionnaire, the
PCOS prostate cancer-specific
HRQOL index, was a novel tool derived
from previously validated instruments
such as the University of
California, Los Angeles (UCLA) prostate
cancer index. Based on three such
instruments used in the prostate cancer
literature, the urinary, bowel, and
sexual domains were addressed. Other
variables included the presence and severity
of comorbid conditions (based
on Medical Outcomes Study Short
Form-36 [SF-36] responses), socioeconomic
status, demographic data, and
satisfaction with treatment chosen. Because
more than 14% of men were
Hispanic, this PCOS scale was translated
into Spanish and pretested in Spanish-
speaking men with prostate cancer.
Overall, 5,672 men were sampled, with
3,533 (62%) responders and 2,139
(38%) nonresponders.[5]

More information about the
PCOS can be obtained at http://

In 1994, the Department of Defense
Center for Prostate Disease Research
(CPDR) began collecting data
on prostate cancer patients treated
within the medical systems of the military
services, first at Walter Reed
Army Medical Center, and since 1997,
at eight other military sites around the
country; three other sites initially
participated but were excluded due to
poor performance. In 1999, a uniform
database entry form was developed.
Data quality is ensured by dedicated
study coordinators at each participating
site in the CPDR. Unique
among the registries, the CPDR
project seeks to standardize clinical
practice across the various military
sites rather than simply to collect clinical

The mean age at diagnosis fell from
68 in 1991 to 65 in 1999. Of CPDR
patients, 39% elected radical prostatectomy,
32% underwent externalbeam
radiation therapy, 29% primary
androgen-deprivation therapy, and the
remainder, miscellaneous therapy
(brachytherapy, cryotherapy, etc).[6]
Including 13,878 patients as of June
2002, CPDR is the only registry to
have reported a significant number of
men who underwent radical perineal
prostatectomy (finding similar oncologic
results to those undergoing retropubic
prostatectomy).[7] More
information on CPDR can be obtained

The Cancer of the Prostate Strategic
Urologic Research Endeavor
(CaPSURE) was initiated in 1995 with
the aims of tracking prostate cancer
epidemiologic trends, practice patterns,
and oncologic and HRQOL outcomes.
A total of 40 urologic practice
sites are represented, 31 of which are
actively enrolling patients. The sites
are primarily community based, although
4.8% and 3.2% of patients are
treated, respectively, at four university-
based and three Veterans Affairs
(VA) medical centers. At each practice
site, all men with biopsy-proven
prostate cancer are invited consecutively
to join CaPSURE.

CaPSURE collects approximately
1,000 clinical and patient-reported
variables. The clinical information,
collected by the treating urologist at
baseline and each time the patient returns
for care, includes history of prostate
cancer diagnosis, biopsies,
pathology, staging tests, primary and
subsequent treatments, clinic procedures,
Karnofsky performance status
scores, and medications. At enrollment
and every 6 months thereafter,
each patient completes a questionnaire
addressing sociodemographic parameters,
comorbidities, and baseline

The HRQOL components of the
questionnaires include extensively
validated survey instruments: the
SF-36 for general HRQOL and the
Prostate Cancer Index (PCI) for disease-
specific HRQOL. Since 1999, the
questionnaires have also included a
survey on patient satisfaction with care
and an assessment of fear of cancer
recurrence. Other sections of the patient
questionnaires assess all use of
health services. Patient-reported hospitalizations
and emergency department
use are verified via review
of discharge summaries.

Patients are treated according to
their physicians' usual practices and
are followed until time of death or
withdrawal from the study. Completeness
and accuracy of the data are ensured
by random sample chart review
at periodic intervals. Additional details
of the project methodology have
been reported previously.[8] The database
is managed by the Urology Outcomes Research Group of the Department
of Urology at the University
of California, San Francisco (UCSF)
and has been funded since its inception
by a grant from TAP Pharmaceutical
Products, Inc.

As of February 2004, there were
10,581 patients enrolled in CaPSURE.
The median patient age at diagnosis
is 67, and nearly 75% of the men are
between 60 and 79 years of age. The
majority are white, with roughly 10%
African-American representation, and
3.5% Hispanic, Asian, or other ethnicities.
There is a fairly even distribution
across socioeconomic strata as
assessed via education and income
level. Over half of CaPSURE patients
are covered by Medicare, with or without
supplemental policies, and the remainder
have a variety of health
coverage types.

Overall, patients are distributed
roughly evenly among low-, intermediate-,
and high-risk groups as defined
by D'Amico et al,[9] although
over time, the proportion with lowrisk
disease has been rising, and that
with high-risk has been falling.[10]
Radical prostatectomy is the most
common primary treatment, followed
sequentially by primary androgendeprivation
therapy, external-beam
radiotherapy, brachytherapy, and
watchful waiting. More information
can be obtained at

The Shared Equal Access Regional
Cancer Hospital (SEARCH) database,
the most recently inaugurated
national prostate cancer registry, comprises
retrospective data from consecutive
patients undergoing radical
prostatectomy at three California VA
medical centers and the San Diego
Naval Medical Center between 1988
and 2001. The VA medical center in
Augusta, Ga, was added later, when
one of the original SEARCH investigators
transferred there.

Exclusion criteria include men who
had preoperative neoadjuvant androgen-
deprivation therapy or radiation
therapy, unknown race, pT0 disease,
or unknown preoperative PSA. From
1988 to 2001, 1,547 patients were enrolled,
and the racial distribution was
1,014 white (66%), 338 African-American
(22%), and 195 Hispanic, Asian,
or Native-American men, who were
collectively termed "non-white, nonblack"

Variables within SEARCH include
clinical and pathologic variables but
no HRQOL data. A principal advantage
to SEARCH is the large proportion
of minority men included (34%),
and the relatively equal access to care
afforded by each of the participating
centers. That said, it should be noted
that the VA medical centers may not
truly be "equal access" centers due to
unmeasured factors such as distance,
availability of transport, and so forth,
which may be unequally distributed
in a systematic fashion (eg, farmers
live farther away than city dwellers,
or men living in the inner city may
have no access to transportation).
However, on average, these centers
are certainly more inclusive of patients
with low socioeconomic status
than those represented in the other

The principal features, advantages,
and disadvantages of each of the
registries are compared in Table 1.

Trends in Prostate Cancer
Presentation and Risk

A recent CaPSURE analysis examined
changes over time in patient risk
characteristics at the time of diagnosis.
Patients presenting with low-risk disease,
defined by PSA ≤ 10 ng/mL,
Gleason score under 7 with no pattern
4 or 5 disease on biopsy, and clinical
stage T1c or T2a,[9] have increased
from 31% of patients in 1989-1990
to 47% in 2001-2002. During the
same time, those with high-risk disease-
PSA > 20 ng/mL, Gleason 8-10
biopsy, or stage T3/4 disease[9]-
decreased from 41% to 15%. T1 tumors
became increasingly prevalent,
as did those with Gleason 7 biopsies
and those diagnosed with a PSA level
of 4 to 10 ng/mL.

In the early years of the study, patients
were most likely to be classified
as high-risk due to a high PSA
level, whereas more recently, they
were more likely to have a low PSA
and a high Gleason score.[10] A report
of trends within CPDR likewise
confirms stage migration, with the
percentage of patients presenting with
metastatic disease declining from
14.1% in 1988 to 3.3% in 1998.[12]

Diagnosis of Gleason 2-5 disease
on biopsy has become quite rare.[10]
Nevertheless, other investigators have
demonstrated that Gleason scores have
been rising over the past decade as a
result of changes in pathologic grading
practices,[13] suggesting that even
as patients are being diagnosed less
commonly with high-risk disease, a
contemporary patient considered to
be at high risk may have a better prognosis
than a high-risk patient in the
early 1990s. Interestingly, a study
from SEARCH compared the time to
failure for biopsy Gleason score 4 (as
referent), 5, and 6 tumors, finding a
monotonically increasing earlier risk
of failure, with a hazard ratio of 1.31
(P = .025).[14] This finding was not
borne out among CaPSURE patients.

Advanced Disease
As a population-based rather than
urology-based database, PCOS includes
greater numbers of patients
with advanced disease. Hoffman et al
described racial differences in men diagnosed with advanced prostate cancer,
defined as clinical T4, N1, and/or
M1 tumors.[15] The PCOS included
2,187 non-Hispanic white men, 539
African-Americans, and 447 Hispanics,
of whom 8% presented with clinically
advanced-stage prostate cancer.
Of men in the PCOS cohort, 12.3% of
African-Americans, 10.5% of Hispanics,
and 6.3% of whites presented with
T4 disease.

Univariate analysis indicated significantly
higher risk for advanced
disease among African-Americans
and Hispanics. Gleason 8-10 cancers
also tended to be more prevalent
among African-American men. Controlling
for age, grade, and socioeconomic
status (as measured by
insurance status and educational level),
African-American but not Hispanic
men remained statistically
significantly more likely to be diagnosed
with advanced disease.[15]

Higher-Risk Populations
The 241 CaPSURE patients enrolled
at VA medical centers are much
more likely than the general patient
population to be African-American,
to have lower income, less education,
and more comorbidity at presentation.
They also have significantly higher
risk disease in terms of PSA at diagnosis
and biopsy Gleason score. They
are more likely to undergo watchful
waiting or receive primary androgendeprivation
therapy, and less likely to
receive definitive local therapy.[16]

African-American patients included
in CaPSURE likewise have consistently
higher-risk characteristics,
but the trends toward better risk at
presentation were identical to those
seen in the general database.[10] A
model of biochemical outcomes in
CaPSURE found African-American
ethnicity to be a predictor of recurrence
in univariate analysis, with the
greatest difference in outcomes between
African-American and white
patients with high-risk characteristics.

The 5-year biochemical diseasefree
survival rates were estimated at
65% among white patients and 28%
among African-American patients.
However, in multivariate analysis controlling
for income and education-
variables that associated closely with
ethnicity-ethnicity was no longer an
independent predictor of outcome.[17]
A similar analysis from SEARCH likewise
found that African-American ethnicity
correlated with outcome but was
not an independent predictor in multivariate

Amling et al reported an association
among CPDR patients between
obesity as measured by body mass
index (BMI) and higher pathologic
stage and grade and younger age at
diagnosis.[18] They then correlated
BMI with ethnicity, and found the
highest mean BMI among African-
American men. They suggest that
higher BMI among African-American
men may offer an explanation for
the higher-risk disease and earlier age
of diagnosis among these patients.[18]
The relationships between ethnicity,
BMI, and risk are complex and intriguing,
and are the subject of numerous
ongoing studies in both

National Practice Patterns

Analyses dating back to the early
1990s have demonstrated minimal
benefit of imaging tests for staging
patients with low-risk disease characteristics
such as PSA < 10 ng/mL.[19]
An analysis of CaPSURE data found
that rates of both bone scan and computed
tomography scan use have fallen
dramatically in recent years, with
the greatest declines among lowerrisk
patients. Indeed, whereas among
early CaPSURE patients, disease risk
exerted no influence on the likelihood
of imaging, in more recent years, rates
of imaging are strongly associated
with risk characteristics.[20]

Due to the extended natural history
of prostate cancer and the HRQOL
impact of all available active treatments,
investigators have recently paid
increased attention to watchful waiting
as a viable alternative for the initial
management of prostate tumors
with favorable risk characteristics.
Among PCOS patients surveyed in
1994-1995, 47.6% underwent radical
prostatectomy; 23.4%, radiation
therapy; 10.5%, primary androgendeprivation
therapy; and 18.5%,
watchful waiting.[21] After adjusting
for age, clinical stage, grade, and PSA level, predictors of conservative therapy
(defined as primary androgendeprivation
therapy alone or watchful
waiting) were age over 75, history of
myocardial infarction, unmarried status,
and preexisting urinary, bowel,
or erectile problems. Men in Connecticut
underwent aggressive treatment
more frequently than did men in the
other five SEER registries examined.

Racial differences in treatment
were absent in men less than age 60,
but uniformly across age strata in men
over age 60. African-American men
underwent aggressive treatment (radical
prostatectomy or radiation therapy)
less frequently than white or
Hispanic men.[21] A more recent
study from PCOS focusing on the
question of racial differences in treatment
patterns found that African-
American men with higher-risk
tumors received aggressive treatment
less frequently.[22]

Watchful Waiting
An early cross-sectional analysis
from CaPSURE found that only 8.2%
of patients in the database pursued
watchful waiting as primary management.[
23] A subsequent study of temporal
trends found that watchful
waiting use has, in fact, fallen from
9.5% in 1992-1994 to 5.5% in 1998-
2000, with the sharpest declines
among low-risk patients.[24] In the
cross-sectional study, over half of the
watchful waiting patients underwent
secondary treatment within 5 years,
especially those who were younger or
had higher PSA scores at diagnosis.
Most of these received androgen-deprivation
therapy.[23] PSA kinetics are
a strong driver of treatment decisions:
Patients whose PSA levels increase
> 5 ng/mL during watchful waiting
were nearly four times as likely to
convert to active treatment as those
with an increase < 2 ng/mL. Highrisk
baseline characteristics were also
significant predictors of eventual active

Low-Risk Patients
The findings of declining watchful
waiting use prompted an examination
in greater detail of treatment trends
among low-risk patients. Preliminary
results suggest that among CaPSURE patients, use of watchful waiting in
low-risk patients has fallen by more
than half, from 20% of patients in
1993-1995 to 8% in 1999-2001. Over
the same time, use of external-beam
radiation therapy fell from 13% to
7%, while that of radical prostatectomy
fell slightly, from 55% to 52%.

In contrast, use of primary androgen-
deprivation therapy and brachytherapy
increased significantly, from
7% to 12% and 4% to 22%, respectively.
Even among patients 75 years
of age or older, watchful waiting use
fell from 52% to 24%, while primary
androgen-deprivation therapy increased
from 23% to 30%, and brachytherapy
from 3% to 31% of patients.
Overall, roughly half of all patients
over age 75 receive either primary
androgen-deprivation therapy or neoadjuvant
androgen-deprivation therapy.
There was no significant influence
of ethnicity on treatment patterns in

Androgen Deprivation Therapy
Another CaPSURE study extended
cross-sectional observations from
PCOS[26] suggesting higher than expected
use of primary androgen-deprivation
therapy among localized
prostate cancer patients. The use of
primary androgen-deprivation therapy
as monotherapy has risen dramatically
across groups over the past decade, from
5% to 14%, 9% to 20%, and 33% to
48% among low-, intermediate-, and
high-risk patients, respectively, from
1989-1990 to 2000-2001. Likewise,
neoadjuvant androgen-deprivation therapy
use has risen from 3% to 8% of
patients undergoing radical prostatectomy,
10% to 75% of those receiving
external-beam radiation therapy, and
7% to 25% of those receiving brachytherapy.[

Cost Analysis
The resource utilization data in
CaPSURE also offer a means of studying
cost implications of various management
strategies for prostate cancer.
Penson et al analyzed the first-year
costs associated with various treatment
options based on Medicare payment
schedules. They found a trend
toward significantly higher costs for
higher-stage patients. Costs were not different between radical prostatectomy
and external-beam radiation therapy
patients, but were significantly
higher for patients receiving neoadjuvant
androgen-deprivation therapy
before either primary treatment.[28]


Loading comments...
Please Wait 20 seconds or click here to close