EURURO Vol. 71 No. 5

730 844

1. Introduction
Prostate cancer is the most common cancer in men and the
second leading cause of cancer-specific death . The vast
majority of men will present with localized disease
amenable to surgery or radiation therapy. Many men,
however, present with indolent cancers that will not
become clinically relevant. While there is evidence to
suggest such men can be safely observed, many still receive
definitive local treatment in the USA At the same time,
many men with more aggressive disease may in fact be
undertreated Given the impact of prostate cancer
treatment on quality of life and significant differences in the
cost of these approaches, a careful understanding of their
modern use is vital
Previous reports have identified significant variability in
patterns of care but are limited by use of databases derived
from small subsets of community-based practices or
inclusion only of particular patient subsets such as older
men . Using the National Cancer Database (NCDB),
we sought to describe recent temporal trends in the
management of localized prostate cancer in the USA and
to identify clinical and sociodemographic variables associ-
ated with the receipt of particular treatment paradigms.
2. Materials and methods
2.1. Data source
The NCDB captures approximately 70% of all newly diagnosed cancer
cases in the USA each year making it the largest and most broadly
representative clinical registry currently available for research De-
tailed descriptions of data coding methods have been documented
previously .
2.2. Inclusion/exclusion criteria and categorization
Male patients aged 18–90 yr with prostate adenocarcinoma diagnosed
between 2004 and 2012 were included. Patients with documented nodal
or distant metastatic disease were excluded. Patients with missing
information on any available demographic variables were excluded.
Patients were categorized into low-, intermediate-, or high-risk prostate
cancer based on 2015 National Comprehensive Cancer Network guide-
lines. Patients missing any variables needed to accurately assign them to
a risk group (prostate-specific antigen, T stage, or Gleason score) were
excluded.
2.3. Variable definition
Primary treatment was defined as within 1 yr of diagnosis and was
categorized as radical prostatectomy (RP), brachytherapy, external beam
radiotherapy (EBRT), cryotherapy, primary androgen deprivation thera-
py (ADT), or observation. Use of ADT along with RP, EBRT, brachytherapy,
or cryotherapy was also analyzed and defined as receipt of ADT up to
6 mo before definitive treatment until such treatment began. Patient
insurance status was defined as private, uninsured, Medicaid (the US
government program for the poor), younger (aged 18–64 yr) Medicare
(the US government program for the elderly or disabled), older Medicare
(aged 65+ yr), or government (such as the Indian Health Service and
Veteran’s Health Administration). Comorbidity was measured by
modified Charlson-Deyo comorbidity score .
2.4. Statistical analysis
Patient demographics and clinical characteristics by risk group were
summarized and compared with chi-square tests with a significance
level of 0.05 (two-sided). Cochran-Armitage trend tests were conducted
to analyze changes in patterns over time. Multivariate logistic regression
models were used to identify factors associated with receipt of surgery
versus radiotherapy and also receipt of observation versus definitive
treatment. Statistical analyses were performed using SAS version 9.4
(SAS, Cary, NC, USA).
3. Results
3.1. Patient characteristics
There were 598 640 patients who met the study criteria
with 36.3%, 43.8%, and 20.0% of patients classified as low,
intermediate, and high risk, respectively . Non-
White race was associated with a significantly higher
likelihood of presenting with high-risk disease (22.9% of
Hispanics, 23.8% of Blacks, and 23.3% of those of other races
compared with 19.0% of Whites, all p < 0.001). Insurance
type was also associated with the likelihood of presenting
with high-risk disease (28.3% of the uninsured or those
insured by Medicaid compared with 15.6% of those with
private insurance, p < 0.001).
3.2. Treatment modalities
RP was the dominant treatment modality received by all
patients, increasing from 32.2% in 2004 to 52.2% in 2012
(ptrend < 0.001). During this time EBRT decreased from
33.3% to 25.0% (ptrend < 0.001). The most dramatic change in
utilization was seen for brachytherapy which decreased
from 20.3% in 2004 to only 7.4% in 2012 (ptrend < 0.001).
Results are displayed in A. When stratified by risk
group, receipt of observation and RP increased for low-risk
patients between 2004 and 2012, from 9.2% to 21.3% and
from 29.5% to 51.1%, respectively ( ptrend both < 0.001).
During the same period, receipt of radiation therapy fell
(24.3% to 14.5% for EBRT and 31.7% to 11.1% for
brachytherapy, ptrend both < 0.001). Management trends
stratified by risk group are shown in B–D.
shows the distribution of management strategies by age
group for low-risk patients. Utilization of RP decreased
monotonically with age while use of EBRT and brachyther-
apy increased until falling in those 75 yr or older.
Observation also increased with age reaching a maximum
of 26.8% in those aged 75 yr and older.
A similar pattern was seen for patients with intermedi-
ate-risk disease where receipt of RP increased from 38.5%
to 57.3% (ptrend < 0.001), while receipt of radiation therapy
fell (35.4% to 26.3% for EBRT and 14.5% to 6.5% for
brachytherapy, ptrend both < 0.001). Rates of observation
initially fell in this subgroup but were overall slightly
increased (5.3% to 6.9%). For patients with high-risk disease,
receipt of RP increased significantly, overtaking EBRT as
the most common form of management (25.1% to 43.4%,
ptrend < 0.001).
E U R O P E A N U R O L O G Y 7 1 ( 2 0 1 7 ) 7 2 9 – 7 3 7 730