Page 139 - An Evidence Review of Active Surveillance in Men With Localized Prostate Cancer

As described for Key Question 2, there is not yet consistency among clinicians or researchers
as to the definitions of AS or of WW, the standard protocols for the interventions, or how to
manage patients whose cancers show signs of progression. In addition, studies based on database
analyses frequently could not segregate patients who had AS from those who had WW. It is also
common for analyses to group together patients who had no treatment with those who had ADT
alone. This was particularly the case for analyses of the SEER database, in which it is only
possible to distinguish nonaggressive therapy (e.g., AS or ADT) from prostatectomy or radiation
therapy.
For the most part, the Key Questions, and thus this review, implicitly assume that it is
possible to identify men who are at sufficiently low risk of progression of their prostate cancer
that AS can be a safe and appropriate option for them. This report did not review studies that
evaluated methods of categorizing patients into risk groups. However, our readings and
discussions with experts lead us to the conclusion that additional basic and clinical research is
needed to more accurately classify or predict those men whose diseases are indeed at a low risk
of progression. Presumably, these are the men for whom it would be most appropriate to
consider offering AS.
Also as described above, there are numerous gaps in the evidence regarding the many
specific factors and subgroups of interest to the conference. This section will not attempt to
delineate all the places where evidence is inadequate, but instead will highlight those areas that
our EPC concluded are in most need of future research. The future research needs will be
addressed in the order of the Key Questions.

Key Question 1. Patient Population and Natural History

Changes in Last 30 Years

While there are several gaps in evidence regarding time-trend analyses of specific factors of
interest to the conference sponsors, better understanding of time-trends in the future can be
gained by improving the data being collected and expanding the scope of the major U.S.
databases. In particular, we found that detailed stage and grade information are often not
available requiring researchers to create broad categories that place major limitations on the
analyses. Likewise, future research would be enhanced if more accurate and specific data were
collected about the therapeutic interventions or observational management strategies. As just
mentioned above, one cannot use the SEER database to accurate analyze true WW (or AS), since
these observational management strategies cannot be distinguished from use of ADT. In
addition, the SEER database is inadequate to analyze data from races other then blacks and
whites, since Hispanics/Latinos, Asians, and others are apparently underrepresented, precluding
complete racial analyses. This may require adding new registries to SEER that better represent
other races.
We were also concerned about a potentially important source of bias in the SEER database
which may require resolution to allow for appropriate future analyses on cancer staging.
Analyses of SEER reported summary stage information using the “a combination of the most
f
precise clinical and pathological documentation of the extent of disease.” This may result in
misclassification, as the accuracy of available information on staging depends on the treatment
patients receive. For patients who have had RP, pathological staging information is available,
whereas for those who have received RT, ADT monotherapy, AS, or WW, only clinical staging

f See the SEER staging manual, available at: http://seer.cancer.gov/tools/ssm/ (last accessed September 30, 2011).

85