350 476
Although there are three randomized trials currently
examining the effect of radical therapy in men with
disseminated prostate cancer (STAMPEDE: clinicaltrials.-
gov, NCT00268476; HORRAD:
trialregister.nl, NTR271;
clinicaltrials.gov, NCT01751438), they will not mature for
many years, and we are currently reliant on observational
data. Our study is unique as it is based on a large, well-
annotated, population-based dataset with validated pa-
tient–tumor characteristics including comorbidity score
and reliable cause of death reporting
[23]. Follow-up is long,
and multiple statistical methodologies were adjusted for
differences in patient–tumor characteristics between treat-
ment cohorts and accounted for competing risks of
mortality. To interrogate the validity of our results further,
a number of additional analyses were performed including
to counter any ‘‘immortal-time’’ bias (Supplementary
Tables 1–7 and Supplementary Fig. 1), and all these
supported our main a priori analyses.
Nonetheless, our study has limitations. We used PSA
>
50 to identify a cohort of men at high risk of disseminated
prostate cancer and performed a subgroup analysis of PSA
>
100 cases; nonetheless, our study population might be
heterogeneous with large localized/locally advanced
tumors as well as the lower end of the metastatic spectrum
[24]. We cannot be certain that polymetastatic patients
would have the same findings, and our results might be
more applicable to those with oligometastatic disease and/
or those with large localized/locally advanced tumors. This
is perhaps even more likely, given the overall low rates of
cancer-specific deaths in the study cohort, consistent with
lower disseminated burdens. In addition, many men were
labeled as Mx and combining these with M1 could have
introduced a bias. Further, N staging poses a problem due to
heterogeneous practices in Sweden during the study period.
Some cases presented for surgery may have had synchro-
nous pelvic lymphadenectomy that likely caused stage
migration; others might have undergone intraoperative
frozen sectioning with variable surgeon judgment as to
which proceeded to radical prostatectomy and which did
not. Regarding radiation therapy, some centers often
performed a pelvic lymphadenectomy first and then gave
radiation for N0 cases, and ADT alone for N1 patients. The
Nx cases, on the other hand, could have consisted of those
felt unlikely to have nodal metastases, or because in the
practitioner’s view the results of N staging would not
change management. As a result of such inconsistency in N
staging, we examined baseline characteristics of the Nx and
N1/N0 cohorts, and found significant differences for
unmatched and matched groups, especially regarding M
stage, suggesting that residual confounding as a result of
true N-stage imbalances might bias our results. Hence, we
evaluated baseline characteristics on NxM0 cases and found
well-balanced treatment groups; repeat analyses with this
subcohort showed similar point estimates as in the main
matched cohort. Hence, any N- and M-stage imbalances in
the main matched cohort did not significantly alter our
finding of a survival benefit for initial radical therapy.
Another limitation is that lack of information on
secondary treatments means that we cannot be certain
whether the radical therapy arm received adjuvant or
salvage systemic therapy comparable with the androgen
deprivation cohort. In addition, an unknown number of
patients in the radical therapy group may have also received
androgen deprivation. We also do not have data on the
relative lengths of ADT in the two groups. Hence, unmea-
sured confounding could explain the large differences in
survival seen in this study, and the ‘‘true’’ point estimate may
indeed be lower. Additionally, although we have previously
suggested that significant differences in oncologic outcome
between surgery and radiation for men at high risk of
disseminated prostate cancer would be improbable
[11], the
majority of radically treated cases in the current study
received radiation. Owing to few surgical cases, it is not
possible to compare radical modalities in this analysis. It is
also not possible to investigate any possible benefit of
surgery regarding palliation from pelvic pain, bladder
outflow/ureteral obstruction, and hematuria
[25,26], al-
though we have previously shown radical prostatectomy to
be safe and technically feasible in the metastatic setting
[27].
5.
Conclusions
These data from a large and comprehensive population-
based study suggest that men at high risk of disseminated
prostate cancer have significant oncologic benefit with
initial radical treatment of the primary lesion rather than
initial ADT alone. Ongoing randomized clinical trials will
determine whether radical therapy should be used in the
future to treat men with disseminated prostate cancer.
Author contributions:
Prasanna Sooriakumaran had full access to all the
data in the study and takes responsibility for the integrity of the data and
the accuracy of the data analysis.
Study concept and design:
Sooriakumaran, Wiklund, Nyberg, Steineck.
Acquisition of data:
Nyberg, Sooriakumaran, Akre, Carlsson, Wiklund.
Analysis and interpretation of data:
Wiklund, Steineck, Akre, Widmark,
Graefen, Sooriakumaran.
Drafting of the manuscript:
Sooriakumaran, Nyberg, Akre.
Critical revision of the manuscript for important intellectual content:
Wiklund, Steineck, Hamdy, Widmark.
Statistical analysis:
Nyberg, Steineck.
Obtaining funding:
None.
Administrative, technical, or material support:
None.
Supervision:
Wiklund, Steineck.
Other:
None.
Financial disclosures:
Prasanna Sooriakumaran certifies that all conflicts
of interest, including specific financial interests and relationships and
affiliations relevant to the subject matter or materials discussed in the
manuscript (eg, employment/affiliation, grants or funding, consultan-
cies, honoraria, stock ownership or options, expert testimony, royalties,
or patents filed, received, or pending), are the following: None.
Funding/Support and role of the sponsor:
None.
Acknowledgments:
The PCBaSe database is funded by the Swedish
Research Council 825-2012-5047, and this project was made possible by
the continuous work of the National Prostate Cancer Registry of Sweden
steering group: Pa¨r Stattin (chairman), Anders Widmark, Camilla
Thellenberg, Ove Andre´n, Anna Bill-Axelsson, Ann-Sofi Fransson, Magnus
To¨rnblom, Stefan Carlsson, Marie Hja¨lm-Eriksson, Bodil Westman, Bill
E U R O P E A N U R O L O G Y 7 2 ( 2 0 1 7 ) 3 4 5 – 3 5 1
350