EURURO Vol. 72 No. 4

E98 656

Letter to the Editor
Re: Sungmin Woo, Chong Hyun Suh, Sang Youn Kim,
Jeong Yeon Cho, Seung Hyup Kim. Diagnostic Perfor-
mance of Magnetic Resonance Imaging for the Detection
of Bone Metastasis in Prostate Cancer: A Systematic
Review and Meta-analysis. Eur Urol. In press.
We read with interest the recent report by Woo et al on
the diagnostic performance of magnetic resonance imaging
(MRI) for detection of bone metastases in prostate cancer.
We commend the authors for undertaking a systematic
review and meta-analysis on the value of a promising
diagnostic imaging procedure. However, we believe that
some points warrant further discussion.
First, this meta-analysis only assessed per-patient
diagnostic performance. Although stratifying patients into
groups with and without metastatic disease may be helpful,
this task is much easier than classifying a specific osseous
lesion as being malignant or benign. It is unlikely that the
high sensitivity and specificity reported would hold up in a
per-lesion analysis. Indeed, one of the studies included
reported patient-based sensitivity of 1.0, but the lesion-
based sensitivity of diffusion-weighted imaging (DWI) was
0.56 . Particularly in the case of DWI, specificity has been
reported to be as low as 8% when diagnosing bone
metastases . Moreover, three of the studies included
applied only a single MR sequence, (eg, DWI). Given the fact
that a variety of different bone lesions may cause signal
alterations, it is obvious that a definitive diagnosis could not
have been established in a lesion-based analysis. For
therapeutic decisions, one needs to know about the nature
of a specific lesion. With patient-based data, we know little
about possible false-negative lesions, limiting the efficacy of
potential salvage radiation treatment of bone metastases.
Second, all the studies included used a combination of
imaging/clinical/biological data as the reference standard.
As far as can be ascertained, definitive histopathological
validation of the suggested specificity was only available for
19 of 1031 patients, which is < 2%. Obviously, clinical and
biochemical follow-up has very limited value in confirming
a MRI-based diagnosis, and clear imaging-based validation
strategies have not been applied. A head-to-head compari-
son of MRI and highly specific prostate-specific membrane
antigen (PSMA) ligands for positron emission tomography
could shed light on the true value of MRI as PSMA
expression may also be used to differentiate between vital
and treated metastases. Indeed, six studies included treated
patients, and specificity was calculated without consider-
ation of the vitality of lesions.
Finally, only two multicenter studies were included,
recruiting only 21 and 30 patients, respectively. Of the
1031 patients, 980 were from diagnostic single-center
studies. We know that there is theoretical and empirical
evidence suggesting that these results are likely to be
seriously flawed owing to limited external validity,
implausible effect sizes, and lack of blinding . Selection
bias is likely in four of the studies included that were
retrospective in nature.
In conclusion, the meta-analysis byWoo et al outlines
the high potential value of MRI for detection of bone
metastases. For all of the aforementioned reasons, we raise
a strong word of caution with regard to interpretation of the
reported results. If it seems too good to be true, it probably is.
Both imaging specialists and urologists have to jointly
undertake an effort to generate robust data using prospec-
tive, multicenter approaches with well-defined patient
cohorts and clear validation strategies.
Conflicts of interest: The authors have nothing to disclose.
References
[1] Woo S, Suh CH, Kim SY, Cho JY, Kim SH. Diagnostic performance of
magnetic resonance imaging for the detection of bone metastasis in
prostate cancer: a systematic review and meta-analysis. Eur Urol. In
press. .
[2]
[3]
[4]
[5]
available at
journal homepage:
DOI of original article: .
0302-2838/# 2017 European Association of Urology. Published by Elsevier B.V. All rights reserved.