87 148
we speculate that MRI is a more superior modality in terms
of the sensitivity and specificity for detecting bone
metastasis in patients with prostate cancer, given the
well-known limited diagnostic performance of BS in the
literature (ie, pooled sensitivity and specificity of 0.71 and
0.91, respectively)
[[27_TD$DIFF]
21]. However, as our meta-analysis was
based on a per-patient basis and we did not assess the per-
lesion diagnostic performance of MRI, whether MRI can be
used to identify the metastatic burden or as a modality for
assessing global treatment response cannot be answered
from the results of our study.
There was significant heterogeneity among the includ-
ed studies. Based on
[14_TD$DIFF]
meta-regression analysis, the only
statistically significant factor that may be attributable to
this heterogeneity was the number of imaging planes
used for the determination of bone metastasis. Specifi-
cally, studies that used two or more planes (0.99) showed
significantly greater sensitivity compared with those
using only one plane (0.87). Regarding MRI technology,
use of DWI or slice thickness ( 4 or
>
4 mm) was not
shown to affect the heterogeneity. Regarding slice
thickness, the cutoff value of 4 mm was based on practical
guidelines and may not have been a significant factor as
all but two of the included studies had a minimum slice
thickness of 5 or less
[[28_TD$DIFF]
17,20,22,23]. However, it is unclear
why the addition of DWI did not result in superior results
compared with using only conventional sequences of
T1WI, T2WI, or STIR. Of note, CE MRI was not separately
analyzed as a covariate in the current meta-analysis, as it
was used in only two studies, where it was not a
dominant sequence but rather one among several MRI
sequences included in their multiparametric prostate
MRI protocol
[[29_TD$DIFF]
6,16]. Further studies may be needed to
verify the added value of DWI and CE MRI in determining
bone metastasis. Collectively, based on these results
with regard to technical aspects of MRI, guidelines
should recommend that at least two different planes be
used when assessing bone metastasis from prostate
cancer.
The current study highlights that regardless of the
coverage of MRI—that is, whether it covered only the pelvis
in routine prostate MRI, or whether a dedicated axial
skeleton or whole-body MRI was utilized—there was
consistently high per-patient sensitivity (0.94, 0.95, and
0.97, respectively) and specificity (0.99, 0.94, and 0.97,
respectively). This is substantiated by the literature in that
although the use of axial skeleton or whole-body MRI may
allow identification of a greater number of metastatic bone
lesions, the probability of missing a patient with metastasis
is negligible due to the fact that isolated peripheral
metastasis is highly uncommon as prostate cancer primari-
ly metastasizes to the lower spine and pelvis before
spreading throughout the whole body
[[30_TD$DIFF]
4,6,24] .The decision
to simply use routine prostate MRI, or perform additional
axial skeleton or whole-body MRI should be based on
further studies assessing cost effectiveness, where the latter
two would be more effective in terms of per-lesion
detection, but would increase cost in terms of MRI
acquisition time and medical costs.
Table 3 – MRI characteristics
First author
Magnet
strength
(T)
Vendor
Machine
Coverage Sequence used No. of
imaging
planes
Minimum
ST (mm)
Explicit
MRI
criteria
DWI
T1WI
T2WI
STIR
Imaging
plane
ST (mm)
B
values
(s/mm
2
)
Imaging
plane
ST (mm) Imaging
plane
ST (mm) Imaging
plane
ST (mm)
Conde-Moreno
[15]
1.5
Siemens MagnetomAvanto Whole body DWI/T1WI/STIR
1
5
No
A
5
50, 900
A 5/NR
A
5/NR
Kitajima
[
[7_TD$DIFF] 16]1.5 or 3 GE
MR750, Signa
Pelvis
DWI/T1WI/T2WI/
DCE-MRI
1
2.5
No
A
5–7/0–1 0, 600, 1000
A 6/1
A/C/S 2.5–3/0–0.5
Lecouvet
[4]
1.5
Philips
Achieva
Axial skeleton T1WI/T2WI
2
5
Yes
S
5–6/0.5–0.6 S/C
5–6/0.5–0.6
Lecouvet
[5]
1.5
Philips
Achieva
Whole body DWI/T1WI/STIR
2
5
Yes
A
5/0.5
0, 800
[13_TD$DIFF]
C
6/1
C
6/1
Mosavi
[ [8_TD$DIFF] 17]1.5
Philips
Gyroscan Intera Whole body DWI
1
6
Yes
A
6/0
0, 1000
[1_TD$DIFF]
Pasoglou
[
[9_TD$DIFF] 18]3
Siemens Verio
Whole body DWI/T1WI
2
1.2
Yes
A
5/0.5
0, 800
C
1.2/0
Piccardo
[9]
1.5
GE
Signa HDxt
Pelvis
STIR
1
5
Yes
A
5/0.5
Vargas
[
[10_TD$DIFF] 19]1.5 or 3 GE
NR
Pelvis
T1WI
1
5
No
A 5/1
Venkitaraman
[
[11_TD$DIFF] 20]1.5
NR
NR
Axial
skeleton
T1WI/STIR
1
8
Yes
S
8/1
S
8/1
Woo
[6]
3
Siemens,
Philips
Trio/Verio,
Ingenia
Pelvis
DWI/T1WI/T2WI/
DCE-MRI
3
3
Yes
A
3.5/0
0, 1000
A 3–3.5/0
A/C/S 3–4/0–0.4
A = axial; C = coronal; DCE = dynamic contrast enhanced; DWI = diffusion-weighted imaging; MRI = magnetic resonance imaging; NR = not reported; ST = slice thickness; T1WI = T1-weighted imaging; T2WI = T2-weighted
imaging; S = sagittal; STIR = short tau inversion recovery.
E U R O P E A N U R O L O G Y 7 3 ( 2 0 1 8 ) 8 1 – 9 1
87