85 148
bias was low, with a
p
value of 0.23 for slope coefficient
( Fig. 5 ).
3.5.
Exploration of heterogeneity
The results of
[14_TD$DIFF]
meta-regression analyses are shown in
Table 4. Only the number of imaging planes used was shown
to be a significant factor affecting heterogeneity (
p
<
0.01).
Specifically, sensitivity was significantly higher when using
two or more imaging planes (0.99 [95% CI 0.98–1.00])
compared with when using only one plane (0.87 [95% CI
0.80–0.94],
p
<
0.01). Although specificity also showed
‘‘statistically’’ higher values when using two or more planes,
the difference was not considered meaningful: 0.99 (95% CI
0.98–1.00) versus 0.95 (95% CI 0.90–1.00;
p
= 0.01). Other-
wise, clinical setting, reference standard, magnet field
strength, MRI coverage, type of MRI sequences used, and
minimum slice thickness were not shown to be significant
factors affecting the heterogeneity.
The results of sensitivity analyses are shown in
Figure 6. The specificity estimates were comparable with
consistently high values across all subgroups (0.91–0.99). In
general, the sensitivity estimates were also comparable
across most subgroups with pooled sensitivity ranging from
0.93 to 1.00; however, a few subgroups showed slightly
lower pooled sensitivity. Specifically, studies that included
only patients with treated prostate cancer (0.89 [95% CI
0.74–1.00]), used both 1.5- or 3-T scanners (0.89 [95% CI
0.80–0.94]), and used only one imaging plane for analysis
(0.87 [95% CI 0.80–0.94]) tended to show lower sensitivity.
However, pooling the results for studies including patients
with any risk and those using both 1.5- or 3-T scanners were
considered unstable as only two studies were included in
these subgroups. In addition, when a sensitivity analysis
was performed using nine studies, excluding a single study
[[6_TD$DIFF]
15]that (1) was considered to have high concern for
applicability regarding patient selection and (2) showed
particularly inferior diagnostic performance in terms of
both sensitivity and specificity, the degree of heterogeneity
was substantially decreased (
I
2
[4_TD$DIFF]
= 57.14 and 75.64 for
sensitivity and specificity, respectively) with sensitivity of
0.97 (95% CI 0.89–0.99) and 0.98 (95% CI 0.95–0.99).
3.6.
Discussion
In the current meta-analysis, we evaluated the diagnostic
accuracy of contemporary MRI (using scanners with a
magnetic field strength of 1.5 T or higher) for the detection of
bone metastasis in patients with prostate cancer. Our results
show that the pooled per-patient sensitivity and specificity
of the 10 included studies were 0.96 (95% CI 0.87–0.99) and
0.98 (95% CI 0.93–0.99), respectively. Based on this excellent
diagnostic performance of MRI for the detection of bone
metastasis, MRI could be used as one of the primary
modalities for triaging patients with newly diagnosed or
treated prostate cancer and to help decide on the most
optimal management. Although we did not directly compare
the performance between MRI and BS, which would be best
addressed by well-designed randomized controlled trials,
Table 1 – Patient characteristics
First author
No. of patients
Clinical setting
Age
PSA
PSA-DT
a [6_TD$DIFF]Clinical T stage
Gleason score
Total
(
n
)
Metastasis
(
n
)
Metastasis
(%)
New or treated Risk of bone
metastasis
Median Range Median Range Median Range Median Range Median Range
Conde-Moreno
[ [6_TD$DIFF] 15]35
25
71.4
Treated
Unclear
b70
52–80 12
4.54–75.86 NR
NR
T3a
T1–T4 7
5–9
Kitajima
[16]
95
16
16.8
Treated
Unclear
c [7_TD$DIFF]65.7
49–87 2.7
0.58–68.3 NR
NR
NR
NR
7
2–10
Lecouvet
[4]
66
41
62.1
Mixed
High
74
46–85 NR
NR
NR
12
NR
NR
NR
NR
Lecouvet
[5]
100
68
68.0
Mixed
High
69
d53–88 32
e12–78
e5.4/6.7 1.2–11.6 T3b
NR
8
NR
Mosavi
[17]
49
5
10.2
Newly diagnosed High
67
57–80 14
1.3–950
N/A
NA
T3
T1c–T4 9
8–10
Pasoglou
[
[9_TD$DIFF] 18]30
10
33.3
Mixed
High
69
d [8_TD$DIFF]NR
31
dNR
NR
12
NR
NR
NR
NR
Piccardo
[9]
21
6
28.6
Treated
High
78
70–85 4.9
2.2–13.4
NR
NR
NR
NR
8
7–9
Vargas
[19]228
57
25.0
Newly diagnosed Any
63
36–83 6.3
0.4–222
N/A
NA
T1c
T1c–T4 7
6– 8
Venkitaraman
[
[11_TD$DIFF] 20]99
14
14.1
Newly diagnosed High
66
44–83 26.5
2–1600
N/A
NA
NR
NR
7
6–10
Woo
[6]
308
21
6.8
Newly diagnosed Any
68.5
d [10_TD$DIFF]38–91 30.9
d1.2–955.5 N/A
NA
NR
NR
7
6–10
N/A = not available; NR = not reported; PSA = prostate-specific antigen; PSA-DT = prostate-specific antigen doubling time; RP = radical prostatectomy.
a
For treated patients.
b
No metastasis or oligometastasis based on conventional imaging.
c
Clinically suspected recurrence after RP.
d
Mean.
e
For newly diagnosed prostate cancer patients.
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
85