Professional Documents
Culture Documents
Woo 2017
Woo 2017
Woo 2017
PII: S0720-048X(17)30305-4
DOI: http://dx.doi.org/doi:10.1016/j.ejrad.2017.07.021
Reference: EURR 7910
Please cite this article as: Woo Sungmin, Suh Chong Hyun, Kim Sang Youn, Cho Jeong
Yeon, Kim Seung Hyup.Diagnostic Performance of MRI for Prediction of Muscle-
invasiveness of Bladder Cancer: A Systematic Review and Meta-analysis.European
Journal of Radiology http://dx.doi.org/10.1016/j.ejrad.2017.07.021
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.
1
Sungmin Woo, M.D. a, Chong Hyun Suh, M.D. b, c, Sang Youn Kim, M.D. a,
Jeong Yeon Cho, M.D. a, d, Seung Hyup Kim, M.D. a, d
a
Department of Radiology, Seoul National University College of Medicine, 101 Daehak-ro,
Jongno-gu, Seoul 110-744, Korea
b
Department of Radiology and Research Institute of Radiology, University of Ulsan College
of Medicine, Asan Medical Center, 86 Asanbyeongwon-Gil, Songpa-Gu, Seoul 138-736,
Republic of Korea
c
Department of Radiology, Namwon Medical Center, 365, Chungjeong-ro, Namwon-si,
Jeollabuk-do 590-702, Republic of Korea
d
Institute of Radiation Medicine and Kidney Research Institute, Seoul National University
Medical Research Center, Seoul 110-744, Korea
Abstract
Purpose: To review the diagnostic performance of ≥1.5-Tesla MRI for local staging of bladder cancer.
Methods: MEDLINE and EMBASE were searched up to February 21, 2017. We included diagnostic accuracy
studies published since 2000 that used ≥1.5-Tesla MRI for local staging (≥T2 [muscle-invasive]) in patients with
bladder cancer, using pathology as the reference standard. The methodological quality was assessed using Quality
Assessment of Diagnostic Accuracy Studies-2. Sensitivity and specificity were pooled and plotted in a hierarchical
summary receiver operating characteristics plot. Sensitivity analyses using several clinically relevant covariates
were performed.
Results: 24 studies (1774 patients) were included. Pooled sensitivity was 0.92 (95% CI 0.88-0.95) with specificity
of 0.87 (95% CI 0.78-0.93). Sensitivity analyses showed that sensitivity estimates were comparable and
2
consistently high across all subgroups, but specificity estimates were variable. Studies using 3-Tesla scanners had
higher specificity (0.93 [95% CI 0.86-0.98]) than those using 1.5-Tesla scanners (0.83 [95% CI 0.74-0.98]).
Studies using multiparametric MRI (conventional + ≥2 functional sequences) showed the highest accuracy with
sensitivity and specificity of 0.94 (95% CI 0.89-1.00) and 0.95 (95% CI 0.89-0.98), respectively.
Conclusions: MRI shows good diagnostic performance for predicting muscle-invasiveness of bladder cancer.
Abbreviations:
CE = contrast-enhanced
Key words: Bladder cancer; Muscle-invasive bladder cancer; Local staging; Magnetic
Resonance Imaging; Systematic review; Meta-analysis
1. Introduction
In bladder cancer, determining the depth of invasion is of paramount importance. Management options differ
considerably for muscle-invasive bladder cancer (MIBC) with stage T2 or higher and non-MIBC with stage T1
or lower. MIBC is usually treated with radical cystectomy, or a combination of radiation and/or chemotherapy
while non-MIBC can be managed with transurethral resection (TUR) [1, 2]. Therefore, it would be valuable if
preoperative imaging studies can accurately assess the depth of invasion of bladder cancer.
3
During the past two decades, there has been great interest in the use of magnetic resonance imaging (MRI) for
local staging of bladder cancer paralleling advances in MRI technique (i.e., widespread use of scanners with
magnet strength of ≥1.5 Tesla). Variable techniques including conventional T1- (T1WI) and T2-weighted imaging
(T2WI) and more advanced sequences such as contrast-enhanced (CE) MRI and diffusion-weighted imaging
(DWI) has shown promising results for determining the depth of invasion [3-5]. However, in clinical practice,
TUR followed by pathological investigation is used for local staging, and current guidelines state that due to (1)
the variability of diagnostic performance for local staging using MRI in the literature (73% to 96%) and (2)
insufficient data on the use of advanced MRI techniques (i.e., DWI), no recommendations can be made regarding
their use in the management of bladder cancer [6, 7]. In this clinical context, a comprehensive review of the
existing literature on the diagnostic performance of MRI in local staging of bladder cancer may be needed to
Therefore, we performed a systematic review and meta-analysis to evaluate the diagnostic performance of
contemporary MRI for local staging of bladder cancer with focus on the prediction of muscle-invasiveness.
The current meta-analysis was written according to the Preferred Reporting Items for Systematic Reviews and
Meta-Analyses (PRISMA) guidelines. The research question for the purpose of this meta-analysis was formulated
based on the following PICOS criteria [8]: What is the diagnostic performance of ≥1.5-Tesla MRI for local staging
in patients with bladder cancer, as compared with pathological results in original articles published since 2000?
A computerized search of MEDLINE and EMBASE databases up to February 21, 2017 was performed to
identify relevant studies. The search query combined synonyms for bladder cancer, MRI, staging, and diagnostic
accuracy as the following: (bladder OR urothelial OR papillary OR transitional) AND (cancer OR carcinoma OR
tumor OR tumour OR neoplasm) AND (MRI OR MR OR "magnetic resonance") AND (stage OR staging OR
TNM) AND (sensitivity OR specificity OR accuracy OR "predictive value"). References of the identified papers
were screened to expand the scope of literature search. The search was not limited to any particular language.
We included studies that met the following PICOS criteria (10): (a) patients diagnosed with bladder cancer; (b)
MRI used as the index test for local staging (assessment of stage ≥T2 [MIBC]); (c) pathology based on cystectomy
or TUR for comparison; (d) sufficient information for reconstruction of 2 x 2 tables regarding sensitivity and
The exclusion criteria were as follows: (1) study population of less than 10 patients; (2) publication type other
than original articles (i.e., review articles, letters, editorials, conference abstracts, and etc.); (3) MRI used for
assessment of bladder cancer, but focusing on other topics (i.e., treatment response, local staging but not MIBC
vs non-MIBC, nodal or distant metastasis staging); (4) studies using <1.5-Tesla MRI scanners; (5) studies
published before 2000; (6) overlapping patient population; and (7) insufficient data for reconstruction of 2 x 2
tables. When multiple studies with overlapping study population were present, we only included the study with
the largest study population. Authors of the studies were contacted when 2 x 2 tables could not be reconstructed
The literature search and study selection process was independently performed by two reviewers and
The following data with regards to patient/tumor, study, and MRI characteristics were extracted from the
(1) Patient and tumor characteristics―number of patients, ethnicity of study population, mean age and range
of patients, number of tumors, prevalence of stage ≥T2 tumors, and histological subtypes (only urothelial
(2) Study characteristics―origin of study (authors, institution, and duration of patient recruitment), publication
year, study design (i.e., prospective, multi-center, and consecutive enrollment), reference standard (based on
cystectomy or TUR), interval between MRI and reference standard, blinding between MRI interpretation and
pathological assessment, and characteristics of readers (number, consensus reading, and experience)
(3) MRI characteristics―scanner manufacturer and model, magnet field strength (1.5- or 3.0-Tesla), type of
MRI sequences used, their corresponding technical parameters, whether preparation for bladder distension was
performed or not, whether MRI was performed before or after a prior TUR or biopsy, and whether criteria for
The methodological quality of the included studies was assessed using the Quality Assessment of Diagnostic
Accuracy Studies-2 (QUADAS-2) tool [9]. Data extraction and quality assessment was performed independently
Two x two tables were reconstructed from the included studies to calculate their sensitivity and specificity. If
several different diagnostic performance values were provided within a single study (i.e., several types of MRI
sequences or multiple independent readers), we selected the result with the highest accuracy to be representative
of the study.
Summary estimates of sensitivity and specificity were calculated with hierarchical logistic regression modelling
including bivariate and hierarchical summary receiver operating characteristic (HSROC) modelling [10, 11].
These results were plotted using HSROC curves with 95 % confidence and prediction regions. Publication bias
was evaluated using visual analysis of the Deeks’ funnel plot and calculating the p value using Deeks’ asymmetry
test [12].
Heterogeneity was assessed with the following methods: (1) Cochran’s Q-test with p <0.05 indicating that
heterogeneity was present; (2) Higgins I2 test with the degree of heterogeneity interpreted using the following
criteria: inconsistency index (I2) = 0%–40%, heterogeneity might not be important; 30%–60%, moderate
heterogeneity may be present; 50%–90%, substantial heterogeneity may be present; and 75%–100%, considerable
heterogeneity [13]; and (3) testing for a threshold effect (positive correlation between sensitivity and false positive
rate) between the included studies. In order to explore the cause of heterogeneity, sensitivity analyses using several
The “metandi” and “midas” modules in Stata 10.0 (StataCorp LP, College Station, TX, USA) and “mada”
package in R software version 3.2.1 (R Foundation for Statistical Computing, Vienna, Austria) were used for
3. Results
The systematic literature search initially yielded 1062 articles. After removing 387 duplicates, screening of the
675 titles and abstracts yielded 70 potentially eligible original articles. Full-text reviews were performed and 46
studies were excluded due to the following: less than 10 patients (n = 1), not in field of interest (n = 4), used <1.5-
6
Tesla MRI (n = 24), published before 2000 (n = 11), insufficient data to reconstruct 2 x 2 tables (n = 3), and shared
study population with other studies (n = 3). Ultimately, 24 studies including 1774 patients evaluating the
diagnostic performance of MRI for local staging of bladder cancer were included in this meta-analysis [3-5, 14-
The patient characteristics are summarized in Table 1. The size of the study population ranged from 19 to 362
patients. The patients were of Asian ethnicity in 13 studies and non-Asian in the other 11. The mean age of the
patients ranged from 61.3 to 72.1 years. The number of bladder cancers ranged from 20 to 362. The prevalence
of stage ≥T2 tumors ranged from 14% to 96%. 10 studies only included patients with urothelial carcinoma, while
other histological subtypes were also included in 10 and details were not provided in four.
The study characteristics are described in Table 2. The study design was prospective in 10 studies, and
retrospective in 14. All but two studies were single-center studies. Patient recruitment was consecutive in all but
two studies. Four studies used only cystectomy as the reference standard; three used only TUR; and 16 used either
cystectomy or TUR. The interval between MRI and the reference standard was provided in 11 studies. MRI was
interpreted blinded to the reference standard in most (17 of 24) of the studies.
The MRI characteristics are shown in Table 3. Six studies used 3-Tesla scanners, while 18 used 1.5-Tesla
scanners. Various sequences were used in the included studies. Three studies used only conventional sequences
(T1WI and/or T2WI); six used only a single functional imaging technique; 10 used one functional imaging
technique in addition to conventional sequences; and five used two or more functional imaging techniques in
prohibiting the patients from urination, oral hydration, or saline injection via Foley catheter was performed in 11
studies. MRI was performed before TUR or biopsy in all patients in 14 studies, in some in three studies, and after
TUR or biopsy in five. The criteria for local staging on MRI was explicit in most (20 of 24) studies.
In general, the quality of the studies was considered moderate, with 21 of the 24 studies satisfying at least four
of the seven QUADAS-2 domains (Figure 2). Regarding the patient selection domain, two studies were considered
to have high risk of bias as they excluded patients with stage T3-4 bladder cancer when differentiating between
≤T1 and ≥T2 tumors [5, 24]. Another two studies were considered to have unclear risk of bias as they did not
explicitly mention whether patient enrollment was consecutive or not [22, 29]. Regarding the index test domain,
7
there was high risk of bias in two studies as the threshold used for determining local staging was not pre-specified
[3, 33]. There was unclear risk of bias in 10 studies, as blinding was unclear in seven studies [15-18, 22, 29, 31]
and it was unclear whether a pre-specified threshold was used in three studies [20, 28, 34]. Regarding the reference
standard domain, all but five [4, 20, 25, 31] of the studies had high risk of bias as the reference standard was based
on TUR is all or some of the patients. However, those five studies had unclear risk of bias, as it was only described
that pathological results were used and detailed methods including the use of TUR or cystectomy was not
mentioned in one study [31] and it was unclear whether the pathology was assessed blinded to MRI in four studies
[4, 20, 22, 25]. Regarding the flow and timing domain, 16 studies had high risk of bias as different reference
standards were applied within the study [3, 5, 14-17, 19, 23, 24, 26, 28-30, 32-34]. Five studies were considered
to have unclear risk of bias as the MRI-reference standard interval was not provided [4, 20, 22, 25, 31]. There was
low concern for applicability with regard to the first three QUADAS-2 domains for all but one study [31], in
which there was unclear concern as the details for deriving the reference standard (TUR or cystectomy) was not
explicit.
The pooled sensitivity and specificity for diagnosing MIBC were 0.92 (95% CI 0.88-0.95) and 0.87 (95% CI
0.78-0.93), respectively. The area under the HSROC curve was 0.95 (95 % CI 0.93–0.97) (Figure 3). The
Cochran’s Q-test suggested that heterogeneity was present (Q = 74.049, p <0.001). Higgins I 2 statistics
demonstrated substantial and considerable heterogeneity with regard to the sensitivity (I 2 = 70.65%) and
specificity (I2 = 90.24%), respectively. No threshold effect was shown upon visualization of the coupled forest
plot of sensitivity and specificity (Figure 4) with a correlation coefficient between sensitivity and false positive
rate of 0.270 (95% CI -0.150-0.607). The Deeks’ funnel plot suggested that publication bias was present, with a
Figure 6 shows the results of sensitivity analyses. In general, sensitivity estimates were comparable and high
while there was notable variation in the specificity estimates. Studies assessing only urothelial carcinoma (n = 10)
showed lower specificity (0.79 [95% CI 0.65-0.98]) than those evaluating urothelial carcinoma and other tumors
(n = 10; 0.91 [95% CI 0.83-0.98]). Studies using only cystectomy as the reference standard (n = 4) showed lower
specificity (0.45 [95% CI 0.11-0.98]) than those using TUR only (n = 3; 0.84 [95% CI 0.61-0.98]) or both (n =
16; 0.91 [95% CI 0.86-0.98]). Studies using 3-Tesla scanners (n = 6) had higher specificity (0.93 [95 % CI 0.86-
8
0.98]) than those using 1.5-Tesla scanners (n = 18; 0.83 [95% CI 0.74-0.98]). There was a trend for higher
sensitivity (0.94 [95% CI 0.89-1.00]) and specificity (0.95 [95% CI 0.89-0.98]) in studies (n = 5) using
multiparametric approach (conventional sequence + ≥2 functional techniques) than those using only conventional
sequence (n = 3; 0.86 [95% CI 0.73-0.99] and 0.83 [95% CI 0.60-0.98], respectively) or a single functional
technique (n = 6; 0.89 [95% CI 0.80-0.97] and 0.81 [95% CI 0.62-0.98], respectively). Two studies [4, 22] were
considered to be outliers showing peculiarly low specificity (0.00 and 0.29) compared with others (0.76-1.00).
Pooling the results of 22 studies excluding these two studies resulted in a sensitivity and specificity of 0.91 (95%
CI 0.87-0.94) and 0.89 (95% CI 0.84-0.93), respectively, with a substantial decrease in the degree of heterogeneity
(I2) from 70.65% to 65.66% and from 90.24% to 75.91% for sensitivity and specificity, respectively.
4. Discussion
In this meta-analysis, we evaluated the diagnostic performance of ≥1.5-Tesla MRI for local staging in bladder.
To the best of our knowledge, this is the only meta-analysis on this subject other than the study by Zhai et al [35],
which evaluated only eight studies using only DWI without any exploration of heterogeneity. The current meta-
analysis provides a comprehensive overview of the literature published since 2000, encompassing a large number
of studies (n = 24) using various MRI techniques, and provides the results of heterogeneity exploration using
extensive sensitivity analysis. Our results show that the pooled sensitivity and specificity of the 24 included studies
for differentiating MIBC from non-MIBC were 0.92 (95% CI 0.88-0.95) and 0.87 (95% CI 0.78-0.93),
respectively. The differentiation between MIBC and non-MIBC is crucial, as they show different prognosis and
generally are treated with different approaches. Therefore, it may be suggested based on the results of our meta-
analysis that MRI be incorporated in guidelines for bladder cancer as one of the primary modalities for local
staging.
There was significant heterogeneity among the included studies. However, extensive sensitivity analysis
provided some insight regarding the potential causes of heterogeneity. In general, the sensitivity estimates were
relatively comparable while variation was more notable with regard to the specificity estimates. Specifically,
studies that only evaluated urothelial carcinomas and those that used only cystectomy as the reference standard
showed lower sensitivity that other subgroups. We speculate that this was attributed to the studies by Daneshmand
et al [4] and Nishimura et al [22], which showed peculiarly lower specificity (0.29 and 0.00, respectively) than
the other 22 studies (0.76-1.00), rather than reflecting a true association with these covariates (pathological
9
subtype and reference standard). In fact, when pooling the results of 22 studies excluding these two studies, there
was a remarkable decrease in the degree of heterogeneity. We cannot fully understand why these two studies
showed such different results from the other included studies. However, the study by Nishimura et al [22] was
based on a study population predominantly (18/27) consisting of chemotherapy-treated bladder cancer patients.
We speculate that this may have one of the major reasons why the study suffered from peculiarly low specificity.
Furthermore, although we performed sensitivity analysis for each subgroup regarding reference standard
(cystectomy only, TUR only, cystectomy or TUR), the 95% confidence intervals for summary estimates were
relatively wide as there were only a few number studies included in the cystectomy only group (n = 4) and TUR
only group (n = 3). Therefore, firm conclusions cannot be made regarding reference standard as a source of
The current study highlights that the MRI protocol may largely influence the diagnostic accuracy of local
conventional sequences (T2WI), resulted in higher accuracy compared with using only conventional sequences
or only a single functional technique. T2WI is able to differentiate the layers of the bladder as it can be acquired
with high spatial resolution. On the other hand, functional techniques also have strengths in that they can better
depict the tumor itself (DWI), or provide other imaging features that aid in the local staging such as tumor stalk
(DWI) and submucosal linear enhancement (CE-MRI) [5, 30]. Therefore, by comprehensively utilizing both the
anatomical information and functional information from T2WI and DWI or CE-MRI, the performance of MRI for
predicting muscle-invasiveness of bladder cancer may be improved. Furthermore, studies using 3-Tesla MRI
showed better specificity than those using 1.5-Tesla MRI. Therefore, it may be suggested that guidelines
recommend multiparametric 3-Tesla MRI in order to improve the diagnostic accuracy of local staging of bladder
cancer.
It is well known that optimal distension of the bladder is needed when using computed tomography or MRI for
local staging of bladder cancer. In the current meta-analysis, there was no significant differences in the accuracy
between studies that performed bladder distension and the others. However, robust conclusions cannot be made
for several reasons: (1) there were no studies that explicitly stated that they did not perform bladder distension;
(2) studies that did perform them used various methods including prohibiting patients from voiding, oral hydration,
saline infusion via Foley catheter or a combination of the above; and (3) we are only aware that those studies did
10
perform such measures, but do not know whether the bladders were “optimally distended” or not. Further studies
As several of the included studies provided multiple diagnostic performance values from multiple readers and
MRI protocols within the identical study population, we had to choose one to represent each study. For this meta-
analysis, we included the results with the highest accuracy, which generally corresponded to the results from the
more experienced reader and advanced MRI technique. As it is generally known the diagnostic accuracy of pelvic
MRI interpretations are more accurate by more experienced readers, there may be concern that the pooled
estimates from our study may not be directly applicable to real life clinical practice [36, 37]. However, the
differences between multiple diagnostic performance values were mostly comparable between readers―therefore,
had we selected the results from the less experienced readers there would only have been minor changes in the
One of the major limitations of our meta-analysis is that most of the studies used either TUR or cystectomy (n
= 16) or only TUR (n = 3) as the reference standard. However, as compared with cystectomy, TUR may be less
accurate in determining the muscle-invasiveness of bladder cancer, therefore resulting in a less robust reference
standard. In order to deal with this issue, additional measures (i.e., secondary TUR or additional deep muscle
biopsy) were taken in 9 of the studies that used TUR. Another limitation of this meta-analysis is that we only
assessed the diagnostic performance for differentiating MIBC and non-MIBC but did not perform a match
between MRI and pathology for each T stage. However, we decided that determination of muscle-invasiveness of
bladder cancer would be clinical relevant as this is one of the most important factors regarding prognosis and
treatment decision. In addition, publication bias was present among the included studies in this meta-analysis. We
believe that this bias may have been present in our meta-analysis as studies with more significant results are more
likely to be published in the literature. Therefore, there may be a possibility that the diagnostic performance of
MRI in this meta-analysis for local staging of bladder cancer may be overestimated.
5. Conclusions
MRI shows good diagnostic performance for prediction of muscle-invasiveness of bladder cancer. Using
References
1. Josephson D, Pasin E, Stein JP. Superficial bladder cancer: part 2. Management. Expert Rev. Anticancer Ther.
7 (2007) 567-581.
2. Hautmann RE, Abol-Enein H, Hafez K, et al. Urinary diversion. Urology. 69 (2007) 17-49.
3. Rosenkrantz AB, Haghighi M, Horn J, et al. Utility of quantitative MRI metrics for assessment of stage and
grade of urothelial carcinoma of the bladder: preliminary results. Am. J. Roentgenol. 201 (2013) 1254-1259.
4. Daneshmand S, Ahmadi H, Huynh LN, Dobos N. Preoperative staging of invasive bladder cancer with dynamic
gadolinium-enhanced magnetic resonance imaging: results from a prospective study. Urology. 80 (2012) 1313-
1318.
5. Wang HJ, Pui MH, Guan J, et al. Comparison of Early Submucosal Enhancement and Tumor Stalk in Staging
Bladder Urothelial Carcinoma. Am. J. Roentgenol. (2016) Aug 1-7 [Epub ahead of print].
6. Babjuk M, Bohle A, Burger M, et al. EAU Guidelines on Non-Muscle-invasive Urothelial Carcinoma of the
7. Alfred Witjes J, Lebret T, Comperat EM, et al. Updated 2016 EAU Guidelines on Muscle-invasive and
8. Liberati A, Altman DG, Tetzlaff J, et al. The PRISMA statement for reporting systematic reviews and meta-
analyses of studies that evaluate health care interventions: explanation and elaboration. J. Clin. Epidemiol. 62
(2009) e1-34.
9. Whiting PF, Rutjes AW, Westwood ME, et al. QUADAS-2: a revised tool for the quality assessment of
10. Suh CH, Park SH. Successful Publication of Systematic Review and Meta-Analysis of Studies Evaluating
11. Lee J, Kim KW, Choi SH, Huh J, Park SH. Systematic Review and Meta-Analysis of Studies Evaluating
Diagnostic Test Accuracy: A Practical Review for Clinical Researchers-Part II. Statistical Methods of Meta-
12. Deeks JJ, Macaskill P, Irwig L. The performance of tests of publication bias and other sample size effects in
systematic reviews of diagnostic test accuracy was assessed. J. Clin. Epidemiol. 58 (2005) 882-893.
13. Higgins J, Green S. Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0. The Cochrane
Collaboration. http://handbook.cochrane.org/chapter_9/9_5_2_identifying_and_measuring_heterogeneity.htm.
12
14. Abdel-Rahman HM, El Fiki IM, Desoky EA, Elsayed ER, Samad KMA. The role of diffusion-weighted
magnetic resonance imaging in T staging and grading of urinary bladder cancer. Egyptian J. Radiol. Nucl. Med.
46 (2015) 741-747.
15. Afifi AH, Maksoud TASA, EL-noueam KI, Ataa MA, Abdallah DM. Multiparametric-MRI as a
comprehensive study in evaluation, characterization & local staging of urinary bladder carcinomas. Egyptian J.
16. Ghafoori M, Shakiba M, Ghiasi A, Asvadi N, Hosseini K, Alavi M. Value of MRI in local staging of bladder
17. Hayashi N, Tochigi H, Shiraishi T, Takeda K, Kawamura J. A new staging criterion for bladder carcinoma
using gadolinium-enhanced magnetic resonance imaging with an endorectal surface coil: a comparison with
18. Kobayashi S, Koga F, Yoshida S, et al. Diagnostic performance of diffusion-weighted magnetic resonance
imaging in bladder cancer: potential utility of apparent diffusion coefficient values as a biomarker to predict
19. Lee M, Shin SJ, Oh YT, et al. Non-contrast magnetic resonance imaging for bladder cancer: fused high b
value diffusion-weighted imaging and T2-weighted imaging helps evaluate depth of invasion. Eur. Radiol. (2017)
20. Lista F, Andres G, Caceres F, Ramon de Fata F, Rodriguez-Barbero JM, Angulo JC. Evaluation of risk of
muscle invasion, perivesical and/or lymph node affectation by diffusion-weighted magnetic nuclear resonance in
the patient who is a candidate for radical cystectomy. Actas. Urol. Esp. 37 (2013) 419-424.
21. Mizuno K, Sasaki T, Saito Y, et al. [Gadolinium-enhanced MR imaging, T2-weighted MR imaging, and
22. Nishimura K, Fujiyama C, Nakashima K, Satoh Y, Tokuda Y, Uozumi J. The effects of neoadjuvant
chemotherapy and chemo-radiation therapy on MRI staging in invasive bladder cancer: comparative study based
on the pathological examination of whole layer bladder wall. Int. Urol. Nephrol. 41 (2009) 869-875.
23. Ohgiya Y, Suyama J, Sai S, et al. Preoperative T staging of urinary bladder cancer: efficacy of stalk detection
and diagnostic performance of diffusion-weighted imaging at 3T. Magn. Reson. Med. Sci. 13 (2014) 175-181.
24. Panebianco V, De Berardinis E, Barchetti G, et al. An evaluation of morphological and functional multi-
13
parametric MRI sequences in classifying non-muscle and muscle invasive bladder cancer. Eur. Radiol. (2017)
25. Rabie E, Faeghi F, Izadpanahi MH, Dayani MA. Role of Dynamic Contrast-Enhanced Magnetic Resonance
26. Rajesh A, Sokhi H, Fung R, Mulcahy K, Bankart M. Bladder cancer: evaluation of staging accuracy using
27. Sevcenco S, Ponhold L, Heinz-Peer G, et al. Prospective evaluation of diffusion-weighted MRI of the bladder
as a biomarker for prediction of bladder cancer aggressiveness. Urol. Oncol. 32 (2014) 1166-1171.
28. Sherif MF. The value of diffusion weighted MR imaging in T staging and correlation with histologic grading
tomography or magnetic resonance imaging - our experiences in determining preoperative TNM staging of
bladder cancer. Pril (Makedon Akad Nauk Umet Odd Med Nauki). 34 (2013) 63-70.
30. Takeuchi M, Sasaki S, Ito M, et al. Urinary bladder cancer: diffusion-weighted MR imaging--accuracy for
diagnosing T stage and estimating histologic grade. Radiology. 251 (2009) 112-121.
31. Tillou X, Grardel E, Fourmarier M, et al. [Can MRI be used to distinguish between superficial and invasive
32. Watanabe H, Kanematsu M, Kondo H, et al. Preoperative T staging of urinary bladder cancer: does diffusion-
weighted MRI have supplementary value? Am. J. Roentgenol. 192 (2009) 1361-1366.
33. Wu LM, Chen XX, Xu JR, et al. Clinical value of T2-weighted imaging combined with diffusion-weighted
imaging in preoperative T staging of urinary bladder cancer: a large-scale, multiobserver prospective study on
34. Yamada Y, Kobayashi S, Isoshima S, Arima K, Sakuma H, Sugimura Y. The usefulness of diffusion-weighted
magnetic resonance imaging in bladder cancer staging and functional analysis. J. Cancer Res. Ther. 10 (2014)
878-882.
35. Zhai N, Wang YH, Zhu LM, et al. Sensitivity and Specificity of Diffusion-Weighted Magnetic Resonance
36. Woo S, Kim SY, Cho JY, Kim SH. Assessment of deep myometrial invasion of endometrial cancer on MRI:
37. Wibmer A, Vargas HA, Donahue TF, et al. Diagnosis of Extracapsular Extension of Prostate Cancer on
Prostate MRI: Impact of Second-Opinion Readings by Subspecialized Genitourinary Oncologic Radiologists. Am.
Figure legends
Fig. 2. Grouped bar charts show risk of bias (left) and concerns for applicability (right) of 24 included studies
using QUADAS-2.
Fig. 3. Hierarchical summary receiver operating characteristic curve of the diagnostic performance of MRI for
Fig. 4. Coupled forest plots of pooled sensitivity and specificity. Numbers are pooled estimates with 95%
confidence intervals (CI) in parentheses. Corresponding heterogeneity statistics are provided at bottom right
Fig. 5. Deeks’ funnel plot for per-patient analysis. P-value of 0.01 suggests that publication bias is present.
Fig . 6. Forest plots of sensitivity analyses showing pooled sensitivity and specificity estimates and
corresponding 95% confidence intervals (CIs) for various subgroups. TUR = transurethral resection; * = unstable
pooled estimates
16
Figure 1
Figure 2
17
Figure 3
18
Figure 4
19
Figure 5
Figure 6
20
Cystectomy
Nishimura K [22] 2009 NR Saga University No No NR NR NR NR NR NR
only
Cystectomy or Yes Yes
Ohgiya Y [23] 2014 2007.6-2009.9 Showa University School of Medicine No No Yes 0-41 2 10/10
TUR a
Cystectomy or Yes Yes
Panebianco V [24] 2017 2016.1-2016.11 Sapienza University of Rome No No Yes 2-8wks 2 10/3
TUR
Shahid Beheshti University of Medical Sciences, Al-Zahra Cystectomy Yes
Rabie E [25] 2016 2013.1-2014.7 No Yes Yes NR 1 N/A NR
Hospital, Shahrekord University of Medical Sciences only
Cystectomy or Yes
Rajesh A [26] 2011 NR Leicester General Hospital No No Yes NR 1 N/A 5
TUR
Cystectomy or Yes
Rosenkrantz AB [3] 2013 NR NYU Langone Medical Center, Mount Sinai School of Medicine No Yes Yes 1-84 2 Yes NR/NR
TUR
Sevcenco S [27] 2014 2012-2013 Medical University of Vienna Yes No Yes TUR only 1-58 2 NR 8/6 Yes
Cystectomy or Yes
Sherif MF [28] 2015 2012.2-2014.4 Tanta University Yes No Yes NR 1 NR NR
TUR
Cystectomy or
Stojovska-Jovanovska E [29] 2013 NR University Radiology Clinic (Skopje, Republic of Macedonia) No No NR NR NR NR NR NR
TUR a
Cystectomy or No
Takeuchi M [30] 2009 2006.7-2007.11 Nagoya City University Graduate School of Medical Sciences Yes No Yes NR 2 21/15 Yes
TUR a
Tillou X [31] 2008 2002.5-2005.11 CHU Sud Yes No Yes NR NR 2 No NR NR
Cystectomy or
Wang HJ [5] 2016 2011.11-2014.12 First Affiliated Hospital of Sun Yat-sen University Yes No Yes <2wks 2 No 29/5 Yes
TUR a
Cystectomy or No Yes
Watanabe H [32] 2009 2007-2008 Gifu University Hospital No No Yes 6-30 2 NR
TUR
Cystectomy or No Yes
Wu LM [33] 2013 2007.11-2012.5 Ren Ji Hospital Yes No Yes <29 3 10/4/2
TUR a
Cystectomy or
Yamada Y [34] 2014 2006.2-2013.1 Mie University School of Medicine No No Yes NR 1 N/A NR Yes
TUR b
N/A = not available; NR = not reported; TUR = transurethral resection
a
With deep muscle biopsy
b
With confirmatory repeat TUR
22
Mizuno K [21] Siemens Magnetom Vision 1.5 T2 or CE* P 4/0.2 P 4/0.2 19 Yes Yes Yes
Ohgiya Y [23] Siemens Magnetom Trio 3 T2/DWI A,S 4/0.6 A 3/0.6 0,500,1000 NR Yes Yes
T2/DWI/DT 0,500,800,
Panebianco V [24] GE Discovery MR750 3 A,S 3/NR A 3/NR NR 3/NR 20-180 Yes Some Yes
I/CE 1000
T1/T2/STIR
Rabie E [25] Siemens Magnetom Avanto 1.5 A,C,S 4/1 A 3.6/0.9 NR NR No Yes
/CE
Magnetom
Rajesh A [26] Siemens 1.5 T2/CE A,C,S 5/1 A 3/0.6 20/70 NR Some Yes
Symphony
Rosenkrantz AB [3] Siemens Magnetom Avanto 1.5 T2 A,S 4/NR NR Yes Yes
50,400,100
Sevcenco S [27] Siemens Tim Trio 3 DWI A 5/NR NR Yes Yes
0
T1/T2/DWI/
Sherif MF [28] GE NR 1.5 A,C,S 3/NR A,S 3/NR 0,1000 A 3/NR NR NR Yes No
CE
Stojovska-
Siemens Magnetom Avanto 1.5 CE A,C,S 6/2 20/40/60 Yes No Yes
Jovanovska E [29]
30/60/90/1
Takeuchi M [30] Philips Gyroscan Intera 1.5 T2/DWI/CE A,S 4/0.4 A,S 4/0.4 0,1000 P 3/0 Yes Yes Yes
20
Tillou X [31] GE NR 1.5 T1/T2/CE Yes Yes Yes
A, O
Wang HJ [5] Siemens Magnetom Trio 3.0 DWI 4/0.4 0,1000 NR Yes Yes
(P)
Watanabe H [32] Philips Intera Achieva 1.5 T1/T2/CE A,S 5/2 P 5/0 30/95/120 NR Some Yes
Wu LM [33] GE Signa HDx 3 T2/DWI A,S 5/1 A 3/1 0,1500 Yes Yes Yes
T1/T2/DWI/
Yamada Y [34] GE Signa Excite HDxt 1.5 A,C,S 5/2 A 5/2 0,1000 A,C,S 5/2 NR Yes NR No
CE
A = axial; C= coronal; CE = contrast-enhanced; DWI = diffusion-weighted imaging; DTI = diffusion-tensor imaging; NR = not reported; OS
= oblique sagittal; P = perpendicular to tumor base; ST = slice thickness; TUR = transurethral resection; T1 = T1-weighted imaging; T2 = T2-
weighted imaging; S = sagittal; STIR = short tau inversion recovery