ORIGINAL RESEARCH

Intravoxel Incoherent Motion Diffusion-

Weighted Imaging in Assessing Bladder
Cancer Invasiveness and Cell Proliferation
Fang Wang, MD, Lian-Ming Wu, MD, PhD, Xiao-Lan Hua, BS, Zi-Zhou Zhao, BS,
Xiao-Xi Chen, MD, and Jian-Rong Xu, MD, PhD*

Background: Nonmuscle-invasive bladder cancer (NMIBC, Stage T1 or lower) is treated with transurethral resection
(TUR), while muscle-invasive bladder cancer (MIBC, Stage T2 or more) requires neoadjuvant chemotherapy before radi-
cal cystectomy. Hence, preoperative differentiation is vital.
Purpose: To investigate whether intravoxel incoherent motion (IVIM) diffusion-weighted imaging (DWI) can differentiate
NMIBC from MIBC and to assess whether there were correlations between IVIM parameters and the Ki-67 labeling index (LI).
Study Type: Retrospective.
Subjects: Thirty-six patients diagnosed with bladder cancer confirmed by histopathological findings.
Field Strength/Sequence: 3.0T magnetic resonance imaging (MRI) DWI with eight b-values ranging from 0 to 1000 s/mm2.
Assessment: Molecular diffusion coefficient (D), perfusion-related diffusion coefficient (D*), perfusion fraction (f), and
apparent diffusion coefficient (ADC) were calculated by biexponential and monoexponential models fits, respectively.
Statistical Tests: Comparisons were made between the MIBC and NMIBC group, and differences were analyzed by
comparing the areas under the receiver-operating characteristic curves (AUCs). The correlations between these parame-
ters and Ki-67 LI were assessed by Spearman’s rank correlation analysis.
Results: The ADC and D value were significantly lower in patients with MIBC compared to those with NMIBC (P <
0.01). No significant (P > 0.05) differences were observed in D* and f. The AUC of D value (0.894) was significantly (P <
0.05) larger than the ADC value (0.786), with sensitivities and specificities of 95% and 87.5% (D) and 80% and 68.7%
(ADC), respectively. In addition, the D and ADC values were significantly correlated with Ki-67 LI (r 5 –0.785, r 5 –0.643,
respectively; both P < 0.01).
Data Conclusion: The D value obtained from IVIM exhibited better performance than conventional DWI for distinguish-
ing NMIBC from MIBC and may serve as a potential imaging biomarker for bladder cancer invasion.
Level of Evidence: 1
Technical Efficacy: Stage 3
J. MAGN. RESON. IMAGING 2017;00:000–000.

C linical management and prognosis of bladder cancer

mainly rely on the degree of bladder wall invasion.1
Nonmuscle-invasive bladder cancer (NMIBC, Stage T1 or
survival outcome of MIBC is less favorable, and patients
tend to have distant metastasis.3 Therefore, the clinically
appropriate treatment of bladder cancer depends on accurate
lower) accounts for 75% of bladder cancer, has a good prog- preoperative staging of the tumor. Therefore, using preoper-
nosis, and is treated with transurethral resection (TUR).1 ative imaging to differentiate NMIBC from MIBC plays an
Muscle-invasive bladder cancer (MIBC, Stage T2 or more), important diagnostic role.
comprising 25% of bladder cancer, requires neoadjuvant Diffusion-weighted imaging (DWI) may be a useful
chemotherapy before radical cystectomy.2 In addition, the method for determining bladder cancer staging. The

View this article online at wileyonlinelibrary.com. DOI: 10.1002/jmri.25839

Received Jun 8, 2017, Accepted for publication Jul 29, 2017.

*Address reprint requests to: J.X., Department of Radiology, Ren Ji Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai, P.R. China.
E-mail: [email protected]

The first two authors contributed equally to this work.

The last two authors contributed equally to this work.

From the Department of Radiology, Ren Ji Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai, P.R. China

C 2017 International Society for Magnetic Resonance in Medicine

V 1
Journal of Magnetic Resonance Imaging

TABLE 1. MRI Protocol for Tumor Assessment

MRI protocol
MRI sequences
T1-weighted SE T2-weighted SE DWI
Plane Axial Axial, sagittal Axial
Fat saturation No Yes No
Time to repeat (msec) 534 4000 6000
Time to echo 8 110 70
Thickness (mm) 3 3 3
FOV (cm) 30 22 20
Matrix (mm3mm) 300 3 226 276 3 239 80 3 142
Intersection gap (mm) 1 1 0.3
Number of excitations 2 2 2
Other 8 (b values 0, 50, 100, 150,
200, 500, 800, 1000 s/mm2)
MRI, magnetic resonance imaging; FOV, field of view; SE, spin echo.

contours of tumors in DWI can indicate stage T1 or lower and brain tumors.9–13 IVIM-DWI with multiple b-values
as a thin, flat, C-shape; stage T2 as a smooth tumor margin; can provide three parameters through biexponential analysis:
stage T3 as an irregular margin extension into the perivesical the pure diffusion coefficient (D), perfusion-related incoher-
fat; and stage T4 as extension into adjacent organs.4 As ent microcirculation (D*), and the microvascular volume
stage T1 vs. T2 clinically differentiates NMIBC and MIBC, fraction (f ). Compared to conditional DWI, IVIM-DWI
the differentiation of stage T1 and T2 by DWI could play a can reflect both perfusion and true diffusion-related effect in
significant role in clinical treatment procedures. The unique tumors, thus more accurately characterizing the actual status
appearance of “the inch worm sign” always indicates T1 or of diffusion in tumors without a contrast agent.
lower, and the stalk of the tumor consists of a submucosa Currently, IVIM magnetic resonance imaging (MRI) is
with edema and fibrosis.4 In addition, smooth tumor mar- rarely used as a tool to stage bladder cancer. Therefore, this
gins without extension into the perivesical space indicates study was designed to investigate whether the parameters of
T2.4 However, the muscle layer is usually thin on DW IVIM can differentiate NMIBC from MIBC.
images, and the contrast difference between the tumor mar-
gin and the bladder muscle layer is not great enough to pre- Materials and Methods
cisely distinguish invasive (T2) from superficial (T1) Study Population
tumors.4 Dynamic contrast-enhanced imaging (DCEI) can The protocols of this study were approved by the Institutional
demonstrate that intact submucosal linear enhancement Review Board, and informed consent was obtained from all
(SLE) adjacent to a tumor indicates stage T1 or a lower patients. Between April 2016 and April 2017, 52 patients sus-
stage. However, approximately 60% of dynamic contrast- pected of bladder cancer underwent MRI with a protocol including
enhanced studies showed that the submucosa and tumor T1WI (weighted imaging), T2WI, and DWI using eight b-values
have similar signal intensity, making it difficult to recognize (0, 50, 100, 150, 200, 500, 800, and 1000s/mm2) before transure-
submucosal linear enhancement.5 In addition, perivesical thral resection (TUR). Patients who underwent TUR after imaging
enhancement can be a reactive or acute inflammatory of pathologically confirmed bladder cancer were enrolled in this
change rather than tumor invasion.6 These phenomena may study. Tumor specimens were sufficient for immunohistochemical
cause false positivity and overestimation of the tumor stage. staining. Patients (n 5 4) with no histological evidence of bladder
cancer and patients (n 5 12) with pathologically confirmed benign
The overall staging accuracy using DCEI has been reported
lesions were excluded.
to be 52–85%.5,7,8 Therefore, further improvements to
diagnostic methods may be desirable for preoperative staging MRI Protocols
of tumors. All patients underwent pelvic MRI with DWI examination on a
Intravoxel incoherent motion (IVIM) DWI is an 3.0T MR scanner (Ingenia; Philips Healthcare, Best, the Nether-
advanced imaging technique that has been explored in vari- lands) with a Torso coil (see protocol in Table 1). Patients were
ous tumor types, including breast, liver, prostate, pancreas, asked to fast for 4–6 hours before examination and were instructed

2 Volume 00, No. 00

 Wang et al.: Assessing Bladder Cancer Invasiveness

TABLE 2. Comparisons Between the NMIBC Group and the MIBC Group

Class N ADC (31023 mm2/s) D (31023 mm2/s) D* (31023 mm2/s) f(%)

NMIBC 16 1.587 6 0.442 1.557 6 0.385 69.175 6 58.596 16.1 6 11.7
MIBC 20 1.209 6 0.222 0.956 6 0.237 12.124[47.986] 17.5 6 10.9
(t/Z) value 3.117 -4.011 -1.580 -0.373
P value 0.005 0.000 0.114 0.711
D* values (median [quartiles]). NMIBC, nonmuscle-invasive bladder cancer.

not to urinate for at least 1 hour before the examination in order not conform to a normal distribution). Receiver operator character-
to moderately distend the bladder. istic (ROC) curves were drawn to compare the areas of these
parameters. The correlation between these parameters and the Ki-
Image Analysis 67 LI was assessed by Spearman’s rank correlation analysis. The
All digital images were transferred to the in-house software devel- intraclass correlation coefficient (ICC) was used to evaluate the
oped in IDL 6.3 (ITT Visual Information Solutions, Boulder, interobserver agreement of all metrics.14 An ICC greater than 0.75
CO) for further analysis. All regions of interest (ROIs) were drawn represented good agreement.15 P < 0.05 was considered statisti-
by two radiologists (X.-X.C. and L.-M.W with 2 and 5 years of cally significant.
experience in pelvic MRI, respectively) who were blind to the path-
ologic outcomes. The interreader agreement in the calculation of
these metrics was also evaluated. For each case, the ROI for IVIM
fitting was placed over the entire largest lesion while avoiding Results
necrosis. The ROI was transferred to the biexponential model to Thirty-six patients (30 male, 6 female; mean age 65 6 12
obtain the values of D, f, and D*. At the same time, the ROI was years) underwent MRI followed by TUR biopsy. Sixteen
transferred to the monoexponential model to obtain ADC values. patients diagnosed with NMIBC underwent TUR, and 20
patients diagnosed with MIBC received two cycles of
Immunohistochemistry
Thirty-six bladder cancer specimens obtained from TUR were
cisplatin-based chemotherapy before radical cystectomy. The
immunohistochemically stained for evaluation of the Ki-67 labeling interobserver agreement was good for the calculation of all
index (LI). By counting the cells in each area of the tumor, the these metrics (ICC range, 0.874–0.989).
percentage of positive nuclear Ki-67 expression was evaluated. The ADC and D value were significantly lower in
patients with MIBC ([1.209 6 0.222] 3 1023 mm2/s and
Statistical Evaluation [0.956 6 0.237] 3 1023 mm2/s, respectively) compared to
Statistical analyses were carried out using SPSS 20.0 (Chicago, IL).
those suffering NMIBC ([1.587 6 0.442] 3 1023 mm2/s
Mean 6 standard deviation, median and interquartile range (IQR)
and [1.557 6 0.385] 3 1023 mm2/s, respectively) (P <
were used to express continuous variables and discrete variables,
respectively. Normality of the distribution was tested using the Kol-
0.01). No significant differences were observed in D* and f
mogorov–Smirnov test. Independent-samples t-tests were used to parameters (both P > 0.05) (Table 2; Fig. 1).
compare differences between the MIBC group and the NMIBC The area under the ROC curves of the D value
group in the case of a normal distribution. Nonparametric Mann– (0.894) was significantly (P < 0.05) larger than that of the
Whitney U-tests were used when appropriate (when parameters did ADC value (0.786). The sensitivities were 95% and 80%

FIGURE 1: Box-and-whisker plots of ADC values (a) and D values (b) between the NMIBC group and the MIBC group.

Month 2017 3
Journal of Magnetic Resonance Imaging

TABLE 3. Results of the Receiver Operating Characteristic Analyses for Discriminating NMIBC From MIBC

Parameters Cutoff value (31023 mm2/s) AUC Sensitivity % Specificity %

ADC 1.343 0.786 80 68.7
D 1.192 0.894 95 87.5

and the specificities were 87.5% and 68.7% for the D and that is much more compact, and thus, the ADC and D val-
ADC value, respectively (Table 3; Fig. 2). ues in MIBC patients were significantly lower than in
The D value and Ki-67 LI were negatively correlated NMIBC patients. The D* value is affected by average blood
(r 5 –0.785, P < 0.01), and the ADC value and Ki-67 LI velocity and capillary segment length.22 The signal intensity
were negatively correlated (r 5 –0.643, P < 0.01) (Table 4; ratio of blood capillaries and tumor tissues determines the f
Figs. 3–5). There were no significant correlations among value.23 However, there were no significant differences in
D*, f value, and Ki-67 LI (both P > 0.05). D* nor the f values between the MIBC and NMIBC
groups, which indicated that extracellular components such
Discussion as blood flow velocity, vascular perfusion, or capillary vol-
Previous studies7,16–21 have shown that ADCs can be used ume contribute little to the aggressiveness of bladder
to stage bladder cancer. ADCs are lower in higher T-stage tumors. In the current study, the D* values in the MIBC
bladder cancer. Kobayashi19 evaluated three T-stage groups group were not significantly lower than the D* values in the
(Ta, T1, and T2 or higher) and found that ADC values NMIBC group. Similarly, previous studies of liver or breast
were significantly lower for higher T stages than for lower T have reported that D* values in malignant lesions were
stages (T2 vs. T1 vs. Ta). However, the ADC value is lower than or indistinguishable from those of benign
affected not only by tissue cellularity density but also by lesions.24–28 The repeatability estimates found for D* in this
blood flow perfusion and membrane permeability. Since the work were relatively good, with an ICC greater than 0.75
D value could represent the true molecular diffusion and for D* using the biexponential model, while D* had previ-
distract the affection of microvascular perfusion, it may ously been found to be less stable, with wide variability in
more accurately reflect the inner construction and invasive tumors.24 IVIM parameters are dependent on magnetic field
information of tumors. strength to a small extent and the relatively low signal-to-
In this study, we found that the ADC and D values noise ratio (SNR) may compromise the precision of the cal-
were significantly lower in patients with MIBC compared to culated IVIM parameters.29 In the current study, we per-
those with NMIBC. These results indicate that higher T- formed this examination at 3.0T MRI, and the SNR was
stages corresponded to greater restriction of water molecular relatively high. In addition, the measurement bias between
two observers may be reduced compared to other organs
diffusion. Tumor cell-active growth results in cell density
because DWI provides high contrast between bladder cancer
and background tissue. Moreover, since D* is considered to
represent blood flow velocity, the wide variability of D* val-
ues may result from varying tumor vascularity among differ-
ent tumor types.28 In the present study, the f value did not
differ significantly between the MIBC and NMIBC groups,
which was similar to other findings.25,30 f was sensitive to
the maximal b value employed and their potential sampling
from non-Gaussian effects.24 A previous study found that
the f value was significantly increased in tumors with maxi-
mal b value below 750 s/mm2 in prostate cancers and

TABLE 4. Correlations Between MR Parameters and

Ki-67 LI

Parameters Ki-67(r) P value

ADC 20.643 0.000
FIGURE 2: Receiver operating characteristic analyses for dis- D 20.785 0.000
criminating NMIBC from MIBC.

4 Volume 00, No. 00

 Wang et al.: Assessing Bladder Cancer Invasiveness

FIGURE 3: Scatterplots showing significant negative correlations between the ADC value (a), the D value (b) and Ki-67 LI.

decreased or indistinguishable from normal tissues with The ADC value obtained from the monoexponential model
higher maximal b value.30 In addition, f could be influenced assumed a Gaussian distribution was influenced not only by
by a T2 contribution.22,31 tissue cellularity but also by microcirculation of the blood
In this study, we found that the D value was superior in the capillary network, so as not to reflect the true water
to the ADC value for discriminating NMIBC from MIBC.

FIGURE 5: Images acquired from a 38-year-old male diagnosed

FIGURE 4: Images acquired from a 63-year-old female diag- with MIBC confirmed by pathological findings. (a) Axial T2-
nosed with NMIBC confirmed by pathological findings. (a) Axial weighted image shows an irregular margin on the right poste-
T2-weighted image shows an irregular margin on the left poste- rior bladder wall (arrow). (b–d) D map, D* map, and f map of
rior bladder wall (arrow). (b–d) D map, D* map, and f map of (a). The bladder wall is disrupted by the tumor on the D map.
(a). The bladder wall is intact on the D map. The D value of the The D value of the tumor is 1.063 3 1023 mm2/s. (e) Ki-67
tumor is 1.476 3 1023 mm2/s. (e) Ki-67 immunohistochemical immunohistochemical staining shows that the percentage of
staining shows that the percentage of positive cells is 2%. positive cells is 50%.

Month 2017 5
Journal of Magnetic Resonance Imaging

molecular diffusion of the tumor. The D value separating References

diffusion and perfusion represents the true molecular diffu- 1. Griffiths TR. Current perspectives in bladder cancer management. Int
sion in the biexponential model. The diffusion pattern of J Clin Pract 2013;67:435–448.

water molecules of high-stage tumors is more heterogeneous 2. Apolo AB, Vogelzang NJ, Theodorescu D. New and promising strate-
gies in the management of bladder cancer. Am Soc Clin Oncol Educ
and non-Gaussian than low-stage tumors. Thus, compared Book 2015;35:105–112.
to the ADC, the D value may more accurately reflect the 3. Stein JP, Skinner DG. Radical cystectomy for invasive bladder cancer:
cell diffusivity and microvascular perfusion of tumors. long-term results of a standard procedure. World J Urol 2006;24:296–
The Ki-67 labeling index (LI) is associated with recur- 304.

rence, progression, aggressiveness, and survival of patients 4. Takeuchi M, Sasaki S, Ito M, et al. Urinary bladder cancer: diffusion-
weighted MR imaging—accuracy for diagnosing T stage and estimat-
with bladder cancer.32–34 Ki-67 LI is an established marker ing histologic grade. Radiology 2009;251:112–121.
of cell proliferation and tumor invasion, which can be a 5. Tekes A, Kamel I, Imam K, et al. Dynamic MRI of bladder cancer: eval-
prognostic biomarker for bladder cancer that predicts clini- uation of staging accuracy. AJR Am J Roentgenol 2005;184:121–127.
cal aggressiveness.32–35 In this study, we found that the D 6. Rabie E, Faeghi F, lzadpanahi MH, Dayani MA. Role of dynamic
and ADC values were negatively correlated with Ki-67 LI contrast-enhanced magnetic resonance imaging in staging of bladder
cancer. J Clin Diagn Res 2016;10:TC01–5.
and that there were no significant correlations between the
7. Kim B, Semelka RC, Ascher SM, Chalpin DB, Carroll PR, Hricak H.
D* and f values and Ki-67 LI. This result confirmed that Bladder tumor staging: comparison of contrast-enhanced CT, T1- and
tumor invasion can be reflected by the extent of water T2-weighted MR imaging, dynamic gadolinium-enhanced imaging,
molecular diffusion restriction rather than microvascular and late gadolinium-enhanced imaging. Radiology 1994;193:239–245.

perfusion. Additionally, in this study there were negative 8. Tanimoto A, Yuasa Y, Imai Y, et al. Bladder tumor staging: compari-
son of conventional and gadolinium-enhanced dynamic MR imaging
correlations between the ADC value and Ki-67 LI. This and CT. Radiology 1992;185:741–747.
finding is similar to other studies.19,34,35 However, there 9. Pesapane F, Patella F, Fumarola EM, et al. Intravoxel incoherent
were no significant correlations between D*, f value, and motion (IVIM) diffusion weighted imaging (DWI) in the periferic pros-
Ki-67 LI. As mentioned above, extracellular components tate cancer detection and stratification. Med Oncol 2017;34:35.

such as blood flow velocity, vascular perfusion, or capillary 10. Klaub M, Maier-Hein K, Tjaden C, Hackert T, Grenacher L, Stieltjes
B. IVIM DW-MRI of autoimmune pancreatitis: therapy monitoring
volume contributed little to the aggressiveness of bladder and differentiation from pancreatic cancer. Eur Radiol 2016;26:2099–
tumors. Therefore, the extent of water molecular diffusion 2106.
restriction, rather than microvascular perfusion, accounted 11. Klauss M, Mayer P, Maier-Hein K, et al. IVIM-diffusion-MRI for the dif-
ferentiation of solid benign and malign hypervascular liver lesions—
for the invasive bladder cancer. As is known, similar to
Evaluation with two different MR scanners. Eur J Radiol 2016;85:
tumors, a reactive or acute inflammatory change but a 1289–1294.
benign reaction of hyperemia and edema could form new 12. Federau C, O’Brien K, Meuli R, Hagmann P, Maeder P. Measuring
capillaries, which may also lead to changes in microvascular brain perfusion with intravoxel incoherent motion (IVIM): initial clinical
experience. J Magn Reson Imaging 2014;39:624–632.
perfusion.36
13. Lee YJ, Kim SH, Kang BJ, et al. Intravoxel incoherent motion (IVIM)-
Additionally, there were limitations in this study. First,
derived parameters in diffusion-weighted MRI: associations with prog-
the number of patients enrolled in this study was rather nostic factors in invasive ductal carcinoma. J Magn Reson Imaging
small. Second, we could not subdivide patients into differ- 2017;45:1394–1406.

ent stages because muscle-invasive bladder cancer requires 14. Shrout PE, Fleiss JL. Intraclass correlations: uses in assessing rater reli-
ability. Psychol Bull 1979;86:420–428.
chemotherapy before radical cystectomy and, thus, hinders
15. Busing KA, Kilian AK, Schaible T, Debus A, Weiss C, Neff KW. Reli-
comparing T stages accurately between cystectomy speci-
ability and validity of MR image lung volume measurement in fetuses
mens and MRI. Third, the IVIM parameters depend on with congenital diaphragmatic hernia and in vitro lung models. Radi-
this imaging protocol, and thus, calculating the values in ology 2008;246:553–561.

the current study is not applicable to other studies that uti- 16. Padhani AR, Liu G, Koh DM, et al. Diffusion-weighted magnetic reso-
nance imaging as a cancer biomarker: consensus and recommenda-
lize different imaging protocols. Standardization of the tions. Neoplasia 2009;11:102–125.
imaging protocol is desired. Finally, it requires more time to
17. Kobayashi S, Koga F, Yoshida S, et al. Diagnostic performance of
perform IVIM-DWI with eight values than routine DWI, diffusion-weighted magnetic resonance imaging in bladder cancer:
and patients with a full bladder may suffer more pain dur- potential utility of apparent diffusion coefficient values as a biomarker
to predict clinical aggressiveness. Eur Radiol 2011;21:2178–2186.
ing this examination.
18. Rosenkrantz AB, Mussi TC, Spieler B, Melamed J, Taneja SS, Huang
In conclusion, adding IVIM to MRI can increase the WC. High-grade bladder cancer: association of the apparent diffusion
diagnostic performance of assessing bladder cancer invasive- coefficient with metastatic disease: preliminary results. J Magn Reson
Imaging 2012;35:1478–1483.
ness compared to conventional ADCs. D values obtained
from IVIM are helpful in distinguishing NMIBC from 19. Kobayashi S, Koga F, Kajino K, et al. Apparent diffusion coefficient
value reflects invasive and proliferative potential of bladder cancer.
MIBC. In addition, the D value could serve as an imaging J Magn Reson Imaging 2014;39:172–178.
biomarker that reflects the invasive and proliferative poten- 20. Daggulli M, Onur MR, Firdolas F, Onur R, Kocakoc E, Orhan i. Role of
tial of bladder cancer. diffusion MRI and apparent diffusion coefficient measurement in the

6 Volume 00, No. 00

 Wang et al.: Assessing Bladder Cancer Invasiveness

diagnosis, staging and pathological classification of bladder tumors. 29. Li YT, Cercueil JP, Yuan J, et al. Liver intravoxel incoherent motion
Urol Int 2011;87:346–352. (IVIM) magnetic resonance imaging: a comprehensive review of pub-
lished data on normal values and applications for fibrosis and tumor
21. El-Assmy A, Abou-El-Ghar ME, Mosbah A, et al. Bladder tumour stag-
evaluation. Quant Imaging Med Surg 2017;7:59–78.
ing: comparison of diffusion- and T2-weighted MR imaging. Eur
Radiol 2009;19:1575–1581. 30. Pang Y, Turkbey B, Bernardo M, et al. Intravoxel incoherent motion
MR imaging for prostate cancer: an evaluation of perfusion fraction
22. Le Bihan D, Breton E, Lallemand D, Aubin ML, Vignaud J, Laval- and diffusion coefficient derived from different b-value combinations.
Jeantet M. Separation of diffusion and perfusion in intravoxel incoher- Magn Reson Med 2013;69:553–562.
ent motion MR imaging. Radiology 1988;168:497–505.
31. Jerome NP, d’Arcy JA, Feiweier T, et al. Extended T2-IVIM model for
23. Sumi M, Van Cauteren M, Sumi T, Obara M, Ichikawa Y, Nakamura T. correction of TE dependence of pseudo-diffusion volume fraction in
Salivary gland tumors: use of intravoxel incoherent motion MR imag- clinical diffusion-weighted magnetic resonance imaging. Phys Med
ing for assessment of diffusion and perfusion for the differentiation of Biol 2016;61:N667–N680.
benign from malignant tumors. Radiology 2012;263:770–777.
32. Margulis V, Shariat SF, Ashfaq R, Sagalowsky AI, Lotan Y. Ki-67 is an
24. Liu C, Wang K, Chan Q, et al. Intravoxel incoherent motion MR imag- independent predictor of bladder cancer outcome in patients treated
ing for breast lesions: comparison and correlation with pharmacoki- with radical cystectomy for organ-confined disease. Clin Cancer Res
netic evaluation from dynamic contrast-enhanced MR imaging. Eur 2006;12:7369–7373.
Radiol 2016;26:3888–3898.
33. Quintero A, Alvarez-Kindelan J, Luque RJ, et al. Ki-67 MIB1 labelling
25. Ichikawa S, Motosugi U, Ichikawa T, Sano K, Morisaka H, Araki T. index and the prognosis of primary TaT1 urothelial cell carcinoma of
Intravoxel incoherent motion imaging of focal hepatic lesions. J Magn the bladder. J Clin Pathol 2006;59:83–88.
Reson Imaging 2013;37:1371–1376.
34. Margulis V, Lotan Y, Karakiewicz PI, et al. Multi-institutional validation
26. Bokacheva L, Kaplan JB, Giri DD, et al. Intravoxel incoherent motion of the predictive value of Ki-67 labeling index in patients with urinary
diffusion-weighted MRI at 3.0T differentiates malignant breast lesions bladder cancer. J Natl Cancer Inst 2009;101:114–119.
from benign lesions and breast parenchyma. J Magn Reson Imaging
2014;40:813–823. 35. Sevcenco S, Haitel A, Ponhold L, et al. Quantitative apparent diffusion
coefficient measurements obtained by 3-Tesla MRI are correlated with
27. Sigmund EE, Cho GY, Kim S, et al. Intravoxel incoherent motion biomarkers of bladder cancer proliferative activity. PLoS One 2014;9:
imaging of tumor microenvironment in locally advanced breast can- e106866.
cer. Magn Reson Med 2011;65:1437–1447.
36. Rosenkrantz AB, Ego-Osuala IO, Khalef V, Deng FM, Taneja SS,
28. Liu C, Liang C, Liu Z, Zhang S, Huang B. Intravoxel incoherent motion Huang WC. Investigation of multisequence magnetic resonance imag-
(IVIM) in evaluation of breast lesions: comparison with conventional ing for detection of recurrent tumor after transurethral resection for
DWI. Eur J Radiol 2013;82:e782–e789. bladder cancer. J Comput Assist Tomogr 2016;40:201–205.

Month 2017 7