The influence of diffusion gradient direction on diffusion-weighted imaging of breast mass-like lesions at 3.0T

Abstract

Background

It has been challenging to achieve ideal breast diffusion-weighted imaging (DWI). The optimization of diffusion gradient direction is of great importance.

Purpose

To evaluate the effect of diffusion gradient direction on the apparent diffusion coefficient (ADC) values of breast mass-like lesions and the visual grades of image quality, lesion visibility, and sharpness of breast contour at 3.0T.

Material and Methods

Sixty consecutive patients with mass-like lesions were enrolled in this study. In addition to typical breast magnetic resonance imaging (MRI) protocols, the breasts were scanned with conventional orthogonal DWI (c-DWI), tetrahedral DWI (t-DWI), and 3in1 DWI (3in1-DWI) sequences. The DW images were observed and visually graded by two radiologists independently. For ADC measurement, one radiographer manually selected the region of interest (ROI).

Results

For both readers, t-DWI had better image quality and sharpness of breast contour than c-DWI. Regarding lesion visibility, no significant differences were observed among three sequences. The mean ADC values were 1.462 × 10⁻³, 1.490 × 10⁻³, and 1.446 × 10⁻³ mm² s⁻¹ for c-DWI, t-DWI, and 3in1-DWI, respectively. The ADC values extracted from both t-DWI and 3in1-DWI were not statistically different compared with those from c-DWI. In all DWI sequences, the ADC of malignant lesions was significantly reduced compared with benign lesions.

Conclusion

DWI with tetrahedral or 3in1 diffusion gradients is a more useful technique in clinical breast MRI than c-DWI because the image quality and sharpness of breast contour are improved. ADC is comparable to c-DWI.

Keywords

Diffusion gradient tetrahedral breast diffusion-weighted imaging (DWI)apparent diffusion coefficient (ADC)

Introduction

Diffusion-weighted imaging (DWI) is a special magnetic resonance imaging (MRI) technique that can be used to evaluate the physical process of the microscopic thermal motion of water molecules in vivo, which is quantified by the apparent diffusion coefficient (ADC) (1). DWI has some advantages, such as speed and the lack of contrast medium injection and ionizing radiation (2). In addition, it can be readily implemented on any MRI platform with little or no software modification (3).

In 1997, Englander et al. used DWI and ADC values to distinguish fibroglandular and fat tissue in normal breasts (4) and they initiated the clinical application of the technique in human breasts. There are some challenges to achieve ideal DWI of the breast. First, the location of both breasts is off the center of the magnetic field, which can impair magnetic field homogeneity. Second, breasts are composed of fatty tissue, which can easily lead to chemical shift artifacts. Finally, physiological movement can cause motion artifacts.

Based on the above situation, many researchers have attempted to optimize breast DWI via the selection of appropriate pulse sequences (5 –9), b-value (10,11), fat suppression techniques (12,13), and signal attenuation models (14,15). However, not many studies have focused on the effect of diffusion gradient direction on breast DWI. Sinha et al. (16) explored the potential of tetrahedral DWI for lesion characterization with a smaller number of subjects. Takashi et al. (17) compared tetrahedral and orthogonal DWI for the detection and localization of breast lesions in 67 patients. Both of these studies were conducted at 1.5 T.

In the present study, tetrahedral DWI (t-DWI) and 3in1 DWI (3in1-DWI) were compared to conventional orthogonal DWI (c-DWI) at 3.0T in the assessment of the image quality, lesion visibility, sharpness of breast contour and ADC value of lesions.

Material and Methods

Participants

Seventy-one patients who had palpable breast masses and underwent breast MRI from December 2014 to August 2015 at the MR department of the first affiliated hospital of Zhengzhou University were included in this prospective study. Before examination, the radiographer explained the examination steps and some attentions, for example, keeping calm breath and trying to avoid movements. The nurse explained the related adverse reactions. Then, the patients provided contrast-enhanced MRI agreement.

The exclusion criteria were as follows: aged less than 18 years (n = 1); lesion size <5 mm on contrast-enhanced MRI (n = 2); renal insufficiency or contrast agent allergy (n = 1); unable to tolerate the prone position during examination (n = 2); claustrophobic (n = 1); no surgery (lesions were not verified by histopathologic examination) (n = 1). Moreover, to facilitate the comparison of ADC values, non-mass-like tumors and simple cysts (n = 3) were excluded from the cohort.

Subsequently, 60 consecutive patients (mean age, 52 years; age range, 21–74 years) were eligible for this study.

Diffusion gradient directions

In the DWI sequence, diffusion gradients are bipolar gradient pulses that are added after excitation and before readout. The role of the pulses is compensation in the case of stationary spins and to simultaneously measure the signal attenuation of moving spins in the given direction.

In c-DWI, diffusion gradients are sequentially applied along three orthogonal directions: (1,0,0)^T, (0,1,0)^T, and (0,0,1)^T. Instead, the diffusion gradients of t-DWI are performed along four different vectors: (1,1,1)^T, (1,–1,–1)^T, (–1,–1,1)^T, and (–1,1,–1)^T. Obviously, the strength of tetrahedral gradients is $\sqrt{3}$ times the orthogonal input components. In 3in1-DWI, diffusion gradients are simultaneously applied along three orthogonal directions: (1,0,0)^T, (0,1,0)^T, and (0,0,1)^T (18). Compared with t-DWI, 3in1-DWI has only one diffusion direction, the (1,1,1)^T vector.

MRI acquisition

Examinations were performed using a dedicated bilateral 8-channel array breast coil (GE Healthcare, Milwaukee, WI, USA) on clinical whole-body scanners at 3.0T (Discovery MR750, GE Healthcare) with a 50 mT/m gradient and 200 mT/m/s slew rate.

MRI protocols were as follows: axial T2-weighted (T2W) fast spin echo-IDEAL sequence (TR/TE, 4314/50 ms; FOV, 32 × 32 cm; matrix, 320 × 192; slice thickness/spacing, 4/1 mm), 3D VIBRANT (Volume Imaging for BReast AssessmeNT) sequence (TR/TE, 3.9 ms/minimum; flip angle, 5°; FOV, 36 × 36 cm; matrix, 320 × 320; slice thickness, 1.4 mm). The VIBRANT sequence included six phases. The first phase was performed before contrast media injection. The other phases were performed approximately 30 s after injection.

DWI sequences were performed before VIBRANT sequence. Local shimming was identically performed on three DWI sequences. Outer volume suppression was used to remove the signal from the chest. Image positioning was copied each other among T2W sequence and different DWI sequences. Table 1 presents the three DWI sequences in sequential order and their main parameters. In three DWI sequences, different signal averages were used to obtain comparative acquisition time and to keep the total number of diffusion directions equal.

Table 1.

MRI sequence parameters of three different DWI protocols.

	DWI protocol 1	DWI protocol 2	DWI protocol 3
Diffusion direction	ALL (orthogonal)	TETRA (tetrahedral)	3 in 1
TR/TE (ms)	3600/77.3 (minimum)	4910/73.5 (minimum)	4910/73.5 (minimum)
Bandwidth (Hz)	250	250	250
b value (s mm^–2)	0, 800	0, 800	0, 800
Signal averages	4	3	12
Acceleration factors	2	2	2
Fat suppression	Spatial-spectral water-only excitation	Spatial-spectral water-only excitation	Spatial-spectral water-only excitation
Phase-encoding direction	Anterior/Posterior	Anterior/Posterior	Anterior/Posterior
Shims	Dual volumes	Dual volumes	Dual volumes
Field of view (mm²)	320 × 320	320 × 320	320 × 320
Acquisition matrix	128 × 128	128 × 128	128 × 128
Slice thickness/Spacing (mm)	4/1	4/1	4/1
Slices (n)	32	32	32
Scan time (min:s)	0:54	1:14	1:14

All images were transferred to the picture archiving and communication systems (PACS). The DWI sequences were post-processed using the commercial FunctionTool software to obtain ADC maps on the Advantage Windows workstation (Version 4.5, GE Healthcare, Buckinghamshire, UK).

Image analysis

Two radiologists (readers A and B) visually assessed image quality, lesion visibility, and the sharpness of breast contour on three DWI sequences through PACS. One radiographer manually selected a region of interest (ROI) on the lesion for ADC measurement. For image analysis, all readers were able to view DW images (b = 0 and 800 s mm⁻²), ADC maps, T2W images, contrast-enhanced MR images, and subtraction images. All observers were allowed to adjust the window settings and blinded to the sequence type (c-DWI, t-DWI, and 3in1-DWI). The cases were arranged in a random order.

Assessment of image quality: Two radiologists subjectively graded the images of three DWI sequences based on the presence of image artifacts (e.g. ghosting artifacts), the visualization of breast parenchyma, and image blurring. Overall image quality was assessed using a four-point scale (1, poor; 2, fair; 3, good; 4, excellent).

Lesion visibility: Using contrast-enhanced images as references, the visibility of lesions (5) was rated according to a scale ranging from 0 to 3 (0, not visible; 1, poor visibility, subtraction images needed for lesion demarcation; 2, good visibility, but inferior to subtraction images; 3, excellent visibility, equivalent to subtraction images).

The sharpness of breast contour: To measure the distance between breast skin line and the lesion, it is necessary to obtain a sharp breast contour. Therefore, two radiologists rated the sharpness of breast contour as: 1, very blurry; 2, slightly blurry; 3, average; and 4, sharp.

ADC evaluation: For a reliable radiologic–pathologic correlation, relevant clinical information and surgical histopathology results were collected prospectively. For multiple masses in the breast, we chose the largest pathologically proven mass as target lesion. The target lesion was identified on dynamic contrast-enhanced (DCE)-MR images. Then, ROIs were drawn freehand at the corresponding location on the b = 800 s mm⁻² DW images.

While defining ROIs, the radiographer should pay attention to the following points: (i) The scope of ROI should contain at least 4 pixels, approximately 25 mm²; (ii) the definition of ROI should be as much of the lesion as possible while staying within the border of the hyperintense region; (iii) the images of other sequences (T2W, 3D Vibrant) serve as a guide to avoid areas of hemorrhage, necrosis, and cysts. The measurements were repeated three times and the average was recorded. A third radiologist (with extensive experience in breast MRI) reviewed all ROI localizations and confirmed that the assessment of ADC by the radiographer was appropriate.

Statistical analysis

All statistical analyses were performed using statistical software (SPSS Inc., version 21.0; Chicago, IL, USA). For all tests, probability values less than 0.05 were considered statistically significant. Quantitative data are presented as means ± standard deviations (SD).

Kruskal–Wallis one-way analysis of variance (ANOVA) test was used to assess the visual grades of image quality, lesion visibility, and the sharpness of breast contour among three DWI sequences. For those independent variables for which the Kruskal–Wallis test is significant, pairwise multiple comparisons are automatically produced using the Dunn–Bonferroni approach. Kappa statistic was used to evaluate agreement between radiologists for visual grades.

ANOVA tests were performed to compare the ADC values among different DWI sequences (t-DWI, 3in1-DWI, and c-DWI). Mauchly’s test of sphericity was applied, and the degrees of freedom were corrected using the Huynh–Feldt Epsilon as appropriate. Bonferroni-corrected multiple comparison tests were used to identify significant differences for variables.

Results

Lesion characteristics

Histopathological confirmation was obtained for 60 lesions via excisional surgery. The lesions comprised invasive ductal carcinoma (n = 26), fibroadenomas (n = 32), benign phylloides (n = 1), and duct papilloma (n = 1).

Comparisons of the visual grades among three DWI sequences

Table 2 presents the results of visual assessment with respect to image quality, lesion visibility, and the sharpness of the breast contour with different DWI sequences by two radiologists. For both readers, t-DWI had better image quality and sharpness of breast contour than c-DWI. Regarding the lesion visibility of different DWI sequences, no significant differences were observed. However, minimal agreement between readers was noted for the assessment of specific visual grades (Table 3). Fig. 1 presents some representative cases.

Table 2.

Mean (SD) of visual grades (n = 60).

	Image quality			Lesion visibility			Sharpness of breast contour
	c-DWI	t-DWI	3in1-DWI	c-DWI	t-DWI	3in1-DWI	c-DWI	t-DWI	3in1-DWI
Reader A	2.1 (0.4)*	2.9 (0.6)	2.9 (0.6)	2.4 (0.5)	2.5 (0.5)	2.4 (0.5)	2.3 (0.6)*	3.1 (0.7)	3.1 (0.6)
Reader B	2.5 (0.5)†	2.9 (0.3)	2.7 (0.5)	2.7 (0.5)	2.7 (0.5)	2.7 (0.5)	3.1 (0.3)*	3.6 (0.6)	3.7 (0.5)

Significantly lower than t-DWI and 3in1-DWI (P < 0.001).

†

Significantly lower than t-DWI (P < 0.001).

Table 3.

Agreement between two readers on specific visual grades using Kappa statistic (n = 60).

	c-DWI			t-DWI			3in1-DWI
	Image quality	Lesion visibility	Sharpness of breast contour	Image quality	Lesion visibility	Sharpness of breast contour	Image quality	Lesion visibility	Sharpness of breast contour
A vs. B	0.217 (0.016)	0.111 (0.263)	−0.004 (0.923)	−0.05 (0.593)	0.075 (0.509)	0.048 (0.577)	0.056 (0.606)	0.111 (0.263)	0.024 (0.741)

Values are presented as Kappa value (P value).

Fig. 1.

Typical patient images depicting complex fibroadenoma with extensive necrosis (a), simple fibroadenoma (b), and invasive ductal carcinoma (c) with contrast-enhanced MRI, c-DWI, t-DWI, and 3in1-DWI (from left to right). (a) The lesion showing rim enhancement and t-DWI provides significantly increased anatomic detail compared with c-DWI and 3in1-DWI. (b) Significantly stronger artifacts are noted on c-DWI compared with t-DWI and 3in1-DWI. (c) The breast contour is clearly depicted on both t-DWI and 3in1-DWI but it is not sharp on c-DWI. Arrow = lesion, artifact, and breast contour in figure (a), (b), and (c), respectively.

Comparisons of the ADC values among three DWI sequences

Mauchly’s sphericity assumption was violated by the test (P = 0.000051) and then univariate tests with Huynh–Feldt Epsilon correction were used for data analysis. The results showed that the interaction between lesion type effects and diffusion direction effects was not significant (F = 1.778, P = 0.181). The main effects of lesion type (F = 86.831, P = 0.000) and the diffusion direction (F = 6.8, P = 0.003) were significant. The results of multiple comparison tests indicated that the differences in the mean ADC values between t-DWI and 3in1-DWI were significant (P = 0.000059). ADC values with 3in1-DWI were significantly reduced compared with t-DWI. Fig. 2 presents error bars showing the mean ADC values of three DWI sequences (Fig. 2a) and box plot showing distributions of the ADC values of all evaluated DWI sequences divided into benign and malignant lesions (Fig. 2b). Table 4 summarized the ADC values of the two lesion types and three DWI sequences.

Fig. 2.

Error bars showing the mean ADC values of different DWI sequences (a) and box plot showing distributions of the ADC values of them divided into benign and malignant lesions (b). *P < 0.01.

Table 4.

Summarization for ADC values of the two lesion types and the three different DWI sequences.

	c-DWI	t-DWI	3in1-DWI
Benign lesions (n = 34)	1.679 ± 0.251	1.717 ± 0.261	1.683 ± 0.253
Malignant lesions (n = 26)	1.178 ± 0.173	1.193 ± 0.167	1.136 ± 0.171

Discussion

DWI with tetrahedral or 3in1 diffusion gradients can be used clinically for breast examinations. Radiologists rated tetrahedral or 3in1-DWI as exhibiting higher image quality and sharpness of breast contour than c-DWI. When comparing the ADC values of breast lesions from t-DWI and 3 in1-DWI to c-DWI, it was found that the ADC values of the former techniques were comparable to c-DWI.

Previous reports have compared image quality between t-DWI and c-DWI. Sinha et al. (16) found that the signal-to-noise ratio (SNR) advantage of t-DWI is obvious through comparing the ADC maps of t-DWI and c-DWI visually. Takashi et al. (17) demonstrated that background tissue was more easily determined with t-DWI compared with c-DWI. Masatoshi et al. (19) considered that the quality of the images acquired using the tetrahedral technique was significantly superior to those obtained using the orthogonal technique in terms of the signal homogeneity in the right lobe of the liver and the overall image quality. Masanori et al. (20) demonstrated that the image quality was visually better for velocity-compensated DWI combined with tetrahedral gradients compared with velocity-compensated DWI and c-DWI. Usha et al. (21) compared diffusion images of human calf muscle acquired with the tetrahedral gradients to orthogonal pattern and confirmed the improvement of the SNR of the average diffusion map using tetrahedral sampling to orthogonal sampling by the decrease in standard deviation of the trace maps. Yamaguchi et al. (18) investigated the geometric distortion of the t-DWI and c-DWI using four different phantoms of coefficient of viscosity, and found that geometric distortion was improved in t-DWI.

As expected and consistent with prior publications, the visual grades of image quality and the sharpness of the breast contour are higher in t-DWI and 3in1-DWI than in c-DWI. The differences were not significant between t-DWI and 3in1-DWI. The primary underlying concern is that a longer TE will result in a lower SNR (22). Thus, the TE should generally be set as short as possible. Table 1 indicates that TE was shorter in t-DWI and 3in1-DWI compared with c-DWI. Three orthogonal gradient pulses are simultaneously applied in the tetrahedral and 3in1 gradient patterns (18,23). Therefore, the resulting TE is shorter in t-DWI and 3in1-DWI compared with c-DWI for a given b value. In addition, the more uniform background anatomical structure with t-DWI and 3in1-DWI allowed greater reliability in radiologic interpretation. The sharper breast contour with t-DWI and 3in1-DWI allowed more accurate positioning of lesions, that is of great significance to the patients who are not suitable for contrast medium injection. In terms of lesion visibility, no significant differences were found among three DWI sequences by both radiologists.

In this study, the ADC values extracted from both t-DWI and 3in1-DWI were not significantly different compared with those of c-DWI. The mean ADC values of 60 lesions were 1.462 × 10⁻³ mm² s⁻¹, 1.490 × 10⁻³ mm² s⁻¹, and 1.446 × 10⁻³ mm² s⁻¹ for c-DWI, t-DWI, and 3in1-DWI, respectively (Fig. 2a). This result is partly concordant with the results of two previous reports by Takashi et al. (17) and Yamaguchi (18). Takashi et al. conducted a phantom study at 1.5 T to verify the consistency of ADC values obtained from tetrahedral and orthogonal DWI. The ADC values obtained by both were strongly correlated, and the slope of linear regression analysis was 1.02, which is higher than 1.00. Yamaguchi et al. compared the ADC values of the t-DWI and c-DWI in the shortest TE in four different phantoms of b = 1000 s mm⁻². The clustered bar charts showed that the mean ADC value of t-DWI was slightly increased compared to c-DWI in all four phantoms. Moreover, the mean ADC value of 60 lesions was significantly higher for t-DWI than for 3in1-DWI. Unfortunately, a reasonable explanation cannot be given for this finding.

Increasing awareness of the potential of overdiagnosis is leading to the evaluation of new noninvasive imaging diagnostic tools to discriminate malignant from benign lesions (24). DWI is a promising technique to differentiate benign from malignant breast lesions (25). Thus, the ADC values between benign and malignant lesions were compared in all three DWI sequences and a statistically significant difference in the ADC values were observed between benign and malignant mass-like lesions. In all three DWI sequences, the ADC value of malignant lesions was significantly reduced compared with benign lesions.

A strength of the current study is that this is the first study to prospectively assess the influence of diffusion gradient direction on DWI of breast lesions at 3.0 T. With the development of hardware and software technology, high field MR systems ≥ 3.0 T are increasingly used clinically. However, further studies are necessary to confirm the preliminary results generated with 1.5 T MR systems. In addition, to the best of our knowledge, few reports are available about 3in1 gradient patterns. The advantages of the 3in1 acquisition scheme were similar to the tetrahedral scheme, e.g. the strength of 3in1 gradients is $\sqrt{3}$ times the orthogonal input components. In addition, 3in1’s unique advantage involves easy control of the scanning time. In our study, although the signal averages were set to 12 in 3in1-DWI, the scanning time was only 1 min 14 s. Of course, 3in1 gradient pattern also has limitations, including only one diffusion direction. In general, more directions help obtain precise information about water molecule diffusion.

The present study also has limitations. First, non-mass-like enhancement lesions were excluded because it is much more challenging to define the boundaries of the lesions of this type. Second, the types of benign and malignant lesions were not multiple. Most benign and malignant lesions were fibroadenomas and invasive ductal carcinoma, respectively. Third, although visual grade is an accepted parameter for assessment, it is a subjective criterion that largely depends on the expertise of the radiologist. Fourth, despite the best efforts of the radiographer, a degree of subjectivity in the selection of the ROIs for ADC measurements cannot be excluded.

In conclusion, DWI with tetrahedral or 3in1 diffusion gradients is a more useful technique for clinical breast MRI than c-DWI because the image quality and the sharpness of the breast contour are improved. The ADC value is comparable to that of c-DWI. In the future, further studies with multiple types of lesions are needed to investigate the influence of diffusion gradient direction on the detection and diagnostic performance of breast lesions at 3.0T.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Le Bihan

. Molecular diffusion nuclear magnetic resonance imaging. Magn Reson Q 1991; 7: 1–30.

Thomassin-Naggara

De Bazelaire

Chopier

et al.

Diffusion-weighted MR imaging of the breast: advantages and pitfalls. Eur J Radiol 2013; 82: 435–443.

Zhu

Arlinghaus

Whisenant

et al.

Sequence design and evaluation of the reproducibility of water-selective diffusion-weighted imaging of the breast at 3.0T. NMR Biomed 2014; 27: 1030–1036.

Englander

Ulug

Brem

et al.

Diffusion imaging of human breast. NMR Biomed 1997; 10: 348–352.

Baltzer

Renz

Herrmann

et al.

Diffusion-weighted imaging (DWI) in MR mammography (MRM): clinical comparison of echo planar imaging (EPI) and half-Fourier single-shot turbo spin echo (HASTE) diffusion techniques. Eur Radiol 2009; 19: 1612–1620.

Granlund

Staroswiecki

Alley

et al.

High-resolution, three-dimensional diffusion-weighted breast imaging using DESS. Magn Reson Imaging 2014; 32: 330–341.

Bogner

Pinker-Domenig

Bickel

et al.

Readout-segmented echo-planar imaging improves the diagnostic performance of diffusion-weighted MR breast examination at 3.0T. Radiology 2012; 263: 64–76.

Bogner

Pinker

Zaric

et al.

Bilateral diffusion-weighted MR imaging of breast tumors with submillimeter resolution using readout-segmented echo-planar imaging at 7T. Radiology 2015; 274: 74–84.

Filli

Ghafoor

Kenkel

et al.

Simultaneous multi-slice readout-segmented echo planar imaging for accelerated diffusion-weighted imaging of the breast. Eur J Radiol 2016; 85: 274–278.

10.

Chen

Jin

et al.

Conspicuity of breast lesions at different b values on diffusion-weighted imaging. BMC Cancer 2012; 12: 334–341.

11.

Nilsen

Fangberget

Geier

et al.

Quantitative analysis of diffusion-weighted magnetic resonance imaging in malignant breast lesions using different b value combinations. Eur Radiol 2013; 23: 1027–1033.

12.

Mürtz

Tsesarskiy

Kowal

et al.

Diffusion-weighted magnetic resonance imaging of breast lesions: the influence of different fat-suppression techniques on quantitative measurements and their reproducibility. Eur Radiol 2014; 24: 2540–2551.

13.

Ouyang

Zhu

et al.

Diffusion-weighted imaging with fat suppression using short-tau inversion recovery: Clinical utility for diagnosis of breast lesions. Clin Radiol 2014; 69: e337–e344.

14.

Tamura

Usui

Murakami

et al.

Comparisons of multi b-value DWI signal analysis with pathological specimen of breast cancer. Magn Reson Med 2012; 68: 890–897.

15.

Liu

Liang

Liu

et al.

Intravoxel incoherent motion (IVIM) in evaluation of breast lesions: comparison with conventional DWI. Eur J Radiol 2013; 82: e782–e789.

16.

Sinha

Lucas-Quesada

Sinha

et al.

In vivo diffusion-weighted MRI of the breast: potential for lesion characterization. J Magn Reson Imaging 2002; 15: 693–704.

17.

Takashi

Sachiko

Naoki

et al.

Breast diffusion-weighted MRI: comparison of tetrahedral versus orthogonal diffusion sensitization for detection and localization of mass lesions. J Magn Reson Imaging 2011; 33: 1375–1381.

18.

Yamaguchi

Ogura

Hayashi

et al.

Evaluation of imaging quality and ADC for the tetrahedral gradients of diffusion weighted imaging. Nihon Hoshasen Gijutsu Gakkai Zasshi 2015; 71: 595–604.

19.

Masatoshi

Tonsok

Takamichi

et al.

Isotropic diffusion-weighted MR imaging with tetrahedral gradients in the upper abdomen. Magn Reson Med Sci 2006; 5: 201–206.

20.

Masanori

Yusuke

Tosiaki

et al.

Motion artifact reduction of diffusion-weighted MRI of the liver:use of velocity-compensated diffusion gradients combined with tetrahedral gradients. J Magn Reson Imaging 2013; 37: 172–178.

21.

Usha

Lawrence

. In vivo diffusion tensor imaging of human calf muscle. J Magn Reson Imaging 2002; 15: 87–95.

22.

Holdsworth

Yeom

Antonucci

et al.

Diffusion-weighted imaging with dual- echo echo-planar imaging for better sensitivity to acute stroke. Am J Neuroradiol 2014; 35: 1293–1302.

23.

Thomas

Rovert

Erbil

et al.

Encoding of anisotropic diffusion with tetrahedral gradients:a general mathematical diffusion formalism and experimental results. Magn Reson Med 1996; 35: 399–412.

24.

Ramírez-Galván

Cardona-Huerta

Ibarra-Fombona

et al.

Apparent diffusion coefficient (ADC) value to evaluate BI-RADS 4 breast lesions: correlation with pathological findings. Clin Imag 2015; 39: 51–55.

25.

Satake

Nishio

Ikeda

et al.

Predictive value for malignancy of suspicious breast masses of BI-RADS categories 4 and 5 using ultrasound elastography and MR diffusion-weighted imaging. Am J Roentgenol 2011; 196: 202–209.