Synopsis
The
effects of different numbers of b values and excitations on the non-Gaussian
DWI parameter estimates in the breast of normal volunteers was assessed. A good agreement of
Do, K and synthetic ADC200-1500 was observed among different number of b values as well as excitations. Lower Do and sADC200-1500
were found in women with lactation period. Higher K and fIVIM were observed in
lactating volunteers with some overlap. A limited protocol using only 5 b
values could be relevant in clinical setting, resulting in a remarkable
reduction in acquisition time.Introduction
Diffusion
weighted imaging (DWI) is increasingly used for the breast cancer diagnosis (1).
Non-Gaussianity for water diffusion can be quantified by means of diffusion
kurtosis model (2,3) beyond ADC, however, the accurate estimation of such
parameters requires many
acquisition with many b values, which is a limitation in
a clinical setting. To establish an image acquisition protocol clinically
relevant in short scanning time, we have investigated the effects of various
scanning schemes (number of b values and excitations) on the non-Gaussian DWI
parameter estimates in the breast of normal volunteers.
Materials and Methods
13 non-lactating
and 3 lactating volunteers were recruited in this IRB approved prospective study. Breast MRI
was performed using a 3-T system (Trio, B17; Siemens AG) equipped with a
dedicated 16-channel breast array coil. The following fat-suppressed DWI （single shot EPI) images
were obtained along three orthogonal axes;
・16 b values of 0, 5, 10, 20, 30, 50, 70, 100, 200, 400, 600, 800, 1000, 1500, 2000, 2500 sec/mm2 with 1 number of excitation (NEX) and a scan time of 3 min 55 sec.
・5 b values of 0, 100, 200, 1500 and 2500 sec/mm2 with 3 NEX and a total scan time of 3min 30sec.
For both protocols acquisition parameters were: repetition time/echo time 4,600/86 ms, flip angle 90°, field of view 160×300 mm2, matrix 80×166, and slice thickness 3.0 mm.
ROIs were placed onto the normal breast tissue and images processing was performed using software implemented in Matlab (Mathwork, Natick, MA) comprising the following steps:
1/Noise correction to handle Rician noise at each b value:
S(b)2 = Sn(b)2 + NCF [1]
where NCF (noise correction factor) is a parameter which characterizes the “intrinsic” non-Gaussian noise contribution within the images (4).
2/The corrected signal acquired with b>200 s/mm² was fitted using the kurtosis diffusion model to estimate ADCo and K:
S/So = exp[-bADCo + K(bADCo)²/6] [2]
where S0 is the theoretical signal acquired at b=0, fIVIM the (T1,T2-weighted) volume fraction of incoherently flowing blood in the tissue, D* the pseudo-diffusion coefficient associated to the IVIM effect, ADCo the virtual ADC which would be obtained when b approaches 0, K the kurtosis parameter.
3/Then, the fitted diffusion signal component was subtracted from the corrected raw signal acquired with b<200s/mm² and the remaining signal was fitted using the IVIM model (4) to get estimates of the flowing blood fraction, fIVIM, and the pseudodiffusion, D*.
Additionally a synthetic ADC encompassing both Gaussian and non-Gaussian diffusion effects (6), sADC200-1500, was defined using only 2 b values as:
sADC200-1500 = ln [Sn(b200)/Sn(b1500)]/1300 [3]
The IVIM/DWI parameters were calculated from the datasets acquired with 16 b values (1 NEX), 5 b values (1 NEX), and 5 b values (3 NEX). The reproducibility for each parameter was assessed using intra-class correlation coefficients (ICCs) with absolute agreement.
Results
No remarkable difference in the DWI image quality was observed regardless
of the number of b values
and excitations (Figure 1).
Overall there was a good agreement of Do, K and
sADC200-1500
among different numbers
of b values or excitations (Tables 1, 2). Notably a good agreement of
diffusion parameters was observed among 16 b values (1
NEX), 5 b values (1 NEX), and 5 b values (3 NEX). Lower Do and
sADC200-1500 were
found in women with lactation period in accordance with the
literature (Figure 2) (5).
Higher K and
fIVIM were observed in lactating volunteers with some overlap.
Discussion&Conclusion
There
was no significant difference between IVIM/non Gaussian parameter estimated
values in the normal breast tissue
regardless the number of b values or excitations used in this study. A limited
protocol using only 5 b values could be relevant in clinical setting, resulting
in a significant reduction in acquisition time. Ultimately a synthetic ADC
calculated from only 2 values could provide information combining Gaussian and
non-Gaussian diffusion effects (6). Protocols with more b values might still be
required in the case of noisy data sets or to improve the accuracy on estimated
non-Gaussian and IVIM parameters. Interestingly some difference of non-Gaussian
diffusion and IVIM parameters was observed with lactation status, and there
might be a need for the consideration of lactation status when assessing breast
DWI data. The results of this preliminary study will need to be extended to a large scale population with
a wide range of breast lesions.
Acknowledgements
This work was supported
by Hakubi Project of Kyoto University and JSPS KAKENHI Grant.References
1. Partridge S et al. Differential diagnosis of mammographically and clinically occult breast lesions on diffusion-weighted MRI. Journal of Magnetic Resonance Imaging. 2010;31:562-570.
2. Chabert S et al. Relevance of the information about the diffusion distribution in vivo given by kurtosis in q-space imaging. Proceedings of the 12th Annual Meeting of ISMRM, Kyoto, Japan, 2004, p.1238.
3. Jensen JH et al. MRI quantification of non-Gaussian water diffusion by kurtosis analysis NMR in Biomedicine. 2010;23:698-710.
4. Iima M et al. Quantitative Non-Gaussian Diffusion and Intravoxel Incoherent Motion Magnetic Resonance Imaging: Differentiation of Malignant and Benign Breast Lesions. Investigative Radiology. 2015;50:205-11.
5. Nissan N et al. Diffusion-Tensor MR Imaging of the Breast: Hormonal Regulation. Radiology.2014:271:672-680.
6. Iima M et al. Clinical Intravoxel Incoherent Motion and Diffusion MR Imaging: Past, Present and Future. Radiology. (in press, Dec 2015)