The Application Of Intravoxel Incoherent Motion Diffusion-weighted Imaging In Evaluation Of Breast Lesions

Background and Purpose:Breast cancer is one of the most common malignant tumors in women, and thesecond leading cause death in women. Because it is quite different in the choice oftreatment and operation mode and the prognosis between benign and malignantlesions, the early diagnosis of breast cancer is very important. In addition to thetraditional breast mammography has been widely used in the diagnosis of benign andmalignant, the application of magnetic resonance imaging is more and more widely.Dynamic contrast-enhanced MRI of the breast is widely used for differentialdiagnosis of benign and malignant lesions of breast, it shows high diagnostic, but lowspecificity.Diffusion weighted imaging is a kind of magnetic resonance imaging technologywhich do not need contrast agent. By the mean of detecting the micro motion of watermolecules, it can reflect the pathological changes. It has been widely used in varioussystems of the body. In vivo, microscopic motion of water molecules detected byDWI is influenced not only by diffusion of water molecules, affected by the structuresof the tissue, but also by microcirculation of blood in the capillary network.Accordingly, the signal attenuation on monoexponential DWI sometimes does notrepresent a linear relationship and it is difficult to calculate the accurate ADC value.It is simple to set up the sequences Muti-b-value diffusion weighted imaging. Bymeans of scanning a series of b value and the analysis of biexponential model basedon intravoxel incoherent motion (IVIM) theory, the diffusion of water molecules andblood perfusion can be separated. On the one hand, it can accurately measure thediffusion of water molecules. Due to the exclusion the influence of blood perfusion,the pure diffusion coefficient (D value) could reflect the true state of diffusion of water molecules better, compared with the conventional ADC value. On the otherhand, it can obtain the information which reflect the flood perfusion of lesionswithout the contrast agent, as the perfusion parameters perfusion-related diffusioncoefficient and the （D*value）and microvascular volume fraction (f).By mean of analysising the image of conventional single b-valuemonoexponential model DWI and Muti-b-value DWI based on IVIM theory of87cases breast patients, we can obtion the conventional ADC value, pure diffusioncoefficient (D value), perfusion-related diffusion coefficient and the （D*value）andmicrovascular volume fraction (f) of the normal breast tissues, benign lesions, cystsand malignant tumors. The purpose of this study is to prospectively investigate thevariation of these parameters based on IVIM theory in different tissues and todetermine whether these parameters can distinguish benign from malignant breastlesions with higher accuracy, comparing with conventional ADC values.Materials and Methods:A total of102patients with breast tumor underwent magnetic resonance axialT2-weighted imaging, axial fat-suppression T2-weighted imaging, single b-factordiffusion-weighted imaging, extender b-factor diffusion-weighted imaging andDynamic contrast-enhanced MRI. A total of87cases (97lesions) were enrolled in thestudy and were analyzed. In the GE Advantage Windows4.5workstation, we makeuse of ADC software and IVIM processing software in the Function Tools kit to dealwith the single b-factor diffusion-weighted imaging and extender b-factordiffusion-weighted imaging. By monoexponential linear model, ADC processingsoftware can generate ADC parametric maps. By bioexponential nonlinear model,ADC processing software can generate D, D*and fparametric maps. We measure the ADC value, D value, D*value and f value ofbreast benign lesions group, cysts group, malignant group and breast contralateralnormal control group. One-way anova and LSD method were used to investigate thevariation of these parameters in different tissues and receiver operating characteristic curve was used to analysis and compare the ability of the parameters in differentialdiagnosis benign and malignant lesions.Results:1. First of all, The signals of the normal breast tissues, cysts, benign lesions andmalignant lesions decayed with a large-range b values, they are more obvious withinthe range of b value less than200s/mm2. Second, The signals decay of cystsrepresent a linear relationship, but the signals decay of normal breast tissues, benignlesions and malignant lesions represent a nonlinear relationship. At last, The signalsof the malignant lesions decayed more obvious within the range of b value less than200s/mm2compared with the benign lesions, normal breast tissues and cysts.2. D value is less than the ADC value in the breast benign lesions group,malignant group and breast contralateral normal control group. But D value is samewith the ADC value in the cysts group.3. The result of one-way anova show that there are significant differences amongbenign group, malignant group and control group in ADC value, D value, D*valueand f value. Further pairwise comparison, in addition to there are no significantdifferences between benign group and control group in D*value, there are significantdifferences in other parameter.ADC value and D value of breast malignant lesions group is less than that ofbenign lesions group, cysts group and breast contralateral normal control group andthere are significant differences between any other groups. F value and D*value ofbreast malignant lesions group is larger than that of benign lesions group, cysts groupand breast contralateral normal control group and there are significant differencesbetween any other groups.4. The AUC of ADC value, D value, D*value and f value for the differentialdiagnosis of benign and malignant lesions of breast are0.845,0.897,0.688and0.709respectively. The AUC of parameter D value is biggest.5. The optimal threshold of ADC value, D value, D*value and f value for thedifferential diagnosis of benign and malignant lesions of breast were1.3×10-3mm2/s、 1.0×10-3mm2/s、6.8×10-3mm2/s、21%if take the maximum Youden’s index as cut-off.The sensitivity of ADC value, D value, D*value and f value were87.2%、94.9%、92.3%、43.6%and the specificity of ADC value, D value, D*value and f value were78.0%、85.4%、41.5%、92.7%for the differential diagnosis of benign and malignantlesions of breast if take this threshold. Combining f and D values had a sensitivity upto99.61%for D value.Conclusion:The parameter （D value, D*value and f value）obtained from biexponentialmodel DWI based on intravoxel incoherent motion theory would help to differentialdiagnosis of benign and malignant lesions of breast. Compare the parameters ofmonoexponential model DWI and biexponential model DWI, The sensitivity and thespecificity of D value which is the parameter of monoexponential model DWI for thedifferential diagnosis of benign and malignant lesions of breast is higher, and thediagnostic performance of it is largest.