Study Of 3T Dynamic Contrast-enhanced And Diffusion-weighted MR Imaging In Differential Diagnosis Between Benign And Malignant Breast Lesions

Objective1.To investigate the feasibility of breast MRI with body surface coil and self-made breast stent in high-magnetic field(3.0T) MR scanner.2.To evaluate the value of dynamic contrast-enhanced MR Imaging(DCE-MRI) in differential diagnosis between benign and malignant breast lesions by using Liver Acquisition with Volume Acceleration(LAVA) technique and analyzing the morphologic features and hemodynamic characteristics of benign and malignant breast lesions.3.To explore the value of quantitative apparent diffusion coefficient(ADC) in differential diagnosis between benign and malignant breast lesions by using diffusion-weighted MR Imaging(DWI) with ASSCT technique at different b values. Partâ… Study of LAVA dynamic contrast-enhanced MR Imaging in differential diagnosis between benign and malignant breast lesionsMaterials and Methods1.Subjects42(totally 47 lesions,10 cases of 11 malignant lesions,32 cases of 36 benign lesions) patients with breast lesions were enrolled during July 2007 to February 2008 in Nanfang Hospital.All the subjects were confirmed by pathology.2.Equipment and contrast mediumâ‘ All the patients were performed with GE Signa EXCITE HD 3.0T superconducting magnetic resonance scanner.â‘¡Abdomen surface 8-channel phased array coil and self-made breast stent were used.â‘¢Gd-DTPA was injected as paramagnetic contrast medium with high pressure injection(Ulrich corp).â‘£DCE-MRI data was post-processed using Functool SER(signa enhance ratio) software package on AW 4.3 workstation.3.Examination technique3.1 Plain scanningBilateral breast axial T₁WI,axial and sagittal FSE T₂WI with fat suppression were performed with scan parameters as followings:axial T₁WI+FS (TR/TE=450/9.7ms) and axial T₂WI+FS(TR/TE=4000/85ms) with 6mm slice thickness,1mm gap,FOV=32cm×32cm,NEX=2.0,320×224 matrix;Sagittal FSE T₂WI+FS(4675/85ms) with 5mm slice thickness,1mm gap,FOV=22cm×22cm, NEX=2.0,320×224 matrix.3.2 LAVA DCE-MRIDynamic contrast-enhanced MRI was performed with LAVA sequence (TR/TE=3.6/1.7ms,20°flip angle,4.4mm slice thickness with 50%overlapping, 128×128 matrix,FOV=32cm×32cm,NEX=0.75,28s acquisition time).Mask imaging was first scanned without contrast medium,then scanning were repeated when injecting of contrast agent Gd-DTPA(flow rate of 2.0ml/s and dose of 0.2mmol/kg) by the first time,scanned again at the 1^th,2^th,3^th,4^th,6^th,8^th minute after injection with the same parameters and acquisition time(28s).4.Data processing and data analysisAider scanning,all the primitive data was sent to GE AW 4.3 workstation.The time-signal intensity curve(TIC) of each ROI was automatically generated with Functool SER(signal enhance ratio) software.ROIs were chosen in slice where lesions displayed most clearly.The morphologic features and hemodynamic characteristics of benign and malignant breast lesions were compared and the En%, time to peak and Slope_max were calculated respectively.5.Statistical analysisStatistical analysis were made using SPSS for windows 13.0.The statistically significant difference was set at P＜0.05.Pearson Chi-Square(x²) Test was used to compare T₂WI+FS signal intensity, time to peak and the numbers of each types of TIC.Early enhancement rate(En%) and Slope_max were recorded as(means±standard deviation) and were compared by the means of Independent-Samples T Test,but if non-homo geneity variance,compared by the two Independent Samples Tests.Descriptive statistical analysis method was used in imaging observation and analysis.The sensitivity,specificity and accuracy of each index(such as morphology, early enhancement rate,time to peak and the types of TIC) in diagnosing breast lesion were calculated.Results1.Pathological results:42 patients with breast lesions underwent DCE-MRI scanning;in which 11 were benign(23.4%) and 36 were malignant(76.6%).2.MRI feature2.1 Signal intensity(SI):Most of the breast lesions showed the same signal intensity as normal breast tissue on T₁WI+FS.On T₂WI+FS,the SI of breast lesions were different from that of normal breast tissue,but there was no statistically difference(x²=2.089,P=0.352) between benign and malignant breast lesions.2.2 Morphologic features:Benign breast lesions were mainly characterized as regular mass(26/36),clear boundaries(27/36),homogeneous enhancement(30/36); on the other hand,the malignant breast lesions as irregular shape(7/11),unclear boundaries(8/11),inhomogeneous enhancement(5/11).With the morphologic features(irregular shape and unclear boundaries) as the diagnostic standards for malignant lesions,the following diagnostic indices emerged:sensitivity 63.6%, specificity 72.2%and accuracy 70.2%.3.LAVA DCE-MRI3.1 Time-signal intensity curve(TIC):Benign breast lesions were characterized by typeâ… curve(10/32),and malignant by typeâ…¢curve(8/11).Typeâ…¡curve could be seen both in benign(19/32) and malignant(3/11) breast lesions.Statistically significant difference was found in TIC between benign and malignant breast lesions (x²=17.930,P=0.000).With the typeâ…¡andâ…¢time-signal intensity curve as the diagnostic standards for malignant lesions,the following diagnostic indices emerged: sensitivity100.0%,specificity 38.9%and accuracy 53.1%.3.2 Slope_max,early enhancement rate(En%) and time to peak(TTP):Slope_max was non-homogeneity variance,and was compared with the two Independent Samples Tests.Statistically significant difference exited between benign and malignant breast lesions(Z=-3.897,P=0.000),the malignant breast lesions had higher Slope_max,value than that of benign lesions.Most of malignant breast lesions were enhanced at early stage(9/11),however, most of benign breast lesions were enhanced laterly(25/32).Statistically significant differences(P=0.000 and 0.001 respectively) exited in En%and TTP.With the early-phased enhancement more than 80%,peak time less than 2 minutes as the diagnostic standards for malignant lesions,the following diagnostic indices emerged: sensitivity 81.8%,81.8%,specificity 75.0%,75.5%and accuracy 76.6%,79.0%, respectively.Conclusions1) Our study indicates that using abdomen surface 8-channel phased array coil and self-made breast stent for breast MRI examination is feasible.2) Morphologic features can be displayed well in DCE-MRI.With the morphologic features(irregular shape and unclear boundaries) as the diagnostic standards for malignant lesions,the sensitivity was 63.6%,the specificity was 72.2% and the accuracy was 70.2%.But there was overlapping between benign and malignant breast lesions in some patients.3) LAVA DCE-MRI is an effective method in differential diagnosis between benign and malignant breast lesions,as it can give more morphologic and hemodynamic information.With the early-phased enhancement more than 80%,peak time less than 2 minutes and typeâ…¡andâ…¢time-signal intensity curve as the diagnostic standards for malignant lesions,the following diagnostic indices emerged: sensitivity 81.8%,81.8%,100.0%,specificity 75.0%,75.5%,38.9%and accuracy 76.6%,79.0%,53.1%respectively.Partâ…¡Study of ASSET diffusion-weighted MR Imaging in differential diagnosis between benign and malignant breast lesionsMaterials and Methods1.Subjects:The same as partâ… .2.Equipment and contrast medium:The same as partâ… .3.Examination technique:All the DWI examinations were performed before DCE-MRI examination,using EPI(echo planner imaging) sequence(TR/TE=2000/52.6ms,6mm slice thickness, 1mm interval gap,matrix 128x128,FOV=32cm×32cm,NEX=4.0) with ASSET (array spatial sensitivity encoding) technique.Two diffusion-weighted imaging were implemented at two different b values(600s/mm² and 1000s/mm²).4.Data processing and data analysisAll the primitive data was sent to GE AW 4.3 workstation for post-processing. ADC imaging were automatically generated with Functool sol,ware.The apparent diffusion coefficient(ADC) values were recorded and compared in different breast lesions(breast cancers,benign lesion and normal breast tissue) with the same b value and same lesions with the different b value(b=600s/mm² and b=1000s/mm²).5.Statistical MethodStatistical analysis were made using SPSS for windows 13.0.The statistically significant difference was set at P＜0.05. All data were recorded as(mean±standard deviation).ADC values were compared with one-way ANOVA at different b value among benign,malignant and normal breast tissue.On the other hand,ADC values of the same lesions with different b value(b=600s/mm² and b=1000s/mm²) were compared with Paired-Sample T Test.The one-side upper limits of 95%confidence interval of mean ADC values were adopted as the standards to separate the malignant from the benign breast lesions.The sensitivity and the specificity were calculated,then using ROC curve to analyze the value of the sensitivity and the specificity in diagnosing malignant breast lesions with different b value.The areas under ROC curves were compared with U Test.Results1.Pathological results:The same as partâ… .2.DWI signal intensityComparing with the normal breast tissue,both of the benign and malignant breast lesions demonstrated high SI on DWI imaging,but were opposite on ADC imaging.3.Quantitative ADC value analysisAt b=600s/mm²,the mean ADC values of malignant breast lesions,benign breast lesions and normal breast tissue were(1.1573±0.3915)×10^-3mm²/s,(1.6093±0.5342)×10^-3mm²/s,and(1.9508±0.3663)×10^-3mm²/s respectively.At b=1000s/mm²,were (0.9135±0.2484)×10^-3mm²/s,(1.4073±0.3860)×10^-3mm²/s,and(1.7237±0.3225)×10^-3mm²/s respectively.Statistically significant differenence(P＜0.05) existed in the mean ADCs value among benign lesions,malignant lesion and normal breast tissue, with the smallest in malignant,the second in benign,and the largest in normal.But there was an overlapping between benign and malignant lesions in some patients.The mean ADC values of normal breast tissue,benign breast lesions,and malignant breast lesions at different b values were compared with Paired-Sample T Test.The results indicated that statistically significant differenence(P=0.022, P=0.001) existed in normal breast tissue and benign breast lesions at different b values but not in malignant breast lesion(P=0.108) with ADC value at high b value (b=1000s/mm²) smaller than that at low b value(b=600s/mm²).4.Application of ADC value with different b value in diagnosing breast lesionsUsing the one-side upper limits of 95%confidence interval of mean ADCs as the standard to separate malignant from benign breast lesions,the sensitivity was 72.7% (b=600s/mm²,1.4230×10^-3mm²/s),45.5%(b=1000s/mm²,1.0864×10^-3mm²/s),and specificity 83.1%,87.0%,accuracy 81.8%,84.1%,respectively.The areas under ROC curves were(Az) 600=0.863±0.043(b=600s/mm²,P =0.000),(Az) 1000=0.905±0.035(b=1000s/mm²,P=0.000).Statistically significant difference(P＜0.05) existed in areas under the ROC curves at both b values,which indicated that these two DWI with different b values could be used in separating malignant from the benign breast lesions.But there was no significant difference (P＞0.05) between these two b values(b=600s/mm² and b=1000s/mm²) in separating malignant from the benign breast lesions.Conclusions1) Diffusion-weighted imaging is an effective and feasible method to diagnosing breast lesions.ADC value can provide important information for differential diagnosis.2) Quantitative ADC value is useful for differential diagnosis between benign and malignant breast lesions.Statistically significant differenence(P＜0.05) existed in the mean ADCs value among benign lesions,malignant lesion and normal breast tissue,with the smallest in malignant,the second in benign,and the largest in normal. But there was an overlapping between benign and malignant lesions.3) With b value increasing in DWI,mean ADC value have the tendency of decreasing,and the image quality was decreased.4) Using the one-side upper limits of 95%confidence interval of mean ADCs as the standard to separate malignant from benign breast lesions,the sensitivity was 72.7%(b=600s/mm²,1.4230×10^-3mm²/s),45.5%(b=1000s/mm²,1.0864×10^-3mm²/s), specificity 83.1%,87.0%and accuracy 81.8%,84.1%respectively.