| Dynamic contrast-enhanced breast MRI is the conventional sequence, the time-signal intensity curve(TIC)would be help in the identification of benign and malignant breast lesions. The current study think that the accuracy of diagnosis of malignant lesions with TIC type Ⅲ may be 90%-93%,which plays an important role in clinical practice.Howerer, the accuracy of diagnosis of malignant lesions with TIC type Ⅱand typeⅠdecrease greatly to 50%-70% and to 15-35%.It is puzzledfrequently to distinguish malignant lesions with TIC typeⅠor Ⅱ from benign lesions. ADCderived fromthe traditional diffusion weighted imaging(DWI) is valuable for identification of benign and malignant breast lesions.Nevertheless,the traditional ADC could not distinguish the effect of microcirculation perfusion from the water moleculesdiffusion,thus,overestimate the real value.Moreover,water molecules diffuse ininhomogeneous organisms follows non-Gaussian distribution.In consequence,the diagnosticvalue of traditional DWI would be limited. This clinical research planned to investigate the diagnosis value of combining IVIM and DKI for distinguishbreast lesions, especially lesions with TIC typeⅠor Ⅱ,in order to improve the ability of identification of benign and malignant lesions which are difficult to diagnose qualitatively.Part I the Application of IVIM and DKI in Evaluation of Breast LesionsObjective To investigate the role of diffusion weighted imaging(DWI) based on IVIM and non-Gaussian distribution DKI in characterizing breast lesionsMaterials and Methods Breast patients who underwentpreoperative diffusion-weighted MRI at 1.5T and multiple b values diffusion weighted scanning varying in February 2015- January 2016 were consecutively included in this prospective study. Of the consecutive 120 patients,a total of 132 lesions, 88 were malignant lesions and 44 were benign lesions. Randomly selected 56 contralateral normal tissues as control.The pathological results were abtained finally. ADC was abtained with monoexponential fitting of the DWI data(b = 0 and 1000 s/mm2). Tissue diffusivity(D), perfusion fraction(f), pseudo-diffusion coefficient(D*) and Blood flow(BF) were calculated using segmented biexponential analysis. Mean kurtosis(MK) and mean diffusivity(MD)were calculated using DKI model.Those parameters werecompared by One-way analysis of variance(ANOVA) with Tukey’s multiple comparison test and further comparison by Least Significant Difference(LSD) comparison.The Receiver operatingcharacteristic(ROC) analysis wasperformed to assess the sensitivity and specificity of these parameters in the diagnosis ofbreast lesions by the areaunder the curve(AUC),compared the AUC difference by Z test. Using parallel experiment computing the sensitivity of the joint diagnosis.Result(1) Dvalues were significantly different among malignant tumors, benign lesions, and normal breast tissues(P = 0.000) and it was the same result for ADC、D*、MK、MD values.There were no significantdifferences among malignant tumors, benign lesions and normalbreast tissues in f and BF values(P>0.05). Further comparisons of these parameters between every single pair were as the following. D、MD and ADC values of malignanttumors were significantly smaller than those of benign lesions and normal tissues(P<0.01,respectively).MK values of malignanttumorswere significantly higher than those of benign lesions and normal tissues(P<0.01,respectively). There was no significant difference between malignant tumors and benign lesions in D* values,while significantdifference was shown in D* value between every otherpair.(2)D values were smaller than ADC values in every group,in contrast,ADC values were smaller than MD values.MD,ADC,D showed a negative correlation with MK.(3)With pathology as the gold standard,the AUC of ADC values was 0.846 for distinguishing malignantlesions from benign lesions.At diagnostic threshold of 1.19 × 10-3mm2/s for diagnosis of malignancy,the sensitivity and specificity were 80.8% and 78.0%.(4)With pathology as the gold standard,the AUC of D values(0.881) was bigger than ADC values for distinguishing malignant lesions from benign lesions,with no significant difference(P>0.05).At diagnostic threshold of 0.92×10-3mm2/s for diagnosis of malignancy,the sensitivity and specificity(88.1%,83.1%)were higher than ADC values(80.8%, 78.0%).The AUC of D* values was 0.679 fordistinguishing malignant lesions from benign lesions and provide moderate sensitivity and specificity.(5)With pathology as the gold standard,the AUC of MD and MK values were 0.936 and 0.911.The AUCof MD was higher than D and ADC values with significant difference(P<0.05)At diagnosis thresholds of 1.48×10-3mm2/s, and 0.78 for MD、MK,the sensitivity were 82.2% and 91.5%,and specificifitywere 98.3% and 85.3%,respectively.Combining MD and MK values had a sensitivity up to 98.4%.(6)Combining MD and D values had a sensitivity up to 99.0%.Conclusions(1) The diagnostic performance of D values derived from IVIM is better than ADC values.(2)The diagnostic efficiency of DKI parameters for breast lesions is the highest among all parameters.MD could increase diagnostic confidence of breast tumors.Part II the Diagnostic Value of IVIM and DKI in Evaluation of Breast Lesions with DCE TIC TypeⅠandⅡObjective To evaluate the diagnostic value of IVIM and DKI in breast lesions with DCE TIC TypeⅠandⅡ.Materials and Methods This prospective studyincluded 96 patients,total 108 breast lesions with DCE TIC TypeⅠandⅡ,containing 42 benign lesions and 66 maligant tumors.ADC was abtained with monoexponential fitting of the DWI data(b = 0 and 1000 s/mm2).Tissue diffusivity(D),perfusion fraction(f), pseudo- diffusion coefficient(D*)and Blood flow(BF)were calculated using segmented biexponential analysis. Mean kurtosis(MK) and mean diffusivity(MD)were calculated using DKI analysis. The difference of all parameters in benign and malignant lesions with TIC type Ⅰ and Ⅱ werecompared by independent-samples T test. ROCanalysis was performed to assess the sensitivity and specificity of these parameters in the diagnosis ofbreast lesions with TIC type ⅠandⅡ,compared the AUC difference by Z test. Using parallel experiment computing the sensitivity of the joint diagnosis.Result(1)There are 23、85、24 lesions, respectively, for DCE-MRI TIC typeⅠ、Ⅱand Ⅲ.The proportion of maglinant tumors within each type was30.4%(7 / 23),69.4%(59 / 85),91.7%(22 / 24).(2)There were significantdifferences between malignant tumors and benign lesions with DCE TIC TypeⅠandⅡin ADC、D、MD and MK values(P<0.05),while no significantdifferences were shown in f, D* and BF.(3) With pathology as the gold standard,the AUC of ADC values for distinguishing malignant lesions from benign lesions with DCE TIC TypeⅠandⅡ was 0.857, At diagnostic threshold of 1.04×10-3mm2/s for diagnosis of malignancy,thesensitivity and specificitywere 84.6% and 76.7%.(4) With pathology as the gold standard,the AUC of D values for distinguishing malignant lesions from benign lesions with DCE TIC TypeⅠandⅡ was 0.905,which was higher than that of ADC,with no significant difference(P>0.05).At diagnostic threshold of 0.98×10-3mm2/s for diagnosis of malignancy,thesensitivity and specificitywere 81.5% and 95.3%.(5)With pathology as the gold standard,the AUC of MD and MK values for distinguishing malignant lesions from benign lesions with DCE TIC TypeⅠandⅡ were 0.896 and 0.876,respectively. At diagnostic threshold of 1.30×10-3mm2/s、0.78 for diagnosis of malignancy,thesensitivity and specificityof MD and MK values were 86.2% and 88.4%, 90.7% and 80.0% respectively.(6)Combining D and MK values had a sensitivity of 98.3%.(7)ADC,D,MD in infiltrating ductal carcinoma(IDC),ductal carcinoma in situ(DCIS), fibroadenoma(FA)and mucinous adenocarcinoma(MAC) increasing gradually, on the contrary, MK valuesdecreasing, the difference was statistically significant.There was significant difference in ADC,D,MK,MD between IDC and DCIS, IDC and MAC, DCIS and MAC, IDC and FA, IDC and lobulated tumor, IDC and intraductal papilloma.Conclusions(1)The diagnostic efficiency of IVIM and DKI model for breast lesions with TIC TYPEⅠandⅡ is higher than those of the conventional DWI.D values shows higher specificity while MD and MK demonstrate higher sensitivity,but the parameters of D*, f, BF associated with microcirculation perfusion provide limited value.(2)MK in various pathological types has important reference value and it is the same result for ADC、D、MD values. |
PDF Full Text Request