| Objective:To demonstrate the distinguishing features of signal attenuation curve in normal breast tissue and in different mass-like lesions of breast in multi-b-value diffution weighted imaging which based on the intravoxel incoherent motion(IVIM) theory; To study the differences of normal breast tissue and mass-like lesions inapparent diffusion coefficient (ADC),the parameter of conventional monoexponential model diffusion-weighted imaging(DWI), and pure diffusion coefficient (D), perfusion fraction(f) andperfusion related diffusion coefficient(D*) value, the parameters of IVIM diffusion-weighted imaging(IVIM-DWI); To evaluate the diagnostic and differential capacity of these parameters to identify malignant lesions from benign lesions and normal gland. By the cut-off point of 20%, all malignant lesions were divided to two groups according to the expression of proliferativebiological marker Ki-67.All parameters above were compared between two groups and the relationship between the Ki-67 level of high expression group and their parameters of DWI and IVIM-DWI was explored.Materials and Methods:By foresight,144 cases of patients examined by breast MRI after finding breast mass via clinical diagnosis, breast X-ray or ultrasound were collected. The examination sequence was successively axial T1 weighted imaging (T1WI), axial and bilateral sagittal SPIR T2 weighted imaging (T2WI), axial conventional DWI (b value was 0 and 800s/mm~2) as well as IVIM-DWI (10 b values was respectively 0,20,65, 110,155,200,300,500,800 and 1000s/mm~2) and dynamic enhancement scanning. ADC value of lesions and normal gland tissues were measured by Philips IntelliSpace Portal workstation; IVIM-DWI image was processed by a third-party software MITK-diffusion. After sketching ROI, the software, by double-exponential model, automatically worked out the value of pure diffusion coefficient D, perfusion fraction f and pseudo-diffusion coefficient D* and signal attenuation image was generated correspondingly. After MRI examination,127 cases (135 lesions in total) with pathological results were acquired through operation or aspiration biopsy. According to the pathological results, the lesions were divided into breast cancer group, cyst group, inflammatory lesion group and group taken contralateral gland as the gland. Statistical analysis was conducted by adopting SPSS 18.0. The differential diagnosibility of ADC value, D value, D* value and f value between breast cancer, benign tumor, cyst, inflammatory lesion and gland to benign and malicious lesions was analyzed and compared. As per receiver operating characteristic curve (ROC), area under the curve (AUC) was calculated; the optimal diagnostic cut-off value as well as corresponding sensitivity and specificity were confirmed.Expression results of antigen marker of cell proliferation Ki-67 in breast cancer group were collected.≥20% was high expression group and <20% was low expression group. Differences of ADC value, D value, D* value and f value between high and low expression groups of breast cancer were compared. The relationship between Ki-67 percentage and various parameters in high expression group was analyzed.Result:In the 127 patients(include 135 lesions), there were 75 cases(75 lesions) of breast cancers,27 cases(32 lesions) of benign tumors,10 cases(13 lesions)of cysts,15 cases(15 lesions)of inflammatory lesions and 127 contralateral breast tissues. The signal of breast gland and lesion tissues attenuated within a big range of b value. Along with the increase of b value, cyst took on single-exponential linear attenuation; gland tissues, inflammatory lesions, benign tumors and breast cancer took on double-exponential non-linear attenuation. Within the range of low b value b≤200 s/mm~2), the signal of normal tissues and lesions except for cyst had a more significant downtrend, of which the attenuation degree of breast cancer signal was quicker compared to that of benign tumor, inflammatory lesion and cyst. Then, the attenuation degree of gland and lesion signal slowed down inordinately along with the increase of b value. When the b was at its maximum (b=1000s/mm~2), all cyst lesions and most of gland signals attenuated to extremely low level; part of glands and benign tumors had relatively low signals; residual signal of inflammatory lesions had a relatively big range; generally breast cancer still had comparatively high signal. One-way analysis of variance showed that the differences of conventional DWI parameters, ADC value, multi-b value parameter D value, f value and D* value between breast cancers, benign tumors, inflammatory lesions, cysts and normal gland tissues were all with statistical significance; ADC values in the five groups, excluded normal gland tissues and benign tumors, were with statistical difference (p<0.05). Cysts had the highest ADC value and breast cancers the lowest. There were statistical difference of D values in any two of all five groups (p<0.05), with the values from high to low respectively: cysts, normal gland tissues, benign tumors, inflammatory lesions and breast cancers. There were no of f values between inflammatory lesions and breast cancers and between normal gland tissues, while other pairs with statistical difference (p<0.05). Malignant group had the highest f value and cyst group the lowest. As for D* value, only comparison between cyst group and other four groups was with statistical difference (p<0.05). Comparing D value and ADC value of different lesions and normal gland, there were statistical difference between the two parameters in malignant group, benign tumor group, inflammatory group and normal gland (p< 0.001), with ADC value higher than D value. Area under the curve obtained from diagnostic experiment of D, ADC and f values was with statistical difference with their AUC respectively being 0.930,0.898 and 0.768. D value had the best diagnostic effect. The optimal cut-off value of the three diagnosing breast gland’s benign and malicious lesions was respectively 0.89×10-3mm~2/s,1.08×10-3mm~2/s and 5.54%; the sensitivity of breast gland’s benign and malicious lesions diagnosed by this threshold value was respectively 95.7%,89.3% and 79.2%. The specificity was respectively 92.9%,87.5% and 71.4%.In the 75 cases of breast cancer group, there were 53 cases in Ki-67 high expression group and 22 in low expression group. ADC value, D value, f value and D* value between two groups were all without statistical difference (P>0.05). The correlation between Ki-67 percentage in high expression group and various parameter values f value (r=0.184, P=0.178), D value (r=-0.005, P=0.968), D* value (r=-0.062, P=0.654) and ADC value (r=-0.095, P=0.489) was not significant.Conclusion:1. In IVIM-DWI, the cyst signal takes on single-exponential linear attenuation; the signals of gland tissues, inflammatory lesions, benign tumors and breast cancer take on double-exponential non-linear attenuation.Within the range of low b value (b≤ 200s/mm~2), attenuation degree of breast cancer signal is faster than that of gland, benign tumor, inflammatory lesion and cyst.2. InIVIM-DWI, pure diffusion coefficient (D), perfusion fraction(f) andperfusion related diffusion coefficient(D*) were beneficial for between benign and malicious lesions of breast.D value has the highest diagnostic efficiency.3. As for differential diagnosis between benign and maglignant lesions in breast, the optimal cut-off of D value for prediction is 0.89mm~2/s (AUC=0.930, P<0.001), with the sensitivity of 95.7% and specificity of 92.9%. Comparing to the ADC value of conversional DWI, with the optimal cut-off of 1.08mm~2/s and the sensitivity of 89.3% and the specificity 87.5%for prediction, D value oflVIM-DWI performed better.4. The values of ADC, D, f and D* between high and low groups of Ki-67 expression were all without significant difference. The correlation between parameters aboved and Ki-67 level in high expression group was not significant. |
PDF Full Text Request