【作者】 胡瑛；

【导师】 张焱；

【作者基本信息】 郑州大学 ， 影像医学与核医学， 2009， 硕士

【摘要】 背景和目的:乳腺癌的发病率在全球范围内居女性肿瘤的首位,在欧美等西方发达国家,每9名妇女中,就有1人在一生中将患乳腺癌。我国原为乳腺癌低发国家,但近年来,随着生育模式、生活方式的改变,乳腺癌的发病率呈逐年上升趋势,且趋向年轻化,严重危害妇女的健康和生命。早期诊断、早期治疗是降低乳腺癌死亡率,提高乳腺癌患者5年生存率的重要措施。影像学检查对乳腺癌的早期诊断和临床分期具有重要价值。本研究旨在探讨3.0 T磁共振扩散加权成像（diffusion weighted imaging,DWI）、动态增强成像（dynamic contrast-enhancedmagnetic resonance imaging,DCE-MRI）及其联合应用对乳腺良、恶性疾病的诊断及鉴别诊断价值。材料与方法:收集郑州大学第一附属医院2008年05月至2009年02月经手术病理证实的乳腺病例55例,女性53例,男性2例;年龄21～67岁,平均44.5±10.9岁,临床主要表现为乳房肿块。磁共振共检出72个病灶,其中恶性肿瘤27例,30个病灶;良性病变28例,42个病灶。检查使用Siemens Magnetom Trio Tim 3.0T磁共振仪、专用双穴乳腺相控阵表面线圈,双侧乳腺、全乳同时扫描。所有病灶均行先行常规MR成像,然后依次行DWI、DCE-MRI扫描。常规MR成像采用快速小角度激发三维梯度回波序列（three-dimensional fast low angle shot spoiled gradient-echo,FLASH-3D）序列,横轴位T₁WI,和快速反转恢复序列（Turbo inversion recovery,TIR）横轴位T₂WI。DWI采用单次激发自旋-平面回波序列（spin echo-echo planner imaging,SE-EPI）,扩散敏感因子b值取0s/mm²、500s/mm²、1000s/mm²,层厚:4.0mm。DCE-MRI采用FLASH-3D序列,横轴位T₁WI,加脂肪抑制,TR/TE:4.23ms/1.57ms,层厚:0.9mm。注射对比剂前先平扫一次,平扫结束后从肘前静脉团注对比剂0.1mmol/kg,并用25ml生理盐水冲管,注射速度2.5ml/s。平扫结束后第20s开始增强扫描,并行连续无间断扫描6次。单次扫描时间60s,全乳扫描时间共计7min23s。将所得DWI及DCE-MRI图像传送Syngo工作站,使用表观扩散系数（apparent diffusion coefficient,ADC）计算软件及Mean Curve动态增强曲线生成软件进行数据处理。根据病变的形态特征、不同b值时的ADC值、早期增强率、强化峰值、达峰时间、时间-信号强度曲线对病变性质进行判断。以术后标本病理结果作为金标准,评价MR形态学、DWI、DCE-MRI及联合成像对乳腺疾病定性的诊断价值。感兴趣区（region of interest,ROI）选择避开出血、液化、坏死、囊变、钙化等区域,于病灶最大层面选取早期强化最显著的区域作为ROI;于对侧正常乳腺腺体组织丰富的相同区域选取相同大小ROI作为对照。结果:55例乳腺疾病患者,MRI检出及病理证实72个病灶,其中恶性肿瘤27例,30个病灶,包括浸润性导管癌23例25个病灶,导管内癌3例4个病灶,浸润性小叶癌1例1个病灶;良性病变28例,42个病灶,其中纤维腺瘤11例21个病灶,导管内乳头状瘤1例1个病灶,囊肿2例4个病灶,炎症5例6个病灶,乳腺腺病7例9个病灶,男性乳房发育1例。31.9%病灶（23/72）在磁共振平扫不能被检出,结合增强后的形态学特征对乳腺癌的检出的敏感性为86.7%,特异性为66.7%,准确性为75.0%。66.7%良性病灶（28/42）呈类圆形或分叶状,边缘光整边界清晰,强化均匀:86.7%恶性病灶（26/30）为不规则形或有毛刺,边界不清,强化不均匀。DWI显示97.6%良性病变（41/42）及所有恶性病灶。在b=500和1000s/mm²时,恶性病变组平均ADC值分别为1.257±0.279×10^-3 mm²/s、1.021±0.192×10^-3 mm²/s;良性病变组平均ADC值分别是1.549±0.341×10^-3mm²/s、1.397±0.216×10^-3 mm²/s;正常组织组分别是1.974±0.492×10^-3 mm²/s、1.764±0.432×10^-3 mm²/s。恶性-良性组、恶性-正常组和良性-正常组之间的平均ADC值差异均有显著性意义（F=50.998,P＜0.001;F=84.646,P＜0.001）。b=500 s/mm²时,以1.5×10^-3mm²/S为界,ADC值诊断的敏感度86.7%、特异度76.2%、准确度80.5%、阳性预测值72.2%和阴性预测值88.9%;b=1000 s/mm²时,以1.2×10^-3 mm²/s为界,ADC值诊断的敏感度80.0%、特异度85.7%、准确度83.3%、阳性预测值80.0%和阴性预测值85.7%。以早期增强率≥80%为恶性、＜80%为良性,良、恶性病变的早期增强率差异有显著性意义（x²=6.73,P＜0.01）,诊断的敏感性为100%,特异性为59.5%。以早期增强率≥100%作为恶性标准,敏感性为93.3%,特异性为73.8%。良、恶性病变强化峰值的差异没有显著性意义（P=0.18）。90.0%恶性病变（27/30）的达峰时间≤3min,14.3%良性病变（6/42）的达峰时间≤3min,良、恶性病变达峰时间的差异有显著性意义（P＜0.001）。46.7%恶性病灶（14/30）的时间-信号强度曲线为Ⅲ型,50.0%恶性病灶（15/30）为Ⅱ型,66.7%良性病灶（28/42）时间-信号强度曲线为Ⅰ型,16.7%良性病灶（7/42）为Ⅱ型,4.7%良性病灶（2/42）为Ⅲ型,11.9%良性病灶（5/42）为Ⅳ型,良、恶性病变的时间-信号强度曲线类型分布的差异有显著性意义（x²=11.84,P＜0.01）。时间-信号强度曲线取Ⅱ、Ⅲ型曲线为恶性病变,Ⅰ型、Ⅳ型曲线为良性病变,诊断的敏感性为96.7%,特异性为78.6%,准确性为86.1%。结论:（1）常规MRI平扫对乳腺良恶性疾病的诊断价值有限。（2）在b=500s/mm²和1000s/mm²时乳腺癌、良性病变和正常乳腺组织的ADC值不同,通过ADC值的测定有助于鉴别乳腺良恶性病变。（3）乳腺癌、良性肿瘤、正常组织的平均ADC值呈逐步递增趋势,在b=500s/mm²和1000s/mm²时设定ADC值为1.5×10^-3mm²/s和1.2×10^-3mm²/s,对乳腺癌检出有较高的敏感性、特异性、准确性、阳性预测值和阴性预测值。（4）DCE-MRI有助于乳腺良、恶性病变的诊断和鉴别诊断。（5）DCE-MRI联合DWI能进一步提高诊断的特异性。更多还原

【Abstract】 Background and purpose:The incidence of breast cancer is located at the top of the female tumor in the global.Every nine women,there is one person in a lifetime risk of breast cancer in the western developed countries.China was a low incidence of breast cancer national,but in recent years,the incidence of breast cancer showed an upward trend year after year, and the trend of younger,with changes in the reproductive patterns and lifestyle.The breast cancer seriously endangers the health and lives of women.Early diagnosis, early treatment is an important measure to reduce breast cancer mortality and improve 5-year survival rate of breast cancer patients.Imaging has a great value for early diagnosis,clinical stage and treatment of breast cancer.The purpose of this study was to evaluate the value of MRI rout-scanning, dynamic contract-enhanced MRI and diffusion-weighted imaging in distinguishing between benign and malignant breast lesions.Materials and methods:We analyzed seventy-two lesions from fifty-five patients in the first affiliated hospital of Zhengzhou University from May 2008 to February 2009,who were operated acquired with histopathological demonstration,of which thirty were malignant and forty-two were benign.53 females and 2 males,with the mean age was 44.5±10.9 years（range 21～67 years）.The main clinical features were breast masses. Using Siemens Magnetom Trio Tim 3.0T magnetic resonance device and a proprietary phase array breast surface coil to check the bilateral and whole breast at the same time. All lesions were examined with MRI plain scan,diffusion-weighted imaging and dynamic contract-enhanced MRI.The sequences of MRI plain scan used were three-dimensional fast low angle shot spoiled gradient-echo sequence（FLASH-3D） T₁WI and Turbo inversion recovery（TIR） T₂WI.Diffusion-weighted imaging was done with the single shot SE-EPI sequence,b=0s/mm²,500s/mm²,1000s/mm²;slice thickness:4mm.Dynamic contrast-enhanced imaging was done with FLASH-3D, with fat-suppression,axial T₁WI.The sequence parameters were as follow:TR/TE: 4.23ms/1.57ms;slice thickness:0.9mm.The same sequence was used pre-contrast one time and post-contrast six times with no gap between each post-contrast sequence. The overall acquisition time was 7min and 23sec.There was 20sec interval between pre-contrast and post-contrast sequence.An intravenous bolus injection of gadolinium diethylene triaminepenta acetate（Gd-DTPA） was give at a standard single dose of 0.1mmol/kg and at a rate of 2.5ml/s.This was followed by a flush of 25ml of saline solution at the same rate.Injection of contrast agent and saline flush was performed using an automatic injector.The data of DWI and dynamic contrast-enhanced MRI were transfer to Leonardo work station.We measured apparent diffusion coefficient （ADC） map and time-signal intensity curve was gain by the ADC and Mean Curve software.The lesions were classified as malignant or benign according the ADC value, the type of time-signal intensity curve（TIC）,early-phase enhancement rate,peak value,peak time,morphology features and indirect features.We calculated the enhancement rate 1 minute after ejection of Gd-DTPA（ΔSI）.The validities of dynamic contrast-enhanced imaging,ADC,and combination of the two methods were evaluated.Select the slice that shows the largest diameter of the lesion which with the maximal enhancement,exclude any hemorrhage,necrotic,cystic or calcification.The same size region of interest（ROI） was selected on the other breast in the same slice. The images were reviewed by two radiologists at workstation. Results:72 lesions of 55 cases that had been proved by pathology include 30 malignant lesions and 42 benign lesions.25 lesions（23 cases,include 1 male breast cancer） were invasive ductal carcinoma,1 lesion was invasive lobular carcinoma,4 lesions（3 cases） were ductal carcinoma in situ.42 benign lesions（28 cases） include 21 fibroadenomas, 1 breast papilloma,4 cysts,6 chronic inflammatory,9 adenosis of breast and 1 breast proliferation of man.31.9 per cent lesions（23/49） can not be detected by MRI plain scanning,while combining of morphology features,the sensitivity is 86.7%,the specificity 66.7% and the accuracy is 75.0%for detection of breast cancer.28 benign lesions were round or lobulated,well-defined,homogenous enhancement.26 malignant lesions were irregular shape or burr,obscure border,inhomogeneous enhancement.DWI showed that 97.6%of benign lesions（41/42） and all malignant lesions.The average ADC values of malignant lesions were 1.257±0.279×10^-3 mm²/s and 1.021±0.192×10^-3 mm²/s.The average ADC value of benign lesions were 1.549±0.341×10^-3 mm²/s、1.397±0.216×10^-3 mm²/s.The average ADC value of normal tissue were 1.974±0.492×10^-3 mm²/s and 1.764±0.432×10^-3 mm²/s.There were statistic significances between benign and malignant lesions,benign lesions and normal tissues,malignant lesions and normal tissues（F=50.998,P＜0.001; F=84.646,P＜0.001）.b=500 s/mm²,the threshold of ADC for diagnosis was 1.5×10^-3 mm²/s,the sensitivity of diagnosis is 86.7%,the specificity is 76.2%,the accuracy is 80.5%,72.2%positive predictive value and 88.9%negative predictive value.b=1000 s/mm²,the threshold of ADC for diagnosis was 1.2×10^-3 mm²/s,the sensitivity,specificity,accuracy of ADC were 80.0%,85.7%,83.3%,positive predictive value was 82.7%and negative predictive value was 85.7%.The difference of the early-phase enhancement rate between benign and malignant lesions was statistically significant（x²=6.73,P＜0.01）.The early-phase enhancement rate≥80%as malignant,diagnostic sensitivity was 100%and specificity was 59.5%.Early enhancement ratio≥100%as malignant,the sensitivity was 93.3%and specificity was 73.8%.The difference of the peak value between benign and malignant lesions had no statistical significance（P=0.18）.The peak time of 90.0 per cent malignant lesions were less than 3 minutes.The difference of the peak time between benign and malignant lesions was statistically significant（P＜0.001）.The distribution of curve types for malignant lesions was washout 46.7%（14/30）,biphasic 50.0%（15/30）,and monophasic 3.3%（1/30）.The distribution of curve types for malignant lesions was washout 4.7%（2/42）,biphasic 16.7%（7/42）,and monophasic 66.7%（28/42）.The difference of the type of time-signal intensity curve between benign and malignant lesions was statistically significant（x²=11.84,P＜0.01）.Diagnostic sensitivity was 96.7%,specificity was 78.6%and accuracy was 86.1%.Conclusion:The value of MRI plain scan in distinguishing between benign and malignant breast lesions is limited.The ADC value among breast cancer,benign lesions with normal breast tissue is different.The ADC values can be used to identify benign and malignant breast lesions.The average ADC values of tissues were increased in b= 500 s/mm² than b=1000 s/mm².b=500 s/mm² and b=1000 s/mm²,the threshold of ADC for diagnosis was 1.5×10^-3 mm²/s and 1.2×10^-3 mm²/s,there was a higher sensitivity,specificity,accuracy.Dynamic contrast-enhanced imaging has value to differentiate benign from malignant breast lesions.Combination of morphology, dynamic contrast-enhanced and DWI improve the diagnostic specificity,has great value to make a correct diagnose.更多还原