【作者】 顾雅佳； 冯晓源； 唐峰； 彭卫军； 毛健； 杨文涛；

【Author】 GU Ya-Jia Feng Xiao-Yuan Tang Feng Peng Wei-Jun Mao Jian Department of Radiology, the Affiliated Cancer Hospital of Fu Dan University, Department of Oncology, Shanghai Medical College, Fudan University, Shanghai 200032, China

【机构】 复旦大学附属肿瘤医院；

【摘要】 目的:探讨不同b值的磁共振弥散成像(DWI)对鉴别乳腺良恶性病变的价值。方法:测定83例95个乳腺病变在b分别取500、1000、2000s／mm2时的ADC值和信号强度, 以手术后病理结果为金标准,比较同一b值时不同乳腺病变、以及不同b值时同一病变平均ADC值和信号强度的差异。结果:1)乳腺病变分为乳腺癌、良性病变和囊肿。就同一种病变来说．平均ADC值、平均信号强度随b值的增大而降低。单向方差分析显示乳腺病变的平均ADC值、平均信号强度在不同b取值时变化趋势一致,差异显著,b=500s／mm2时前者 F=16．019,后者F=30．409,P均小于0．001;b=1000s／mm2 时F分别为16．578和23．092,P均小于0．001;b=2000s／mm2 时ADC比较F为3．122(P=0．049),信号强度比较F=23．888 (P<0．001)。各种乳腺病变两两比较显示在b=500s／mm2和 1000s／mm2时,大部分病变之间平均ADC值差异显著,而在 b=2000s／mm2时仅在乳腺癌和良性病变间有差异。2)以各 b值时乳腺癌平均ADC 95％可信区间的上限作为确定乳腺癌的上限域值点,则取这些域值点确定乳腺癌的敏感度、特异度、准确度、和阴性预测值在b=500s／mm2时都是最高的,并随b值的升高而降低。ROC曲线分析AZ500=0．775±0．046, AZ1000=0．780±0．044,P均小于0．001,提示b取500s／mm2 和1000s／mm2时对诊断乳腺癌有统计学意义。而AZ2000= 0．620±0．062,P=0．056,提示b=2000s／mm2对鉴别乳腺良恶性病变无统计性差异。3)各b值时平均信号强度和信躁比不同,在b为500s／mm2时最大,差异显著(F=65．878和 F=1．355,P均小于0．001),而噪声则与b取值无关(F=1．355, P=0．262)。结论:b取500s／mm2和1000s／mm2的磁共振弥散成像(DWI)可以用于鉴别乳腺良恶性病变。更多还原

【Abstract】 Objective: To evaluate the vary b value of diffusion-weighted (DW) MRI in distinguishing between benign and malignant breast lesions. Methods: Three diffusion-weighted sequences were implemented with 500, 1000 and 2000 s/mm2 b-value respectively on 83 patients 95 breast lesions. All lesions conflrmed by pathology. The apparent diffusion coefficient (ADC) values and signal intensity (SI) were recorded and compared between breast cancer, benign lesion, cyst and normal beast tissue in same b value DWI. The mean ADC value and mean SI of each type breast lesion in different b values were also analyzed. Results: 1) For same breast lesions, the mean ADC value and mean SI declined with the b values increased. Oneway ANOVA analysis appeared that the change trend of the different breast lesions mean ADC value and mean SI in different b values were consistent. The significant difference was found in mean ADC value of breast lesions at b values were 500,1000, and 2000 s/mm2. F=16.019 (P< 0.001), F=16.578, (P<0.001), and F=3.122 ( P=0.049 ) respectively. The significant difference was also indicated in mean SI at these b values. F=30.409 (P<0.001), F= 23.092 (P<0.001), and F=23.888 ( P<0.001) respectively. The difference in the mean ADC value was significant between tow type lesions (cancer and benign lesion, cancer and cyst, cancer and normal breast tissue) at b values were 500 and 1000 s/mm2. But at 2000 s/mm2, the significant difference was only seen between the cancer and the benign lesions. 2 ) The one-sided upper limit of 95% permissible interval of malignant ADCs was used as a upper threshold, the following diagnostic indices included sensitivity, specificity, accuracy and negative predictive value were higher at 500 s/mm2, and these indices reduced with the b value elevate. The areas under ROC curves in 500 s/mm2 and 1000 s/mm2 ADCs were AZ500 = 0.775 ± 0.046 (P<0.001) and AZ1000= 0.780 ± 0.044 (P<0.001), indicating significant difference of the mean ADC value among the breast lesions. But AZ2000 = 0.620 ± 0.062 (P= 0.056), which revealed no difference was indicated at 2000s/mm2. 3 ) The mean SI and signal noise ratio were significant difference between DWI of 500 s/mm2 and other’s b value DWI. The former was F=65.878 (P<0.001), and the latter F=1355 (P<0.001). The difference of the DWI noise was not significant among the different b values. Conclusion: DWIs, which b value was 500 s/mm2 and 1000 s/mm2, have the potential in diagnosis and differentiate the breast lesions.更多还原