Objective To evaluate the value of quantitative analysis of diffusion-weighted MRI in the diagno-sis of breast lesions. Methods 72 cases with pathologically conifrmed breast diseases (85 lesions) and 10 healthy volunteers underwent DWI. The diffusion-sensitive factor (b) values were 0, 500, 1,000 s/mm2. The ADC values of 40 normal breasts, 22 malignant lesions, 50 benign lesions and 13 inlfammatory lesions at different b values were measured and analyzed comparatively. Two methods were used to determine the ADC diagnosis threshold value of benign and malignant breast lesions, and the efifcacy of the two methods in the diagnosis of malignant breast lesions were analyzed comparatively. Results ①When b value was 500 s/mm2, the mean ADC values of the malignant lesions, inlfammatory lesions, benign lesions and normal glands were as follows, respectively, (1.10±0.13), (0.90±0.19), (1.68±0.30), (2.03±0.30) ×10-3mm2/s;When b value was 1,000 s/mm2, the mean ADC values of the malignant lesions, inlfammatory lesions, benign lesions and normal glands were as follows, respec-tively, (0.96±0.16), (0.68±0.20), (1.53±0.33), (1.85±0.31) ×10-3mm2/s.②If the average ADC value of breast ma-lignant lesions in the upper 95%conifdence interval was selected as the threshold value, when b value was 500 s/mm2, the ADC threshold was 1.16 ×10-3mm2/s;when b value was 1,000 s/mm2, the ADC threshold was 1.03 ×10-3mm2/s. If the ADC threshold was drawn by ROC curve, when b value was 500 s/mm2, the ADC threshold was 1.30 ×10-3mm2/s;when b value was 1,000 s/mm2, the ADC threshold was 1.14 ×10-3mm2/s. Conclusion Quantitative analysis of diffusion-weighted imaging is helpful in the diagnosis of breast benign lesions (inlfammatory lesions were not in-cluded) and malignant lesions. The ADC threshold determined by ROC curve has better accuracy, sensitivity and speciifcity in the diagnosis of breast malignant lesions when b value is 1,000 s/mm2.%目的：评价磁共振弥散加权成像（Diffusion-weighted Imaging，DWI）量化分析在乳腺病变诊断中的价值。方法对经病理证实的72例乳腺疾病患者（85个病灶）以及10名健康自愿者行DWI，扩散敏感系数（b）值分别为0、500、1,000 s/mm2。分别测量40个正常乳腺，22个恶性病灶，50个良性病灶，13个炎性病灶在不同b值时的ADC值，比较分析各组ADC值；采用两种方法确定乳腺良恶性病变ADC诊断阈值，并分析比较两者在乳腺恶性病变诊断中的效能。结果①b=500 s/mm2时，恶性病变、炎性病变、良性病变、正常腺体的平均ADC值分别为（1.10±0.13）、（0.90±0.19）、（1.68±0.30）、（2.03±0.30）×10-3mm2/s；b=1,000 s/mm2时，恶性病变、炎性病变、良性病变、正常腺体的平均ADC值分别为（0.96±0.16）、（0.68±0.20）、（1.53±0.33）、（1.85±0.31）×10-3mm2/s。②取乳腺恶性病变平均ADC值的95%可信区间上限作为阈值时，b=500 s/mm2时ADC阈值为1.16×10-3mm2/s；b=1,000 s/mm2时ADC阈值为1.03×10-3mm2/s。采用受试者工作特征（ROC）曲线确定ADC阈值，b=500 s/mm2时ADC阈值为1.30×10-3mm2/s；b=1,000 s/mm2时ADC阈值为1.14×10-3mm2/s。结论DWI量化分析对乳腺良（不包括炎性病变）、恶性病变的诊断有较高的价值；采用ROC曲线确定的b=1,000 s/mm2时的ADC阈值对诊断乳腺恶性病变的准确性、敏感性、特异性均较高。
展开▼
