3．0T磁共振扩散加权成像多b值在乳腺良恶病变诊断中的价值

目的：探讨3.0T磁共振扩散加权成像多b值在乳腺良恶性病变鉴别诊断中的应用价值。方法：分析乳腺良恶性病灶48个，测量病灶在不同b值（600、1000、1500s/mm2）下的ADC值并进行统计学分析。以病理诊断为金标准，采用ROC曲线获得曲线下面积（AUC）及最佳临界点，分析不同b值下ADC值诊断价值。结果：同一b值下乳腺良性、恶性病变的ADC值与不同b值下乳腺良性、恶性病变 ADC值间差异均有统计学意义。采用ROC曲线分析方法显示：b＝600s/mm2时，AUC为0.882，最佳临界点为1.43×10-3mm2/s，诊断的敏感性88.9%、特异性71.4%；b＝1000s/mm2时，AUC为0.897，最佳临界点为1.19×10-3mm2/s，诊断的敏感性为77.8%、特异性为81.0%；b＝1500s/mm2时，AUC为0.915，最佳临界点为1.04×10-3mm2/s，诊断的敏感性为70.4%、特异性为90.5%。结论：（1）ADC值定量分析有助于乳腺良、恶性病变的鉴别，同质病变ADC值随着b值的增大而变小。（2）以ROC曲线分析得到的最佳临界点作为诊断恶性病变的阈值，随着b值的增加，诊断的敏感性下降，特异性增加。

The clinical value of diffusion weighted imaging with multiple b value at 3T MR in diagnosis of breast diseases

Objective：To study the clinical value of diffusion weighted imaging with multiple b value at 3T MR in differentiation of benign and malignant breast diseases. Methods：48 benign and malignant breast lesions were used for as-sessment. The ADC of the lesions at different b values （600、1000、1500s/mm2） was measured, respectively for further eval-uation. Pathological findings were considered as the reference standard. ROC analysis was used to calculate the area under curve （AUC） as well as the best cut-off value for differentiation. The diagnostic performance of ADC at different b values was determined accordingly. Results： ADC measurements of benign breast lesions differed significantly from malignant le-sions both at the same and different b-values. With comparison to pathological results, the AUC was 0.882 and best cut-off was 1.43 ×10-3mm2/s when b value was 600s/mm2. The corresponding sensitivity and specificity for differentiation were 88.9% and 71.4%, respectively. The AUC was 0.897 and best cut-off was 1.19×10-3mm2/s when b value was 1000s/mm2. The corresponding sensitivity and specificity for differentiation were 77.8%and 81%, respectively. The AUC was 0.915 and best cut-off was 1.04×10-3mm2/s when b value was 1500s/mm2. The corresponding sensitivity and specificity for differentia-tion were 70.4% and 90.5%, respectively. Conclusion： ADC measurement is helpful for differentiation of benign and ma-lignant breast lesions. ADC decreases when b value increases. The best cut-off for lesion differentiation can be originated from ROC analysis. With the increase of b value, the sensitivity decreases and specificity increases.