乳腺病灶不同强化形态及大小的MR扩散加权成像研究和参数选择

MR diffusion-weighted imaging in differential diagnosis of breast lesions with different enhancement shapes or size and parameter selection

目的 评价ADC值在不同强化形态及不同肿块大小的乳腺病灶中的诊断价值.方法 136个经手术病理证实的乳腺病灶,根据其不同的强化形态及大小分为3组,分别为非肿块样强化组（G1）、最大径≤2.0 cm的肿块样强化组（G2a）、最大径＞2.0 cm的肿块样强化组（G2b）.采用单激发EPI序列,3个扩散敏感梯度,b值分别为0、800、1000 s/mm2.两样本比较t检验分析各组内恶性与非恶性病灶平均ADC值的差异有无统计学意义,并绘制ROC曲线检验诊断效能.计算不同阈值下,ADC值诊断的敏感度、特异度、阴性预测值、阳性预测值、诊断符合率,并与形态学评价相结合,确定合适的b值和阈值.结果 G1组恶性与非恶性病灶的平均ADC值的差异无统计学意义[恶性病灶与非恶性病灶b=800 mm2/s时,平均ADC值分别为（1.13±0.23）×10-3和（1.28±0.27）×10-3mm2/s,t=1.636,P=0.112;b=1000 mm2/s时,平均ADC值分别为（1. 05±0.20）×10-3和（1.20±0.23）×10-3mm2/s,t=1.720,P=0.109];G2a组恶性与非恶性病灶平均ADC值的差异有统计学意义[恶性病灶与非恶性病灶b=800 mm2/s时,平均ADC值分别为（1.07±0.15）×10-3和（1.37±0.37）×10-3mm2/s,t=4.803,P=0.000;b=1000 mm2/s时,平均ADC值分别为（0.99±0.14）×10-3和（1.30±0.34）×10-3mm2/s,t=5.235,P=0.000];G2b组恶性与非恶性病灶平均ADC值的差异有统计学意义[恶性病灶与非恶性病灶b=800 mm2/s时,平均ADC值分别为（0.97±0.14）×10-3和（1.40±0.39）×10-3mm2/s,t=4.227,P=0.000;b=1000 mm2/s时,恶性病灶与非恶性病灶的平均ADC值分别为（0.93±0.14）×10-3和（1.35±0.36）×10-3mm2/s,t=4.329,P=0.000].b选取800或1000 s/mm2时,ADC值在肿块样强化组中的诊断效能相同（x2=0.36,P=0.5460）.当b值取1000 s/mm2,阈值取1.25×10-3s/mm2时,ADC值诊断乳腺恶性病灶的敏感度和阴性预测值最高,分别为97.7%和97.1%.结论 ADC值对于肿块样强化的乳腺病灶具有诊断价值,但不适用于非肿块样强化灶的诊断.

Objective To investigate the diagnostic value of ADC for breast lesions with different enhancement shape or mass size. Methods One hundred and thirty-six breast lesions confirmed by histopathology were included in this study. According to enhancement shape and size of the lesion, all lesions were divided into 3 groups： non-masslike enhancement （ G1 ）, masslike enhancement with the largest diameter 〈 2. 0 cm （G2a） and masslike enhancement with the largest diameter 〉 2. 0 cm （G2b）. Echo planar imaging DWI was performed and three b-values （0,500 and 1000 s/mm2） were applied. The t-test was used for testing the difference of ADC between malignant and non-malignant breast lesions in each group. ROC curve was deduced to test the diagnostic efficiency of ADC. The sensitivity, specificity, negative predictive value（ NPV）, positive predictive value（PPV） and accuracy of ADC for the diagnosis of breast lesions were calculated under the different threshold. Appropriate b value and threshold were determined with the combination of morphologic evaluation. Results There were no significant differences for the mean ADC values between malignant [b =800 mm2/s： ADC value = （ 1.13 ±0. 23） × 10-3 mm2/s,b=1000 mm2/s： ADC value = （1.05 ±0.20） × 10-3 mm2/s] and non-malignant breast lesions [b =800 mm2/s： ADC value = （ 1.28 ±0. 27） × 10-3 mm2/s, t = 1. 636, P =0. 112,b = 1000 mm2/s： ADC value=（1.20 ±0.23） × 10-3 mm2/s, t = 1.720, P =0. 109] in Group 1. The mean ADC values of malignant breast lesion [b =800 mm2/s： ADC value = （1.07 ±0. 15） × 10-3 mm2/s,b = 1000 mm2/s：ADC value = （0. 99 ±0. 14） × 10-3 mm2/s] were significantly lower than that of non-malignant lesion [b =800 mm2/s： ADC value = （ 1.37 ± 0. 37 ） × 10-3 mm2/s, t = 4. 803, P = 0. 000; b = 1000 mm2/s： ADC value= （1.30 ±0.34） × 10-3 mm2/s, t =4.227, P =0.000] in Group 2a. The mean ADC values of malignant breast lesion [b =800 mm2/s： ADC value = （0. 97 ±0. 14） × 10-3 mm2/s; b = 1000 mm2/s：ADC value = （0. 93 ±0. 14） × 10-3 mm2/s] were significantly lower than that of non-malignant lesion [b =800 mm2/s： ADC value = （ 1.40 ± 0. 39） × 10 -3 mm2/s, t = 4. 227, P = 0. 000; b = 1000 mm2/s： ADC value = （ 1.35 ±0. 36） × 10-3 mm2/s, t =4. 329, P =0. 000] in Group 2b. The diagnostic efficiency was equal（ x2 =0. 36,P =0. 5460） whenever b value of 800 or 1000 s/mm2 was selected. The highest sensitivity （97.7%） and NPV （97. 1%） were obtained with b value of 1000 s/mm2 and threshold of 1.25 ×10 -3 s/mm2. Conclusion MR DWI is useful for the differential diagnosis of breast lesions with masslike enhancement rather than nonmasslike enhancement.