多参数MRI检查BI-RADS分类对乳腺疾病鉴别诊断价值

The value of multi parameter MRI in the differential diagnosis of breast diseases by BI-RADS classification

目的建立一种新的与BI-RADS分类相对应的多参数乳腺MRI诊断模式,探讨多参数乳腺MRI检查对乳腺良恶性疾病的鉴别诊断价值。方法回顾分析290例乳腺疾病患者,共309个经病理证实的乳腺病灶,根据其多平面重组、ADC值测量和TIC测量曲线数据进行BI-RADS分类并与病理结果进行对照分析,分别计算癌前病变纳入良、恶性组的灵敏度和特异度。结果本组病例中多参数MRI联合BI-RADS分类对乳腺疾病鉴别诊断的敏感度和特异度分别为:1)将癌前病变纳入良性病灶、在MRI将4类归为良性组为86.6%和85.0%;2)将癌前病变纳入恶性病灶、在MRI将4类归为良性组为93.3%和66.9%,高于其它组合。病理诊断为癌前病变的治疗建议为手术切除或随访,BI-RADS1-3类对随访病灶的阳性预测值(PPV)为88.2%,BI-RADS 5类对可手术切除病灶的PPV为93.3%,而将癌前病变纳入恶性病灶后,BI-RADS 4类对恶性的PPV为48.0%。结论建立乳腺疾病MRI多参数诊断模型能更精确地判断乳腺病灶良性、癌前病变和恶性特征,对鉴别乳腺良恶性疾病有良好的应用价值。

Objective To establish a new multi parameter MRI diagnostic model for breast cancer with BI-RADS classifica-tion, and to explore the value of multi parameter MRI in the differential diagnosis of benign and malignant breast diseases. Meth-ods 290 patients with breast disease, with 309 lesions, confirmed by pathology were retrospectively analyzed, BI-RADS classifi-cation was performed according to the data of multi planar reconstruction, ADC value and TIC curve, and the results were com-pared with pathological results. Then, we calculated the sensitivity and specificity of precancerous lesions into benign and malig-nant groups. Results In this study, the sensitivity and specificity of multi parameter magnetic resonance imaging combined with BI-RADS classification in differential diagnosis of breast diseases were：1 ） if the precancerous lesions included into benign le-sions, BI-RADS 4 in the MRI classified as benign group was 86. 6% and 85%;2） If the precancerous lesionsincluded in malig-nant lesions, BI-RADS 4 in the MRI classified as benign group were 93. 3% and 66. 9%, higher than other combinations. If the pathology treatment principle of precancerous lesions was defined as local excision or short-term follow-up, the BI-RADS 3 and the following categories of lesions the positive predictive value was 88. 2%, positive prediction value of resectable lesion in BI-RADS 5 was 93. 3%, and positive prediction value of malignant and precancerous lesions in BI-RADS 4 was 48%. Conclusion The MRI multi parameter diagnostic model established in this study can be used to determine the benign, high risk and malignant fea-tures of breast lesions.