超微血管成像技术联合超声BI-RADS分级在鉴别乳腺良恶性肿物中应用价值

The combination of superb microvascular imaging and BI-RADS in distinguishing breast benign lesions from malignant lesions

目的:利用超微血管成像(SMI)技术,并与CDFI对比,以明确BI-RADS分级与SMI结合是否可以提高超声对恶性病变的诊断效能。方法:超声检出并术后病理证实的乳腺肿物138例,良性肿物78例,恶性肿物60例。按照Adler分级标准,分别观察同一肿物在CDFI和SMI技术所显示血流情况,并对肿物进行常规二维灰阶超声BI-RADS分级,使用受试者工作特征(ROC)曲线计算出BI-RADS,SMI及CDFI的Adler分级的鉴别良恶性乳腺病灶的最佳诊断界点,并根据此诊断界点得出以上单一模式及联合模式的诊断效能等相关指标。结果:SMI与CDFI的内部观察者间一致性较好(CDFI:κ=0.82;SMI:κ=0.76),两种检测方法在显示全部肿物血流丰富程度有明显差异(κ=0.38);在恶性肿物中上述差异更加明显(κ=0.028);而在良性肿物中差异不明显(κ=0.61)。分别得到的ROC曲线下面积分别为0.867(BI-RADS),0.669(CDFI),0.871(SMI),0.710(BI-RADS与CDFI联合模式)及0.899(BI-RADS与SMI联合模式)。其敏感性、特异性分别为0.950,0.692(BI-RADS);0.900,0.821(SMI);0.600,0.641(CDFI)。BI-RADS与SMI的联合模式的敏感性与BI-RADS单一模式相近,差异无统计学意义(0.900 vs 0.950,P>0.05);而BI-RADS与SMI的联合模式的特异性明显高于BI-RADS单一模式(0.897 vs 0.692,P<0.01)。结论 :与CDFI相比较,SMI检测肿物的微血管效果更好,SMI与BI-RADS联合可提高超声对乳腺恶性肿物的诊断效能。

Objective: To explicate the diagnostic performance of the combination of superb microvascular imaging(SMI)and BI-RADS in distinguishing breast benign lesions from malignant lesions. Methods: One hundred and thirty-eight breast lesions were detected by ultrasonography and pathologic types were confirmed by histopathology. Both SMI and CDFI were used to detect blood flows of breast masses according to Adler's grading criteria, respectively. The optimal threshold drawing from ROC curve and diagnostic efficacy of single and combinative modality were calculated. Results: The inter-observer agreement of SMI and CDFI were favorable(CDFI: κ=0.82; SMI: κ=0.76). There was a significant difference between SMI and CDFI for detecting blood flows in all lesions(κ=0.38), which was more significant in malignant group(κ=0.028) but not significant in benign lesions(κ=0.61). The areas under ROC curves were 0.867(BI-RADS), 0.669(CDFI), 0.871(SMI), 0.710(BI-RADS+CDFI) and 0.899(BI-RADS+SMI) respectively. The sensitivities and specificities were 0.950 and 0.692(BI-RADS), 0.821 and0.900(SMI), 0.600 and 0.821(CDFI), respectively. There was no significant difference between combinative modality(BI-RADS and SMI) and SMI(0.900 vs 0.950, P>0.05) for the sensitivity; but the specficity of combinative modality(BI-RADS and SMI)was significantly higher than that of SMI(0.897 vs 0.692, P<0.01). Conclusion: Comparing with CDFI, SMI provides better microvascular information in breast lesions. The combination modality of BI-RADS and SMI can improve diagnostic performance of breast malignancy.