MRI对乳腺X线摄影表现为局部结构扭曲病变的诊断价值

Value of MRI in differential diagnosis of primary architectural distortion detected by mammography

目的探讨MRI对乳腺x线摄影表现为局部结构扭曲病变的诊断价值。方法回顾性分析行全视野数字化乳腺x线摄影（FFDM）显示为局部结构扭曲病变，于FFDM检查后2周内行MRI检查，且经病理或随访2年证实的乳腺病变患者的临床及影像资料。60例（61个病变）纳入研究，恶性病变30个（30例），良性病变31个（30例）。患者均行双侧乳腺FFDM和MRI检查，MRI方法包括常规MRI、DWI和动态增强MRI（DCE—MRI）。根据乳腺影像报告和数据系统（BI．RADS）标准，对FFDM、DCE-MRI和DCE．MRI联合相对ADC值（nADC值）进行BI—RADS分类，BI—RADS1～3类为阴性病变，BI—RADS4～5类为阳性病变。测量病变ADC值和正常腺体ADC值，计算病灶ADC值与正常腺体ADC值的比值（nADC值），并采用两独立样本t检验进行比较。根据ROC曲线下面积确定鉴别乳腺良、恶性病灶的ADC和nADC阈值，计算利用ADC和nADC鉴别乳腺良、恶性病变的效能；恶性病变以病理结果为金标准，良性病变以病理或随访结果为金标准，计算不同影像检查方法（FFDM、DCE．MRI和DCE—MRI联合nADC值）诊断乳腺良、恶性病变的效能。结果良、恶性病变的ADC值分别为（1．35±0．31）×10-3mm2／s和（1．07±0．40）×10-3mm2／s，nADC值分别为0．83±0．17和0．59±0．25，差异均有统计学意义（t值分别为2．82、4．54，P均〈0．01）。ADC值及nADC值鉴别乳腺良、恶性病变的ROC曲线下面积分别为0．753和0．829。以ADC值为1．19×10-3mm2／s鉴别良、恶性病变，敏感度和特异度分别为71．0％和86．7％；以nADC为0．589鉴别良、恶性病变，敏感性和特异性分别为93．5％和76．7％。FFDM诊断乳腺良、恶性病变的敏感度、特异度和准确度分别为80．0％（24／30）、9．7％（3／31）和44．3％（27／61），DCE．MRI诊断的敏感度、特异度和准确度分别为90．0％（27／30）、41．9％（13／31）和65．6％（40／61），DCE—MRI联合nADC值诊断的敏感度、特异度和准确度分别为93．3％（28／30）、77．4％（24／31）和85．2％（52／61）。结论DCE-MRI联合nADC值对于FFDM检出的局部结构扭曲病变诊断的敏感度和特异度较高，具有较高的鉴别诊断价值。

Objective The aim was to evaluate the diagnostic value of MRI in lesions with architectural distortion manifested in mammography. Methods A retrospective analysis of MRI was performed in 60 patients with 61 lesions manifested as architectural distortion in full- field digital mammography （FFDM） and subsequently confirmed by pathology or two year＇s follow-up, 30 were malignant and 31 were benign. All the patients underwent MRI within 2 weeks of mammography. MRI protocol included conventional MR, DWI and dynamic contrast-enhanced MRI. The breast imaging reporting and data system （BI-RADS） was used as the reference standard. BI-RADS categories 1 to 3 were considered as negative for malignancy and BI-RADS categories 4 to 5 were considered as positive for malignancy. ADCs of suspicious lesion of interest and glandular tissue were calculated, nADC was then calculated using the following formula： nADC = ADC （lesion）/ADC （glandular tissue）. ADC and nADC were compared by using t test. ROC analysis was carried out to define the most effective threshold ADC and nADC value to differentiate malignant from benign lesion in the breast. Diagnostic performance of the FFDM, DCE-MRI and DCE-MRI combined nADC were calculated. Results ADC value of malignant and benign lesions was （1.35±0.31）×10-3 mm2/s and （1.070.40）×10-3 mm2/s, respectively, nADC values were 0.83±0.17 and 0.59± 0.25, respectively （t values were 2.82 and 4.54, P〈0.01）. Area under the curve of ADC and nADC were 0.829 and 0.753 respectively. When threshold of ADC was set at 1.19x 103mm2/s, sensitivity and specificity were 71.0% and 86.7%, respectively. For a nADC value threshold of 0.589, sensitivity and specificity were 93.5% and 76.7%, respectively. Sensitivity, specificity and accuracy with FFDM were 80.0%（24/30）, 9.7% （3/31） and 44.3%（27/61）, Sensitivity, specificity and accuracy with DCE-MRI were 90.0%（27/30）, 41.9% （13/31） and 65.6%（40/61）, Sensitivity, specificity and accuracy with DCE-MRI combined nADC were 93.3% （28/30）, 77.4%（24/31） and 85.2%（52/61）, respectively. Conclusion Sensitivity and specificity with DCE-MRI combined nADC is higher, and DCE-MRI combined nADC values is helpful to differentiate malignant from benign lesions with architectural distortion manifested in FFDM.