平扫T2脂肪抑制序列图像纹理可提高诊断乳腺良恶性结节的准确率

目的探讨平扫T2脂肪抑制序列（FS T2WI）图像纹理分析鉴别诊断乳腺良恶性结节的价值。方法回顾性分析经手术病理证实的60例患者共61个乳腺结节FS T2WI图像。绘制纹理参数鉴别诊断良恶性乳腺结节的ROC曲线，并与病理结果进行对照。结果60例患者的61个乳腺结节，FS T2WI纹理参数灰度区域矩阵重点运行高灰度级判断乳腺结节良恶性的AUC值（0.701）最大且诊断准确率高，其诊断恶性乳腺结节的敏感度为65.52%（19/29），特异度为71.88%(23/32)，误判率为31.15%（19/61）。FS T2WI图像诊断乳腺恶性结节敏感度为71.43%（20/28）、特异度为63.64%（21/33）、误判率为32.79%（20/61）；两者联合应用敏感度为64.29%(18/28)，特异度78.79%（26/33），误判率为27.87%（15/61）,与单独T2比较，差异有统计学意义（χ2=72.255，P=0.000）。结论平扫FS T2WI序列结合纹理分析能够提高诊断乳腺良恶性结节的特异度，降低误判率，提高乳腺平扫诊断的准确度。 

Value of T2 fat inhibition sequence image texture analysis in diagnosis of benign and malignant breast nodules

ObjectiveTo investigate the value of image texture analysis of T2 Fat-suppression sequence (FS T2WI) in differential diagnosis of benign and malignant breast nodules.MethodsA total of 61 images of breast nodules FS T2WI were retrospectively analyzed in 60 patients with surgical pathology. The ROC curves for differential diagnosis of benign and malignant breast nodules were drawn and compared with the pathological results.Results61 breast nodules in 60 patients, The FS T2WI texture parameter gray region matrix focuses on the operation of high gray level to determine the AUC value (0.701) of benign and malignant breast nodules with high diagnostic accuracy. The sensitivity of the diagnosis of malignant breast nodules was 65.52 % (19/29), and the specificity was 71.88 % (23/32), and the misjudgment rate was 31.15 % (19/61). The diagnostic sensitivity of FS T2WI images to malignant nodules in the breast was 71.43 %(20/28), and the specificity was 63.64 % (21/33), and the miscalculation rate was 32.79 %(20/61); The combined application sensitivity of the two is 64.29 %(18/28), the specificity is 78.79 % (26/33), and the miscalculation rate is 27.87 %(15/61). Compared with individual T2, the difference is statistically significant (χ2=72.255, P=0.000).ConclusionFS T2WI sequence combined with texture analysis can improve the specificity of diagnosis of benign and malignant breast nodules, which can reduce the rate of misjudgment, and improve the accuracy of diagnosis.