常规超声联合S-Detect技术对乳腺病灶的诊断价值

目的旨在比较常规超声BI-RADS分类联合S-Detect技术前后的诊断价值。方法选择435例患者行常规超声及S-Detect技术检查并留图像,由低、中、高年资医师分别阅读图像并进行BI-RADS分类评估。通过两种不同的联合诊断方法对BI-RADS分类进行调整及重新评估。以病理结果作为金标准。通过卡方检验及绘制ROC曲线比较联合诊断前后差异。结果 435例患者共563个病灶(良性403个,恶性160个)。S-Detect诊断灵敏度显著高于低年资医师(P<0.001),与中年资、高年资医师比较均无显著差异。S-Detect特异度显著低于三名医师(P均<0.001)。S-Detect与低年资医师比较AUC无统计学差异,但显著低于中年资及高年资医师(P均<0.000 1)。高年资医师通过联合方法二较原分类诊断乳腺恶性病灶的AUC显著增大;中年资医师联合诊断前后AUC均无差异;低年资医师通过联合方法一较原分类诊断AUC显著性增大。结论 S-Detect技术具有较高的灵敏度,通过不同方式的应用调整BI-RADS分类,S-Detect可以提高低年资及高年资医师的诊断水平。

Diagnostic Value of Conventional Ultrasound Combined with S-Detect in Breast Lesions

Objective The purpose of this study was to compare the diagnostic value of BI-RADS category by conventional ultrasound combined without and with S-Detect.Methods 435 patients for breast surgery were enrolled,and examed by conventional ultrasound and S-Detect.Ultrasound images were reviewed and analyzed by three ultrasound physicians and performed BI-RADS category.The BI-RADS category results were adjusted and re-evaluated by two different combining methods.The pathological results were taken as gold standard.The difference before and after the combining diagnosis was compared with the ROC curve and chi-square test.Results 435 patients had a total of 563 lesions(403 benign and 160 malignant).S-Detect diagnostic sensitivity is significantly higher than that of low yearexperience physicians(P<0.001).There was no significant difference between middle year-experience and high yearexperience physicians.Specificities of S-Detect were significantly lower than that of three physicians(every P<0.001).There was no significant difference between S-Detect and low-experience physicians compared with AUC,but it is significantly lower than middle year-experience and high year-experience physicians(both P<0.0001).The AUC of malignant breast lesions was significantly increased by the second combining method compared with the original classification.There was no difference in AUC before and after the combining diagnosis of middle year-experience physicians.AUC of low-experience physicians have significantly increased by the first combining method compared to the original classification.Conclusions S-Detect had high sensitivity.The BI-RADS category results of high year-experience physicians and low year-experience physicians were adjusted combing with S-Detect results,which can increase the original diagnostic level.