摘要目的：探讨数字乳腺断层合成X线成像（DBT）结合合成2D图像（C-view）诊断乳腺良、恶性病变的价值。方法前瞻性纳入经临床或超声检查乳腺发现可疑病灶拟行穿刺活检或手术的93例患者，行全屏数字化乳腺X线成像（FFDM）和DBT检查，DBT检查中通过重建获得C-view图像。分别对FFDM、DBT结合C-view、DBT结合FFDM图像进行分析，记录腺体平均辐射剂量，并对病灶可见性进行评级。对穿刺及手术标本进行病理分析，获得病变的组织类型。采用方差分析比较3种类型图像的平均腺体剂量，采用配对秩和检验比较不同类型图像间显示评级的差异。对3种类型图像上的病灶进行乳腺影像报告和数据系统（BI-RADS）分类，并以病理结果为金标准，采用ROC曲线分析3种类型图像的诊断效能。结果93例均为单发病灶。诊断16例经穿刺病理证实，77例经手术病理证实。良性病灶27例，恶性病灶66例。3种类型图像间的单个体位腺体平均辐射剂量差异有统计学意义（P均<0.01），其中FFDM、DBT结合C-view比DBT结合FFDM降低50.0%（1.95/3.90）。DBT结合C-view、与DBT结合FFDM图像对病灶的可见性结果完全相同，优于FFDM图像，差异有统计学意义（P<0.01）。54例患者（58.1%，54/93）对乳腺病灶的可见性提高。采用DBT+C-view图像鉴别诊断乳腺良、恶性病变的敏感度和特异度均最高（分别为90.9%和85.2%）。FFDM、DBT结合C-view、DBT结合FFDM诊断乳腺良、恶性病变的ROC曲线下面积分别为0.756、0.880、0.862。结论 DBT结合C-view较FFDM检查能够提高乳腺病灶的显示及诊断效能，且无需增加辐射剂量。

abstractsObjective To investigate the diagnostic value of digital breast tomosynthesis (DBT) plus synthesized two-dimensional(2D) images (C-view) in benign and malignant breast disease. Methods The prospective study enrolled 93 women with a clinically suspicious breast lesion confirmed by pathology. All patients underwent full-field digital mammography (FFDM) and DBT examinations before surgery or biopsy. Synthesized 2D images (C-view) were reconstructed in DBT. The images in FFDM alone, DBT plus C-view, and DBT plus FFDM were observed, the average glandular dose was recorded and the visibility grade of the disease was evaluated. Pathological analysis was performed on biopsy and surgical specimen to acquire histopathologic type. The average glandular dose among three examinations was compared by analysis of variance, and the visibility grade was compared by Wilcoxon matched-pairs signed-ranks test. The manifestation of breast disease in different examinations was evaluated according to breast imaging reporting and data system (BI-RADS). Taking pathologic results as the gold standard, receiving operating characteristic(ROC)was performed to analyze the diagnostic efficiency of three examinations. Result All lesions in 93 cases were solitary, in which 27 were benign and 66 were malignant. Sixteen cases were pathologically confirmed by biopsy and 77 cases by surgery. For the average glandular dose, there was significant difference among three examinations (P<0.01); it decreased by 50.0%(1.95/3.90)in FFDM alone and DBT plus C-view than DBT plus FFDM. All cases showed the same visibility between DBT plus C-view and DBT plus FFDM, in which 54 cases (58.1%,54/93) showed increased visibility than FFDM (P<0.01). The sensitivity and specificity were both highest in DBT plus C-view (90.9% and 85.2% respectively) to differentiate benign and malignant breast disease. The area of ROC curve in FFDM alone, DBT plus C-view, and DBT plus FFDM were 0.756, 0.880 and 0.862 respectively. Conclusion The combination of synthesized 2D images and DBT can improve both detection of breast lesion and diagnostic efficiency compared with FFDM, without an increase in radiation dose to the breast.