中国癌症杂志 ›› 2021, Vol. 31 ›› Issue (10): 912-919.doi: 10.19401/j.cnki.1007-3639.2021.10.006

• 论著 • 上一篇    下一篇

双模式数字乳腺体层合成成像对乳腺疾病的诊断价值

姜婷婷,汤 伟,尤 超,肖 勤,沈茜刚,彭卫军,顾雅佳   

  1. 复旦大学附属肿瘤医院放射诊断科,复旦大学上海医学院肿瘤学系,上海 200032
  • 出版日期:2021-10-30 发布日期:2021-11-10
  • 通信作者: 顾雅佳 E-mail: cjr.guyajia@vip.163.com
  • 基金资助:
    上海市2019年度“科技创新行动计划”产学研医合作领域项目子课题(19DZ1930502);上海市抗癌协会“翱翔”计划(SACA-AX-201903);国家自然科学基金(NSFC 82071878);上海2021年度“科技创新行动”计划医学创新研究专项(21Y11910200)。

Diagnostic value of dual mode digital breast tomosynthesis in breast diseases

JIANG Tingting, TANG Wei, YOU Chao, XIAO Qin, SHEN Xigang, PENG Weijun, GU Yajia   

  1. Department of Radiology, Fudan University Shanghai Cancer Center, Department of Oncology, Shanghai Medical College, Fudan University, Shanghai 200032, China
  • Published:2021-10-30 Online:2021-11-10
  • Contact: GU Yajia E-mail: cjr.guyajia@vip.163.com

摘要: 背景与目的:全数字化乳腺X线摄影(full-field digital mammography,FFDM)可显著降低乳腺癌的死亡率,但作为一种二维成像方式,对于致密型乳腺构成分类患者,其诊断灵敏度和特异度明显较低。探讨数字乳腺体层合成成像(digital breast tomosynthesis,DBT)结合FFDM对乳腺病变的鉴别诊断价值以及高分辨率扫描(high-resolution,HR)模式(扫描角度为40°,DBT-HR)和标准扫描(standard,ST)模式(扫描角度为15°,DBT-ST)诊断乳腺病变的价值。方法:前瞻性收集2016年7月—9月经临床或超声检查怀疑有乳腺异常病变,且经穿刺活检或术后病理学检查证实的175例女性乳腺疾病患者,行FFDM和DBT检查,随机进入DBT-HR组和DBT-ST组,对FFDM图像和DBT结合FFDM图像进行分析。参照2013版《乳腺影像报告和数据系统》(Breast Imaging Reporting and Data System,BI-RADS)分类标准,以病理学检查结果为金标准,采用受试者工作特征(receiver operating characteristic,ROC)曲线分析FFDM和DBT结合FFDM的诊断效能,根据不同的乳腺构成分类分别分析DBT-HR和DBT-ST的灵敏度和特异度。结果:175例患者共检出181个病灶,良性病灶50个,恶性病灶131个,所有入组病例的FFDM和FFDM结合DBT的灵敏度为81.92%和91.15%,特异度为77.45%和87.25%,两种方法的ROC曲线的曲线下面积(area under curve,AUC)差异有统计学意义(0.88 vs 0.94,Z值=5.37,P<0.01)。随机分入DBT-HR组病灶86个而DBT-ST组病灶95个。DBT-HR组FFDM和DBT主要征象表现一致率为83.14%,仅DBT可见率为3.49%,DBT-ST组FFDM和DBT主要征象表现一致率为92.63%,仅DBT可见率为0.53%。致密型乳房在结合DBT后灵敏度和特异度提高均较非致密型乳房明显,尤其是HR组致密型乳房灵敏度(18.72%)和ST组的致密型乳房的特异度(14.28%)均有所提高。结论:FFDM结合DBT的诊断效能优于FFDM,且DBT-HR优于DBT-ST,尤其是对致密型乳房的诊断。

关键词: 乳腺癌, 数字乳腺体层合成成像, 高分辨率扫描模式, 标准扫描模式

Abstract: Background and purpose: The use of full-field digital mammography (FFDM) can significantly reduce the mortality rate of breast cancer. However, it is a two-dimensional imaging method, the sensitivity and specificity of diagnosis are significantly reduced in patients with dense breast. This study compared the detectability of FFDM alone versus FFDM plus digital breast tomosynthesis (DBT), and the detectability of scan angle of 40° (HR-mode) DBT versus scan angle of 15° (ST-mode) DBT for the diagnostic evaluation of breast lesion. Methods: The prospective study enrolled 175 women with clinically suspected breast lesion from July 2016 to September 2016 in Fudan University Shanghai Cancer Center. One hundred and seventy-five patients confirmed by biopsy or surgery were included. All the patients underwent full-field digital mammography (FFDM) and DBT examinations, and patients were randomly allocated into DBT-ST group or DBT-HR group. The images of FFDM and FFDM plus DBT were analyzed. According to Breast Imaging Reporting and Data System (BI-RADS, version 2013), the receiver operating characteristic (ROC) curve was used to evaluate the diagnostic efficiency of FFDM and FFDM plus DBT, referring to the pathology as the gold standard. The sensitivities and specificities of DBT-HR and DBT-ST were analyzed according to different breast compositions. Results: A total of 181 lesions were detected in 145 patients, including 50 benign lesions and 131 malignant lesions. The sensitivities of FFDM and FFDM plus DBT were 81.92% and 91.15%, and the specificities were 77.45% and 87.25%, respectively. There was significant difference in the area under curve (AUC) between the two methods (0.88 vs 0.94, Z value was 5.37, P<0.01). There were 86 lesions randomly allocated into DBT-HR group and 95 lesions into DBT-ST group. For DBT-HR group, the consistency rate of finding between FFDM and DBT was 83.14%, and the only DBT detected rate was 3.49%. For DBT-ST group, the consistency rate of finding between FFDM and DBT was 92.63%, and the only DBT detected rate was 0.53%. Compared with the non-dense breast after combining the DBT, the improvement in sensitivity and specificity was more obvious in dense breast, especially the improvement in sensitivity of dense breast in DBT-HR (18.72%) and specificity of dense breast in DBT-ST (14.28%). Conclusion: The diagnostic efficiency of FFDM plus DBT is superior to that of FFDM, and the DBT-HR better than DBT-ST, especially in dense breast.

Key words: Breast cancer, Digital breast tomosynthesis, High-resolution-mode, Standard-mode