Institute of Radiology, Department of Medicine, University of Udine, Azienda Sanitaria Universitaria Integrata di Udine, p.le S. Maria della Misericordia 15, 33100, Udine, Italy.
Eur Radiol. 2018 Mar;28(3):1000-1008. doi: 10.1007/s00330-017-5074-7. Epub 2017 Oct 10.
To compare automated breast volume scanner (ABVS), ultrasound (US) and MRI in measuring breast cancer size, and evaluate the agreement between ABVS and US in assessing lesion location and sonographic features.
We retrospectively included 98 women with 100 index cancers who had undergone US and ABVS followed by 1.5T MRI. Images were interpreted by a pool of readers reporting lesion size, location and breast imaging reporting and data system (BI-RADS) features. Bland-Altman analysis (with logarithmic data transformation), intraclass correlation coefficient (ICC) and Cohen's kappa statistic were used for statistical analysis.
MRI showed the best absolute agreement with histology in measuring cancer size (ICC 0.93), with LOA comparable to those of ABVS (0.63-1.99 vs. 0.52-1.73, respectively). Though ABVS and US had highly concordant measurements (ICC 0.95), ABVS showed better agreement with histology (LOA 0.52-1.73 vs. 0.45-1.86, respectively), corresponding to a higher ICC (0.85 vs. 0.75, respectively). Except for posterior features (k=0.39), the agreement between US and ABVS in attributing site and BI-RADS features ranged from substantial to almost perfect (k=0.68-0.85).
ABVS performs better than US and approaches MRI in predicting breast cancer size. ABVS performs comparably to US in sonographic assessment of lesions.
• ABVS approaches MRI in predicting breast cancer size. • ABVS is equivalent to US in localising and characterising breast cancer. • ABVS is more accurate than US in assessing breast cancer size. • ABVS has the potential to replace US in breast cancer staging.
比较自动乳腺全容积扫描（ABVS）、超声（US）和磁共振成像（MRI）在测量乳腺癌大小方面的差异，并评估 ABVS 和 US 在评估病变位置和超声特征方面的一致性。
我们回顾性纳入了 98 名 100 个病灶的女性患者，这些患者均接受了 US 和 ABVS 检查，随后还进行了 1.5T MRI 检查。由一组报告病变大小、位置和乳腺影像报告和数据系统（BI-RADS）特征的读者对图像进行解读。采用 Bland-Altman 分析（对数数据转换）、组内相关系数（ICC）和 Cohen's kappa 统计量进行统计学分析。
MRI 在测量癌症大小方面与组织学的一致性最好（ICC 0.93），其 LOA 与 ABVS 的 LOA 相当（分别为 0.63-1.99 和 0.52-1.73）。尽管 ABVS 和 US 的测量结果高度一致（ICC 0.95），但 ABVS 与组织学的一致性更好（LOA 0.52-1.73 与 0.45-1.86，分别），对应的 ICC 更高（0.85 与 0.75，分别）。除了后部特征（k=0.39）外，US 和 ABVS 在归因于病变位置和 BI-RADS 特征方面的一致性范围从高度一致到几乎完美（k=0.68-0.85）。
ABVS 在预测乳腺癌大小方面优于 US，与 MRI 相当。ABVS 在超声评估病变方面与 US 相当。
ABVS 在预测乳腺癌大小方面接近 MRI。
ABVS 在定位和描述乳腺癌方面与 US 相当。
ABVS 在评估乳腺癌大小方面比 US 更准确。
ABVS 有潜力替代 US 进行乳腺癌分期。
