Department of Radiology, Faculty of Medicine, Gazi University, 06500, Beşevler, Ankara, Turkey.
Department of Radiology, Faculty of Medicine, Gazi University, Yenimahalle, Ankara, 06500, Turkey.
Ir J Med Sci. 2022 Aug;191(4):1891-1897. doi: 10.1007/s11845-021-02744-7. Epub 2021 Sep 1.
Although several studies proved that SM could substitute for FFDM, the efficacy of SM in microcalcification evaluation remains controversial.
To investigate the diagnostic performance of synthetic mammography (SM) in the evaluation of microcalcifications in comparison with full-field digital mammography (FFDM).
In this retrospective study, 76 mammograms of 76 patients who underwent FFDM and digital breast tomosynthesis (DBT) acquisitions concomitantly between 2018 and 2019 and whose final mammography interpretation revealed microcalcifications (28 malignant microcalcifications and 48 benign microcalcifications) were included. All mammograms were reviewed independently by three radiologists with different levels of breast imaging experience. Readers were blinded to patient outcomes and interpreted each case in two separate reading sessions (first FFDM, second SM + DBT), according to the BI-RADS lexicon. The area under the receiver operating characteristic (ROC) curve (AUC) was calculated using ROC analysis in all cases for FFDM and SM + DBT sessions. The readers also assigned conspicuity scores to mammograms. The interobserver agreement was calculated using intraclass correlation coefficients (ICC).
The overall AUCs for malignant microcalcifications were 0.80 (95% CI: 0.75-0.85) in FFDM and 0.85 (95% CI: 0.80-0.89) in SM, and no significant difference was found between the groups (p = 0.0603). The sensitivity of the readers increased slightly with experience. The ICC values of BI-RADS categorization between readers were 0.93 (95% CI: 0.90-0.95) and 0.94 (95% CI: 0.91-0.96) for FFDM and SM, respectively.
SM had similar diagnostic performance in the evaluation of breast microcalcifications in comparison with FFDM, regardless of reader experience levels.
尽管有几项研究证实了 SM 可以替代 FFDM，但 SM 在微钙化评估中的效果仍存在争议。
探讨合成乳腺摄影术（SM）在评估微钙化方面与全数字化乳腺摄影术（FFDM）的诊断性能。
本回顾性研究纳入了 2018 年至 2019 年间同时接受 FFDM 和数字乳腺断层合成术（DBT）采集的 76 例患者的 76 张乳腺 X 线照片，这些患者的最终乳腺 X 线摄影检查结果显示有微钙化（28 例恶性微钙化和 48 例良性微钙化）。所有乳腺 X 线照片均由 3 名具有不同乳腺成像经验水平的放射科医生独立进行回顾。在两次单独的阅读会议（第一次为 FFDM，第二次为 SM+DBT）中，读者对每个病例均进行了阅读，阅读时并不知道患者的结果，根据 BI-RADS 词汇进行解读。使用 ROC 分析计算所有病例的 FFDM 和 SM+DBT 会话的接收者操作特征（ROC）曲线下面积（AUC）。读者还为乳腺 X 线照片分配了明显性评分。使用组内相关系数（ICC）计算观察者间的一致性。
恶性微钙化的总体 AUC 在 FFDM 为 0.80（95%CI：0.75-0.85），在 SM 为 0.85（95%CI：0.80-0.89），两组之间无显著差异（p=0.0603）。读者的敏感度随着经验的增加略有提高。读者间 BI-RADS 分类的 ICC 值分别为 0.93（95%CI：0.90-0.95）和 0.94（95%CI：0.91-0.96），用于 FFDM 和 SM。
无论读者经验水平如何，SM 在评估乳腺微钙化方面与 FFDM 具有相似的诊断性能。
