Maastricht University Medical Center, Department of Radiology and Nuclear Medicine, P.O. Box 5800, 6202AZ Maastricht, the Netherlands; Zuyderland Medical Center, Department of Medical Imaging, P.O. Box 5500, 6130MB Sittard-Geleen, the Netherlands; GROW School for Oncology and Reproduction, P.O. Box 616, 6200MD Maastricht, the Netherlands.
Maastricht University Medical Center, Department of Radiology and Nuclear Medicine, P.O. Box 5800, 6202AZ Maastricht, the Netherlands; GROW School for Oncology and Reproduction, P.O. Box 616, 6200MD Maastricht, the Netherlands.
Eur J Radiol. 2023 Jul;164:110881. doi: 10.1016/j.ejrad.2023.110881. Epub 2023 May 13.
Breast MRI is considered the best modality for preoperative staging of invasive lobular carcinoma (ILC). However, contrast-enhanced mammography (CEM) shows comparable diagnostic performance to MRI, but evidence of CEM's accuracy in women diagnosed with ILC is scant. We aimed to retrospectively evaluate CEM and MRI accuracy in preoperative staging of ILC.
ILC cases diagnosed between 2013 and 2021 were collected. For both modalities, tumour diameter was extracted from the reports. Bland-Altman plots were used to assess discrepancies between size measurements according to imaging and histopathological findings. CEM and MRI's ability to detect multifocal/contralateral cancer was expressed as sensitivity, specificity, and diagnostic odds ratios (DORs). Pairwise comparison of women undergoing both CEM and MRI was not performed.
305 ILC-cases fulfilled preset inclusion criteria. Mean age was 63.7 years. Preoperative staging was performed using MRI or CEM in 266 (87.2%) and 77 (25.2%) cases, respectively. MRI and CEM overestimated tumour size by 1.5 and 2.1 mm, respectively. Sensitivity to detect multifocal disease was higher for MRI than for CEM (86% versus 78%), but specificity was lower for MRI (79% versus 92%). For detection of contralateral breast cancer, sensitivity for MRI was 96% versus 88% for CEM, and specificity was 92% and 99%, respectively. For both indications, DOR was higher for CEM, but differences were non-significant (p = 0.56 and p = 0.78).
CEM and MRI overestimate ILC size with comparable systematic and random errors. MRI's higher sensitivity for detection of multifocal/contralateral cancers is accompanied by lower specificity, but discriminative ability for both modalities was non-significant.
乳腺磁共振成像（MRI）被认为是浸润性小叶癌（ILC）术前分期的最佳方式。然而，对比增强乳腺摄影（CEM）与 MRI 具有相当的诊断性能，但 CEM 在诊断为 ILC 的女性中的准确性证据很少。我们旨在回顾性评估 CEM 和 MRI 在 ILC 术前分期中的准确性。
收集了 2013 年至 2021 年期间诊断的 ILC 病例。对于两种方式，均从报告中提取肿瘤直径。Bland-Altman 图用于评估根据影像学和组织病理学发现的大小测量值之间的差异。CEM 和 MRI 检测多灶/对侧癌症的能力用灵敏度、特异性和诊断优势比（DOR）表示。未对同时接受 CEM 和 MRI 的女性进行两两比较。
305 例 ILC 病例符合预设纳入标准。平均年龄为 63.7 岁。术前分期分别在 266 例（87.2%）和 77 例（25.2%）病例中使用 MRI 或 CEM 进行。MRI 和 CEM 分别高估肿瘤大小 1.5 和 2.1 毫米。MRI 检测多灶性疾病的灵敏度高于 CEM（86%比 78%），但特异性低于 CEM（79%比 92%）。对于检测对侧乳腺癌，MRI 的灵敏度为 96%，CEM 为 88%，特异性分别为 92%和 99%。对于这两种情况，CEM 的 DOR 更高，但差异无统计学意义（p=0.56 和 p=0.78）。
CEM 和 MRI 高估了 ILC 的大小，具有相当的系统和随机误差。MRI 检测多灶性/对侧癌症的灵敏度较高，但特异性较低，但两种方式的鉴别能力无显著差异。
