Discipline of Medical Imaging Sciences, Sydney School of Health Sciences, Faculty of Medicine and Health, University of Sydney, Level 7, Susan Wakil Health Building D18, Sydney, NSW 2006, Australia..
Discipline of Medical Imaging Sciences, Sydney School of Health Sciences, Faculty of Medicine and Health, University of Sydney, Level 7, Susan Wakil Health Building D18, Sydney, NSW 2006, Australia.
Acad Radiol. 2022 Aug;29(8):1228-1247. doi: 10.1016/j.acra.2021.09.025. Epub 2021 Nov 16.
Breast cancer is a highly complex heterogeneous disease. Current validated prognostic factors (e.g., histological grade, lymph node involvement, receptor status, and proliferation index), as well as multigene tests (e.g., Oncotype DX and PAM50) are helpful to describe breast cancer characteristics and predict the chance of recurrence risk and survival. Nevertheless, they are invasive and cannot capture a complete heterogeneity of the entire breast tumor resulting in up to 30% of patients being either over- or under-treated for breast cancer. Furthermore, multigene testings are time consuming and expensive. Radiomics is emerging as a reliable, accurate, non-invasive, and cost-effective approach of using quantitative image features to classify breast cancer characteristics and predict patient outcomes. Several recent radiomics reviews have been conducted in breast cancer, however, specific mammography-based radiomics studies have not been well discussed. This scoping review aims to assess and summarize the current evidence on the potential usefulness of mammography-based (i.e., digital mammography, digital breast tomosynthesis, and contrast-enhanced mammography) radiomics in predicting factors that describe breast cancer characteristics, recurrence, and survival.
PubMed database and eligible text reference were searched using relevant keywords to identify studies published between 2015 and December 19, 2020. Studies collected were screened and assessed based on the inclusion and exclusion criteria.
Eighteen eligible studies were included and organized into three main sections: radiomics predicting breast cancer characteristics, radiomics predicting breast cancer recurrence and survival, and radiomics integrating with clinical data. Majority of publications reported retrospective studies while three studies examined prospective cohorts. Encouraging results were reported, suggesting the potential clinical value of mammography-based radiomics. Further efforts are required to standardize radiomics approaches and catalogue reproducible and relevant mammographic radiomic features. The role of integrating radiomics with other information is discussed.
The potential role of mammography-based radiomics appears promising but more efforts are required to further evaluate its reliability as a routine clinical tool.
乳腺癌是一种高度复杂的异质性疾病。目前已验证的预后因素（如组织学分级、淋巴结浸润、受体状态和增殖指数）以及多基因检测（如 Oncotype DX 和 PAM50）有助于描述乳腺癌的特征，并预测复发风险和生存机会。然而，这些方法具有侵袭性，无法全面捕捉整个乳腺肿瘤的异质性，导致多达 30%的患者接受过度或不足的乳腺癌治疗。此外，多基因检测既耗时又昂贵。放射组学作为一种可靠、准确、无创且具有成本效益的方法，正在兴起，它使用定量图像特征来对乳腺癌特征进行分类并预测患者的结局。目前已经有多项关于乳腺癌的放射组学综述，但针对基于乳腺 X 线摄影的放射组学研究尚未得到很好的讨论。本综述旨在评估和总结基于乳腺 X 线摄影（即数字乳腺 X 线摄影、数字乳腺断层合成以及对比增强乳腺 X 线摄影）的放射组学在预测描述乳腺癌特征、复发和生存的因素方面的潜在作用的现有证据。
使用相关关键词在 PubMed 数据库和合格的文本参考文献中进行搜索，以确定 2015 年至 2020 年 12 月 19 日期间发表的研究。收集的研究经过筛选并根据纳入和排除标准进行评估。
纳入了 18 项符合条件的研究，并将其分为三个主要部分：放射组学预测乳腺癌特征、放射组学预测乳腺癌复发和生存以及放射组学与临床数据相结合。大多数出版物报道的是回顾性研究，而有 3 项研究是前瞻性队列研究。研究结果令人鼓舞，提示了基于乳腺 X 线摄影的放射组学具有潜在的临床价值。需要进一步努力来标准化放射组学方法并编目可重复和相关的乳腺 X 线摄影放射组学特征。还讨论了整合放射组学与其他信息的作用。
基于乳腺 X 线摄影的放射组学的潜在作用似乎很有前景，但需要进一步努力来进一步评估其作为常规临床工具的可靠性。
