A robust and efficient AI assistant for breast tumor segmentation from DCE-MRI via a spatial-temporal framework.
A robust and efficient AI assistant for breast tumor segmentation from DCE-MRI via a spatial-temporal framework.
复制标题
DOI：
10.1016/j.patter.2023.100826
复制
发表时间：
2023-09-08
期刊：
影响因子：
6.5
通讯作者：
Shen, Dinggang
中科院分区：
文献类型：
--
作者：
Zhang, Jiadong;Cui, Zhiming;Shi, Zhenwei;Jiang, Yingjia;Zhang, Zhiliang;Dai, Xiaoting;Yang, Zhenlu;Gu, Yuning;Zhou, Lei;Han, Chu;Huang, Xiaomei;Ke, Chenglu;Li, Suyun;Xu, Zeyan;Gao, Fei;Zhou, Luping;Wang, Rongpin;Liu, Jun;Zhang, Jiayin;Ding, Zhongxiang;Sun, Kun;Li, Zhenhui;Liu, Zaiyi;Shen, Dinggang
动态对比增强磁共振成像 (DCE-MRI) 可以高灵敏度地筛查、随访和诊断乳腺肿瘤。 DCE-MRI 准确的肿瘤分割可以提供肿瘤位置和形状的关键信息，这对下游临床决策产生重大影响。在本文中，我们的目标是开发一种人工智能（AI）助手，通过时空框架捕获多相 DCE-MRI 的动态变化来自动分割乳腺肿瘤。我们的人工智能助手的主要优势包括（1）稳健性，即我们的模型可以处理具有不同相数和成像间隔的MR数据，如来自七个医疗中心的大规模数据集所证明的那样，以及（2）效率，即我们的人工智能助手将手动注释所需的时间显着减少了20倍，同时保持了与医生相当的准确性。更重要的是，作为构建人工智能辅助乳腺癌诊断系统的基础步骤，我们的人工智能助手将推动人工智能在更多乳腺癌临床诊断中的应用。为乳腺癌分割开发了强大而高效的人工智能助手来自七个医疗中心的大量 DCE-MRI 数据用于训练我们的人工智能助手专门设计的时空转换器用于捕获动态我们的人工智能助手有潜力构建自动化乳腺癌诊断系统乳腺癌是影响全球女性的最常见癌症。早期检测和诊断对于更好的治疗结果和提高生存率至关重要。然而，乳腺癌的高发病率给临床医生带来了沉重的负担，而培训经验丰富的乳腺影像临床医生非常耗时。这就提出了开发强大的人工智能助手来协助临床医生诊断乳腺癌的需求。为此，我们从七个医疗中心收集了非常大的乳腺 DCE-MRI 数据，用于开发这样的乳腺 AI 助手，这也可以作为构建自动化乳腺癌诊断系统以促进未来智能医疗的基础步骤。临床乳腺癌诊断可能既费时又费力。在这项研究中，我们开发了一个人工智能助手，可以从 DCE-MRI 数据中自动分割乳腺肿瘤，并使用专门设计的时空转换器对大量数据进行训练。实验结果表明，我们的人工智能助手甚至可以比临床医生产生更准确的结果，同时使用的时间也显着减少。这表明了我们的人工智能助手在实际临床应用中的潜力。
Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) allows screening, follow up, and diagnosis for breast tumor with high sensitivity. Accurate tumor segmentation from DCE-MRI can provide crucial information of tumor location and shape, which significantly influences the downstream clinical decisions. In this paper, we aim to develop an artificial intelligence (AI) assistant to automatically segment breast tumors by capturing dynamic changes in multi-phase DCE-MRI with a spatial-temporal framework. The main advantages of our AI assistant include (1) robustness, i.e., our model can handle MR data with different phase numbers and imaging intervals, as demonstrated on a large-scale dataset from seven medical centers, and (2) efficiency, i.e., our AI assistant significantly reduces the time required for manual annotation by a factor of 20, while maintaining accuracy comparable to that of physicians. More importantly, as the fundamental step to build an AI-assisted breast cancer diagnosis system, our AI assistant will promote the application of AI in more clinical diagnostic practices regarding breast cancer. A robust and efficient AI assistant is developed for breast cancer segmentation A large set of DCE-MRI data from seven medical centers is used to train our AI assistant A specifically designed spatial-temporal transformer is used to capture dynamics Our AI assistant has potential for building an automated breast cancer diagnosis system Breast cancer is the most common cancer affecting women worldwide. Early detection and diagnosis are crucial for better treatment outcomes and improved survival rates. However, high incidence of breast cancer puts a significant burden on clinicians, while training experienced breast imaging clinicians is time-consuming. This raises the need for developing a robust AI assistant for assisting clinicians for breast cancer diagnosis. To this end, we collected a very large set of breast DCE-MRI data from seven medical centers for developing such a breast AI assistant, which can also be used as a foundational step for building an automated breast cancer diagnosis system to facilitate future smart medicine. Clinical breast cancer diagnosis can be time-consuming and laborious. In this study, we have developed an AI assistant to automatically segment breast tumors from DCE-MRI data, trained on a very large set of data with a specially designed spatiotemporal transformer. Experimental results show that our AI assistant can produce more accurate results even than clinicians while also using significantly less time. This indicates the potential of our AI assistant in real clinical applications.
登录
查看更多内容
DOI：
10.1186/s13058-017-0846-1
发表时间：
2017-05-18
期刊：
Breast cancer research : BCR
影响因子：
--
作者：
Braman NM;Etesami M;Prasanna P;Dubchuk C;Gilmore H;Tiwari P;Plecha D;Madabhushi A
通讯作者：
Madabhushi A
影响因子：
5.9
作者：
Kim, Jin Joo;Kim, Jin You;Park, Heeseung
通讯作者：
Park, Heeseung
影响因子：
64.8
作者：
McKinney, Scott Mayer;Sieniek, Marcin;Shetty, Shravya
通讯作者：
Shetty, Shravya
影响因子：
30.8
作者：
Leibig, Christian;Brehmer, Moritz;Bunk, Stefan;Byng, Danaiyn;Pinkert, Katja;Umutlut, Lale
通讯作者：
Umutlut, Lale
影响因子：
6.2
作者：
Ginsburg, Ophira;Yip, Cheng-Har;Anderson, Benjamin O.
通讯作者：
Anderson, Benjamin O.