Predicting Breast Cancer Events in Ductal Carcinoma In Situ (DCIS) Using Generative Adversarial Network Augmented Deep Learning Model.
Predicting Breast Cancer Events in Ductal Carcinoma In Situ (DCIS) Using Generative Adversarial Network Augmented Deep Learning Model.
复制标题
用生成性对抗网络增强深度学习模型预测导管原位癌(DCIS)中的乳腺癌事件
作者：
导管原位癌（DCIS）患者有很好的总体生存率，但由于潜在的副作用，过度治疗一直是一个令人担忧的问题。标准临床病理参数在预测乳腺癌事件（bce）和高危和低危患者分层方面价值有限。在此，我们开发了一个深度学习（DL）分类框架来预测DCIS患者的bce。生成对抗网络（GAN）增强了与侵袭性疾病相关的组织学特征的深度学习（DL）分类，并在DCIS的苏木精和伊红（h&e）组织微阵列（TMA）图像上进行了训练，以预测bce。验证集中BCE的曲线下面积（AUC）为0.82。早期和准确预测DCIS bce将有助于个性化的治疗方法。标准临床病理参数（年龄、生长模式、肿瘤大小、边缘状态和分级）在预测导管原位癌（DCIS）患者复发方面的价值有限。早期准确的复发预测有助于对高危患者（乳房切除术或辅助放疗）采取更积极的治疗政策，同时减少对低危患者的过度治疗。生成式对抗网络（GAN）是一类深度学习模型，其中两个对抗神经网络生成器和鉴别器相互竞争以生成高质量的图像。在这项工作中，我们开发了一个深度学习（DL）分类网络，使用苏木精和伊红（H & E）图像预测DCIS患者的乳腺癌事件（bce）。使用来自实际DCIS核心的图像贴片和GAN生成的图像贴片对67例患者进行DL分类模型训练，以预测乳腺癌事件（bce）。保留验证数据集（n = 66）的AUC为0.82。贝叶斯分析进一步证实了模型与经典临床病理参数的独立性。h&e图像的DL模型可作为DCIS患者个性化治疗的风险分层策略。
Ductal carcinoma in situ (DCIS) patients have an excellent overall survival rate and over-treatment is always a cause for concern due to potential side-effects. Standard clinicopathological parameters have limited value in predicting breast cancer events (BCEs) and stratification of high and low risk patients. Herein, we have developed a deep learning (DL) classification framework to predict BCEs in DCIS patients. A generative adversarial network (GAN) augmented deep learning (DL) classification of histological features associated with aggressive disease was trained on hematoxylin and eosin (H & E) tissue microarray (TMA) images of DCIS to predict BCEs. The area under the curve (AUC) for BCE’s in the validation set was 0.82. Early and accurate prediction of DCIS BCEs would facilitate a personalized approach to therapy. Standard clinicopathological parameters (age, growth pattern, tumor size, margin status, and grade) have been shown to have limited value in predicting recurrence in ductal carcinoma in situ (DCIS) patients. Early and accurate recurrence prediction would facilitate a more aggressive treatment policy for high-risk patients (mastectomy or adjuvant radiation therapy), and simultaneously reduce over-treatment of low-risk patients. Generative adversarial networks (GAN) are a class of DL models in which two adversarial neural networks, generator and discriminator, compete with each other to generate high quality images. In this work, we have developed a deep learning (DL) classification network that predicts breast cancer events (BCEs) in DCIS patients using hematoxylin and eosin (H & E) images. The DL classification model was trained on 67 patients using image patches from the actual DCIS cores and GAN generated image patches to predict breast cancer events (BCEs). The hold-out validation dataset (n = 66) had an AUC of 0.82. Bayesian analysis further confirmed the independence of the model from classical clinicopathological parameters. DL models of H & E images may be used as a risk stratification strategy for DCIS patients to personalize therapy.
登录
查看更多内容
影响因子：
5.2
作者：
Badve, Sunil S.;Cho, Sanghee;Lu, Xiaoyu;Cao, Sha;Ghose, Soumya;Thike, Aye Aye;Tan, Puay Hoon;Ocal, Idris Tolgay;Generali, Daniele;Zanconati, Fabrizio;Harris, Adrian L.;Ginty, Fiona;Gokmen-Polar, Yesim
通讯作者：
Gokmen-Polar, Yesim
DOI：
10.1056/nejmoa2108873
发表时间：
2021-12-16
期刊：
The New England journal of medicine
影响因子：
--
作者：
通讯作者：
--
影响因子：
5.4
作者：
Lee G;Veltri RW;Zhu G;Ali S;Epstein JI;Madabhushi A
通讯作者：
Madabhushi A
DOI：
10.1038/s41374-018-0095-7
发表时间：
2018-11
期刊：
Laboratory investigation; a journal of technical methods and pathology
影响因子：
--
作者：
Lu C;Romo-Bucheli D;Wang X;Janowczyk A;Ganesan S;Gilmore H;Rimm D;Madabhushi A
通讯作者：
Madabhushi A
影响因子：
28.4
作者：
Lehman, Constance D.;Gatsonis, Constantine;Comstock, Christopher
通讯作者：
Comstock, Christopher