【文献题目】Prediction of breast cancer molecular subtypes on DCE-MRI using convolutional neural network with transfer learning between two centers.

[Yang Zhang, Jeon-Hor Chen, Yezhi Lin, Siwa Chan, Jiejie Zhou, Daniel Chow, Peter Chang, Tiffany Kwong, Dah-Cherng Yeh, Xinxin Wang, Ritesh Parajuli, Rita S. Mehta, Meihao Wang, Min-Ying Su. Prediction of breast cancer molecular subtypes on DCE-MRI using convolutional neural network with transfer learning between two centers. European Radiology (2021) 31:2559–2567]

目的：应用传统卷积神经网络（CNN）和递归CNN的深度学习算法，在MRI上区分三种乳腺癌分子亚型。

方法：根据激素受体（HR）和HER2受体分为3个亚型：（HR+/HER2−), HER2+ 和三重阴性（TN）；利用传统CNN和卷积长短期记忆网络（CLSTM），以覆盖肿