Deep learning analysis of breast MRIs for prediction of occult invasive disease in ductal carcinoma in situ. (December 2019)

• Record Type: Journal Article
• Authors: Zhu, Zhe; Harowicz, Michael; Zhang, Jun; Saha, Ashirbani; Grimm, Lars J.; Hwang, E. Shelley; Mazurowski, Maciej A.
• Is Part Of: Computers in biology and medicine. Volume 115(2019)
• Publication Date: 2019-12
• Subjects: DCIS (ductal carcinoma in situ) -- SVM (support vector machine) -- AUC (area under the ROC curve) -- CNB (core needle biopsy) -- GUI (graphical user interface) -- (SGD) stochastic gradient descent; Medicine -- Data processing -- Periodicals; Biology -- Data processing -- Periodicals; 610.285
• DOI: 10.1016/j.compbiomed.2019.103498
• Languages: English
• ISSNs: 0010-4825

Abstract

Abstract: Purpose: To determine whether deep learning-based algorithms applied to breast MR images can aid in the prediction of occult invasive disease following the diagnosis of ductal carcinoma in situ (DCIS) by core needle biopsy. Materials and methods: Our study is retrospective. The data was collected from 2000 to 2014. In this institutional review board-approved study, we analyzed dynamic contrast-enhanced fat-saturated T1-weighted MRI sequences from 131 patients with a core needle biopsy-confirmed diagnosis of DCIS. We explored two different deep learning approaches to predict whether there was an occult invasive component in the analyzed tumors that was ultimately identified at surgical excision. In the first approach, we adopted the transfer learning strategy. Specifically, we used the pre-trained GoogleNet. In the second approach, we used a pre-trained network to extract deep features, and a support vector machine (SVM) that utilizes these features to predict the upstaging of DCIS. We used nested 10-fold cross validation and the area under the ROC curve (AUC) to estimate the performance of the predictive models. Results: The best classification performance was obtained using the deep features approach with GoogleNet model pre-trained on ImageNet as the feature extractor and a polynomial kernel SVM used as the classifier (AUC = 0.70, 95% CI: 0.58–0.79). For the transfer learning based approach, the highest AUC obtained was 0.68 (95% CI: 0.57–0.77). Conclusions: Convolutional neural networks might be used to identify occult invasive disease in patients diagnosed with DCIS by core needle biopsy.