Multi-task transfer learning deep convolutional neural network: Application to computer-aided diagnosis of breast cancer on mammograms

By Ravi K. Samala, Heang-Ping Chan, Lubomir M. Hadjiiski, Mark A. Helvie, Kenny H. Cha, Caleb D. Richter. 2017

203 citations recorded by the source corpus.

Transfer learning in deep convolutional neural networks (DCNNs) is an important step in its application to medical imaging tasks. We propose a multi-task transfer learning DCNN with the aim of translating the ‘knowledge’ learned from non-medical images to medical diagnostic tasks through supervised training and increasing the generalization capabilities of DCNNs by simultaneously learning auxiliary tasks. We studied this approach in an important application: classification of malignant and benign breast masses. With Institutional Review Board (IRB) approval, digitized screen-film mammograms (SFMs) and digital mammograms (DMs) were collected from our patient files and additional SFMs were obtained from the Digital Database for Screening Mammography. The data set consisted of 2242 views with 2454 masses (1057 malignant, 1397 benign). In single-task transfer learning, the DCNN was trained and tested on SFMs. In multi-task transfer learning, SFMs and DMs were used to train the DCNN, which was then tested on SFMs. N-fold cross-validation with the training set was used for training and parameter optimization. On the independent test set, the multi-task transfer learning DCNN was found to have significantly (p = 0.007) higher performance compared to the single-task transfer learning DCNN. This study demonstrates that multi-task transfer learning may be an effective approach for training DCNN in medical imaging applications when training samples from a single modality are limited.

Subfields: Artificial Intelligence, Radiology, Nuclear Medicine and Imaging

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Radiology, Nuclear Medicine and Imaging

Citations in the corpus

Distributed deep learning networks among institutions for medical imaging (2018, methodology): “In comparison, the concept of single weight transfer is very similar to that of transfer learning, which is derived from that idea that a model can solve new problems faster by using knowledge learned from solving previous problems in other domains.(23,24) In practice, this involves training a model”
An overview of deep learning in medical imaging focusing on MRI (2018, background): “For example the comprehensive review [101],24 covering both medicine and biology and spanning from imaging applications in healthcare to protein-protein interaction and uncertainty quantification; key concepts of deep learning for clinical radiologists [102–111], including radiomics and imaging geno”
Deep learning in medical imaging and radiation therapy. (2018, background): “The latter has been shown by Samala et al. 199”
A Review of Deep Learning on Medical Image Analysis (2020, background): “[68] 2017 Mammograms Fine-tuning (Multi-task)”
Computer-Aided Diagnosis in the Era of Deep Learning (2020, methodology): “Initial efforts have been made to use these tools to visualize the detected location of abnormalities 67,71 or to visualize the characteristics of the deep features 72 on medical images.”
A Novel Deep-Learning Model for Automatic Detection and Classification of Breast Cancer Using the Transfer-Learning Technique (2021, methodology): “An appropriate TL method is a model transfer; ﬁrst, the network parameters are pre-trained using the source data, then these parameters are applied in the target domain, and ﬁnally the network parameters are adjusted for better performance [79].”
Deep learning-based breast cancer classification through medical imaging modalities: state of the art and research challenges (2019, methodology): “…studies, nine (Bakkouri and Afdel 2017; Carneiro et al. 2015, 2017; Jaffar 2017; Kumar et al. 2017b; Leod and Verma 2016; Rouhi et al. 2015; Samala et al. 2017; Sert et al. 2017) used DDSM datasets and six (Arefan et al. 2015; Duraisamy and Emperumal 2017; Jaffar 2017; Khan 2017; Kumar et…”
GPU accelerated t-distributed stochastic neighbor embedding (2019, background): “Visualizing such large dimensional data can be extremely important in domains such as natural language processing [3], image processing for breast-cancer detection [4], spatio-temporal video data analysis ([5], [6]), and many others.”