Training Strategies for Radiology Deep Learning Models in Data-limited Scenarios

Data-driven approaches have great potential to shape future practices in radiology. The most straightforward strategy to obtain clinically accurate models is to use large, well-curated and annotated datasets. However, patient privacy constraints, tedious annotation processes, and the limited availability of radiologists pose challenges to building such datasets. This review details model training strategies in scenarios with limited data, insufficiently labeled data, and/or limited expert resources. This review discusses strategies to enlarge the data sample, decrease the time burden of manual supervised labeling, adjust the neural network architecture to improve model performance, apply semisupervised approaches, and leverage efficiencies from pretrained models. (C) RSNA, 2021.

Automated summary

Data-driven approaches have great potential in shaping future practices in radiology, but face challenges such as patient privacy concerns, tedious annotation processes, and limited expert resources. This review discusses model training strategies in scenarios with limited data, insufficiently labeled data, and/or limited expert resources, including enlarging data samples, decreasing manual labeling burdens, adjusting network architectures, and leveraging pretrained models.