English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Data scarcity and class imbalance are two fundamental challenges in manymachine learning applications to healthcare. Breast cancer classification inmammography exemplifies these challenges, with a malignancy rate of around 0.5%in a screening population, which is compounded by the relatively small size oflesions (~1% of the image) in malignant cases. Simultaneously, the prevalenceof screening mammography creates a potential abundance of non-cancer exams touse for training. Altogether, these characteristics lead to overfitting oncancer cases, while under-utilizing non-cancer data. Here, we present a novelgenerative adversarial network (GAN) model for data augmentation that canrealistically synthesize and remove lesions on mammograms. With self-attentionand semi-supervised learning components, the U-net-based architecture cangenerate high resolution (256x256px) outputs, as necessary for mammography.When augmenting the original training set with the GAN-generated samples, wefind a significant improvement in malignancy classification performance on atest set of real mammogram patches. Overall, the empirical results of ouralgorithm and the relevance to other medical imaging paradigms point topotentially fruitful further applications.

Synthesizing lesions using contextual GANs improves breast cancer classification on mammograms