MammoGANesis: Controlled Generation of High-Resolution Mammograms for Radiology Education
This addresses the need for accessible training data for radiology trainees, though it is incremental as it applies existing GAN methods to a new medical domain.
The paper tackles the problem of limited access to medical images for radiology education by training a GAN to synthesize high-resolution mammograms, achieving an average AUC of 0.54 in a double-blind study where experts struggled to distinguish generated from real images.
During their formative years, radiology trainees are required to interpret hundreds of mammograms per month, with the objective of becoming apt at discerning the subtle patterns differentiating benign from malignant lesions. Unfortunately, medico-legal and technical hurdles make it difficult to access and query medical images for training. In this paper we train a generative adversarial network (GAN) to synthesize 512 x 512 high-resolution mammograms. The resulting model leads to the unsupervised separation of high-level features (e.g. the standard mammography views and the nature of the breast lesions), with stochastic variation in the generated images (e.g. breast adipose tissue, calcification), enabling user-controlled global and local attribute-editing of the synthesized images. We demonstrate the model's ability to generate anatomically and medically relevant mammograms by achieving an average AUC of 0.54 in a double-blind study on four expert mammography radiologists to distinguish between generated and real images, ascribing to the high visual quality of the synthesized and edited mammograms, and to their potential use in advancing and facilitating medical education.