Mammography

Mammography uses breast X-rays to detect cancer, characterize abnormalities, and help guide decisions such as follow-up or biopsy. AI approaches include convolutional networks and vision transformers for lesion detection, segmentation, and classification, alongside vision-language models and systems that combine images with clinical information. Researchers also use multiple views and prior screening exams to estimate future risk, and explore synthetic images and self-supervised learning to reduce reliance on dense annotations. A central challenge is ensuring models generalize across hospitals, vendors, and populations; external testing and comparisons with radiologists remain important, especially for safe screening decisions.