Topological Conditioning for Mammography Models via a Stable Wavelet-Persistence Vectorization
This work addresses the challenge of false negatives and positives in breast cancer screening, offering a method to enhance model robustness across diverse mammography datasets, though it is incremental as it builds on existing detection pipelines.
The paper tackled the problem of improving external performance of mammography models across different scanners and populations by proposing a topological conditioning signal using wavelet-based vectorization of persistent homology, which increased patient-level AUC from 0.55 to 0.75 on the INbreast dataset.
Breast cancer is the most commonly diagnosed cancer in women and a leading cause of cancer death worldwide. Screening mammography reduces mortality, yet interpretation still suffers from substantial false negatives and false positives, and model accuracy often degrades when deployed across scanners, modalities, and patient populations. We propose a simple conditioning signal aimed at improving external performance based on a wavelet based vectorization of persistent homology. Using topological data analysis, we summarize image structure that persists across intensity thresholds and convert this information into spatial, multi scale maps that are provably stable to small intensity perturbations. These maps are integrated into a two stage detection pipeline through input level channel concatenation. The model is trained and validated on the CBIS DDSM digitized film mammography cohort from the United States and evaluated on two independent full field digital mammography cohorts from Portugal (INbreast) and China (CMMD), with performance reported at the patient level. On INbreast, augmenting ConvNeXt Tiny with wavelet persistence channels increases patient level AUC from 0.55 to 0.75 under a limited training budget.