VMRA-MaR: An Asymmetry-Aware Temporal Framework for Longitudinal Breast Cancer Risk Prediction

2025-06-20

AI Summary

This study addresses two key challenges in long-term (4–5 year) breast cancer risk prediction: the difficulty of modeling dense breast tissue in mammograms and the underutilization of longitudinal dynamic information. We propose VMRNN—the first state-space model integrating Vision Mamba with RNNs—and introduce two novel components: the Spatial Asymmetry Detector (SAD) and the Longitudinal Asymmetry Tracker (LAT), enabling, for the first time, clinically interpretable joint modeling of bilateral structural and evolutionary asymmetry. Our method synergistically combines Vision Mamba’s long-range visual modeling capability, state-space models’ efficient temporal representation, and LSTM-style memory mechanisms. On dense-breast subsets, VMRNN achieves superior long-term prediction performance: its 4–5 year AUC improves by over 8 percentage points relative to the best Transformer-based baseline. This work establishes a new paradigm for precision early breast cancer screening.

Technology Category

• Computer Vision: Multi-modal Vision
• Machine Learning: Large Multimodal Models (LMMs)
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Breast cancer remains a leading cause of mortality worldwide and is typically detected via screening programs where healthy people are invited in regular intervals. Automated risk prediction approaches have the potential to improve this process by facilitating dynamically screening of high-risk groups. While most models focus solely on the most recent screening, there is growing interest in exploiting temporal information to capture evolving trends in breast tissue, as inspired by clinical practice. Early methods typically relied on two time steps, and although recent efforts have extended this to multiple time steps using Transformer architectures, challenges remain in fully harnessing the rich temporal dynamics inherent in longitudinal imaging data. In this work, we propose to instead leverage Vision Mamba RNN (VMRNN) with a state-space model (SSM) and LSTM-like memory mechanisms to effectively capture nuanced trends in breast tissue evolution. To further enhance our approach, we incorporate an asymmetry module that utilizes a Spatial Asymmetry Detector (SAD) and Longitudinal Asymmetry Tracker (LAT) to identify clinically relevant bilateral differences. This integrated framework demonstrates notable improvements in predicting cancer onset, especially for the more challenging high-density breast cases and achieves superior performance at extended time points (years four and five), highlighting its potential to advance early breast cancer recognition and enable more personalized screening strategies. Our code is available at https://github.com/Mortal-Suen/VMRA-MaR.git.

Problem

• Improves breast cancer risk prediction using temporal data
• Captures breast tissue evolution with Vision Mamba RNN
• Identifies bilateral differences via asymmetry-aware modules

Innovation

• Uses Vision Mamba RNN for temporal trend capture
• Integrates asymmetry module with SAD and LAT
• Improves prediction for high-density breast cases