Abstract
Accurate segmentation of small lesions in Breast Dynamic Contrast-EnhancedMRI (DCE-MRI) is critical for early cancer detection, especially in high-riskpatients. While recent deep learning methods have advanced lesion segmentation,they primarily target large lesions and neglect valuable longitudinal andclinical information routinely used by radiologists. In real-world screening,detecting subtle or emerging lesions requires radiologists to compare acrosstimepoints and consider previous radiology assessments, such as the BI-RADSscore. We propose LesiOnTime, a novel 3D segmentation approach that mimicsclinical diagnostic workflows by jointly leveraging longitudinal imaging andBIRADS scores. The key components are: (1) a Temporal Prior Attention (TPA)block that dynamically integrates information from previous and current scans;and (2) a BI-RADS Consistency Regularization (BCR) loss that enforces latentspace alignment for scans with similar radiological assessments, thus embeddingdomain knowledge into the training process. Evaluated on a curated in-houselongitudinal dataset of high-risk patients with DCE-MRI, our approachoutperforms state-of-the-art single-timepoint and longitudinal baselines by 5%in terms of Dice. Ablation studies demonstrate that both TPA and BCR contributecomplementary performance gains. These results highlight the importance ofincorporating temporal and clinical context for reliable early lesionsegmentation in real-world breast cancer screening. Our code is publiclyavailable at https://github.com/cirmuw/LesiOnTime