Preoperative Imaging Assessment of Lymphovascular Invasion in Breast Cancer: Current Evidence, Technical Advances, and Future Directions

Lymphovascular invasion (LVI) in breast cancer is a pivotal prognostic factor that directly influences patient survival outcomes and guides the formulation of individualized treatment strategies, such as the selection of adjuvant chemotherapy regimens or the extent of surgical intervention. Accurate preoperative assessment of LVI status is therefore indispensable for optimizing clinical decision-making and improving patient management. This review systematically synthesizes the latest advancements in imaging techniques for LVI evaluation in breast cancer, including conventional ultrasound, contrast-enhanced ultrasound (CEUS), elastography, mammography (digital and contrast-enhanced), magnetic resonance imaging (MRI, encompassing DCE-MRI, DWI, and novel functional sequences like IVIM), and artificial intelligence (AI)-assisted diagnostic tools (radiomics and deep learning). A critical analysis of these modalities reveals distinct strengths and limitations: for instance, ultrasound offers broad accessibility and real-time imaging but exhibits lower sensitivity (60-75%) for subtle LVI lesions compared to MRI (sensitivity 78-90%), which provides superior soft tissue contrast but is constrained by higher cost and longer scan times. Radiomics models, while demonstrating promising AUC values (0.74-0.896) in predicting LVI, suffer from inconsistencies in feature extraction protocols and limited external validation. Additionally, this review highlights key knowledge gaps, such as the lack of standardized imaging parameters for LVI assessment across centers and the underrepresentation of rare breast cancer subtypes (eg, inflammatory breast cancer) in existing studies. We emphasize the clinical urgency of addressing these limitations-including reducing false-positive rates caused by benign vascular proliferations and false negatives from obscured lesions-to enhance diagnostic accuracy. Furthermore, we propose targeted future directions, such as the development of multicenter, prospective trials to validate AI-integrated multimodal imaging workflows and the establishment of consensus guidelines for imaging protocol standardization, ultimately aiming to translate technical advancements into improved patient outcomes.