Work

Lymphedema imaging and AI: A review of diagnostic modalities, biomarkers, and clinical integration

Year: 2025

Type: article

Abstract: Lymphedema, a chronic lymphatic disorder characterized by swelling, fibrosis, and adipose tissue accumulation, requires precise, stage-specific imaging for effective diagnosis and management. This review evaluates conventional imaging modalities, including indocyanine green lymphography (ICG-L), lymphoscintigraphy, magnetic resonance imaging (MRI), and computed tomography (CT), alongside emerging artificial intelligence (AI) applications to enhance diagnostic accuracy and treatment planning. We analyze their capabilities in assessing lymphatic function and tissue changes through quantitative biomarkers, comparing their strengths across disease stages. ICG-L excels in detecting early lymphatic dysfunction, while MRI and CT provide detailed visualization of advanced fibrotic and adipose changes. AI-driven tools, such as automated segmentation and biomarker quantification, show promise in improving tissue characterization and supporting surgical planning. However, clinical integration of AI is hindered by data heterogeneity, lack of interpretability, and regulatory challenges. To address these, we propose a strategic framework incorporating federated learning for privacy-preserving model training, explainable AI for clinical transparency, and standardized imaging protocols. Future efforts should prioritize multicenter validation and harmonized guidelines to enhance reproducibility and ensure equitable, scalable adoption of advanced imaging technologies for lymphedema management worldwide. • Review of imaging (ICG, MRI, CT, lymphoscintigraphy) for lymphedema diagnosis and staging. • Quantitative biomarkers: fat fraction >30%, fibrosis ratio >25% indicate severity. • AI aids segmentation (Dice >0.945), classification, clinical reporting. • Challenges: data heterogeneity, interpretability, regulations. • Framework: explainable AI, federated learning, standardized protocols for adoption.

Source: Biomedical Journal

DOI: 10.1016/j.bj.2025.100932

Authors: Bushra Urooj, Sabir Ali, Syed Kumail Hussain Naqvi, Furen Xiao, Po-Cheng Huang

Institutions: National Taiwan University, Jeonbuk National University, National Taiwan University Hospital

Language: English

Topic: Lymphatic System and Diseases

Subfield: Oncology

Field: Medicine

Domain: Health Sciences

SDG: Good health and well-being

Open Access status: gold

Estimated APC paid: $2,700