Artificial Intelligence In The Diagnosis And Prediction Of Breast Cancer-Related Lymphedema: A Scoping Review

Area of Science:

• Oncology
• Medical Informatics
• Biostatistics

Background:

• Lymphedema is a chronic, debilitating complication following breast cancer treatment.
• Early diagnosis and intervention are crucial for managing lymphedema and improving patient well-being.

Purpose of the Study:

• To conduct a scoping review evaluating the application of artificial intelligence (AI) in diagnosing and predicting breast cancer-related lymphedema.
• To assess the effectiveness of AI-based techniques in enhancing diagnostic precision and facilitating timely interventions for lymphedema.

Main Methods:

• A comprehensive literature search was conducted across major databases (PubMed, Scopus, EMBASE, Web of Science, Cochrane) up to September 2024.
• Included studies focused on AI techniques for lymphedema detection, severity assessment, and early diagnosis, requiring English full-text availability.
• Excluded were editorials, review papers, and studies with inaccessible full texts.

Main Results:

• Thirteen studies met the inclusion criteria, utilizing diverse AI models and data inputs (EHR, clinical data, patient-reported symptoms, imaging).
• Support Vector Machine (SVM) was the most frequent AI algorithm (61.5%), often combined with other supervised learning methods.
• Risk prediction models achieved high accuracy (81%-93.75%), sensitivity (95.65%), specificity (91.03%), and AUC (0.751). Lymphedema severity models showed accuracy up to 99.4% with AUCs from 0.889 to 0.931.

Conclusions:

• AI holds significant potential for improving the diagnosis and prediction of breast cancer-related lymphedema.
• AI applications can lead to enhanced diagnostic capabilities and personalized treatment strategies.
• Future research should focus on data standardization, robust model validation, and clinical implementation frameworks.