Leveraging immune and clinicopathological profiles with machine learning to predict axillary lymph node metastasis in breast cancer patients
A machine learning model using immune cell patterns predicts lymph node cancer spread with fair accuracy, potentially reducing unnecessary surgeries.
A machine learning model integrating immune and clinicopathological profiles from luminal breast cancer patients achieved AUC 0.84 for predicting axillary lymph node metastasis, with immune features dominating over traditional prognostic variables. The identification of CD21+ follicular dendritic cells as the strongest predictor is a novel finding suggesting immune microenvironment profiling may improve surgical staging decisions.
What the study was
- Study design
- Retrospective; Random Forest ML; two independent datasets; SHAP interpretability
- Population
- Luminal breast cancer patients; Dataset 1: n=83; Dataset 2: n=344; Spain multicenter
- Sample size
- 427
- Category
- Diagnostics
- Maturity
- Exploratory
- Journal
- Breast Cancer Research
Why it surfaced
CD21+ as top predictor is novel and pathobiologically interesting. Limited by retrospective design and single-institution datasets 1995-2008; external validation in modern cohorts required.
A plain-language summary of published research — not medical advice. Talk to a clinician about your own care.