A multimodal ultrasound-based model combining tumor radiomics and axillary lymph node morphologic classification for predicting axillary nodal burden in breast cancer - Report - MDSpire
Advertisement
A multimodal ultrasound-based model combining tumor radiomics and axillary lymph node morphologic classification for predicting axillary nodal burden in breast cancer
Clinical Report: Integrated Ultrasound Model for Predicting Nodal Involvement
Overview
This study developed and validated a multimodal ultrasound model that integrates tumor radiomics and axillary lymph node morphological assessment to predict axillary nodal metastasis in breast cancer. The model demonstrated improved predictive performance, particularly in distinguishing nodal burden based on Ki-67 subgroups.
Background
Accurate assessment of axillary lymph node involvement is crucial for staging, treatment decisions, and quality of life in breast cancer patients. Traditional methods, such as sentinel lymph node dissection, can be invasive and carry risks of complications. Recent advances in ultrasound and radiomics offer potential for non-invasive assessment of nodal status, which is essential for personalized treatment planning.
Data Highlights
Model Type
AUC
Radiomics Model (First Level)
0.79
Combined Model (First Level)
0.90
Radiomics Model (Second Level)
0.74
Combined Model (Second Level)
0.78
Key Findings
The combined model incorporating ALN classification improved AUC from 0.79 to 0.90 for predicting ALN metastasis.
At the second level, the combined model increased AUC from 0.74 to 0.78 for distinguishing axillary nodal tumor burden.
Ki-67 was identified as an independent predictor of nodal involvement.
The combined model performed consistently well across Ki-67 subgroups, with superior performance in the low Ki-67 subgroup (AUC 0.82).
Ultrasound-based radiomics may enhance preoperative assessment of ALN status.
Clinical Implications
The multimodal ultrasound model may facilitate non-invasive prediction of axillary nodal metastasis, aiding in individualized risk stratification and treatment planning. This approach could optimize axillary management and reduce unnecessary surgical interventions.
Conclusion
The integration of tumor radiomics with lymph node morphology presents a promising strategy for accurate preoperative prediction of axillary nodal involvement in breast cancer. Further validation in clinical settings is warranted.
