Predicting axillary lymph node metastases in breast carcinoma patients

Area of Science:

• Oncology
• Biostatistics

Background:

• Routine axillary dissection in breast carcinoma staging can lead to unnecessary morbidity in node-negative patients.
• Accurate prediction of lymph node metastasis is crucial for appropriate treatment planning and avoiding overtreatment.

Purpose of the Study:

• To develop and evaluate a probabilistic model for predicting lymph node metastasis in primary breast carcinoma patients.
• To identify tumor-based parameters that correlate with the risk of axillary lymph node involvement.

Main Methods:

• Utilized Maximum Entropy Estimation (MEE) to construct probabilistic models.
• Analyzed data from 217 invasive breast carcinoma patients with known axillary node status.
• Performed multivariate analysis correlating tumor size, age, and vascular invasion with nodal status.

Main Results:

• Tumor size was significantly correlated with axillary lymph node status (P < 0.001).
• The MEE model predicted <20% risk of nodal metastases in 38 patients, with only 4 (10%) actually having metastases.
• The MEE model demonstrated superior predictive quality compared to Multivariate Logistic Regression (MLR).

Conclusions:

• The developed MEE model shows promise in accurately assessing the probability of lymph node metastasis.
• This approach may help select patients for or against axillary dissection, potentially reducing patient morbidity.
• Further validation in larger populations is recommended to confirm the practical application of the MEE model.