A model for predicting non-sentinel lymph node metastatic disease when the sentinel lymph node is positive

Area of Science:

• Oncology
• Surgical Oncology
• Breast Cancer Research

Background:

• Women with sentinel lymph node (SLN)-positive breast cancer often undergo completion axillary lymph node dissection (ALND).
• Not all patients with positive SLNs have further axillary nodal disease, suggesting completion ALND may be avoidable in select cases.
• The Memorial Sloan-Kettering Cancer Center (MSKCC) developed a nomogram to predict the risk of non-SLN disease.

Purpose of the Study:

• To critically appraise the predictive accuracy of the MSKCC nomogram for non-SLN disease.
• To refine the existing nomogram and develop an improved predictive model for further axillary lymph node involvement.
• To enhance the accuracy of identifying patients who can safely avoid completion ALND.

Main Methods:

• The MSKCC nomogram was applied to 118 patients with positive SLN biopsy who underwent completion ALND.
• Predictive accuracy was assessed using the area under the receiver-operator characteristic (ROC) curve.
• A revised predictive model was developed using detailed pathological information and backward stepwise multiple logistic regression, validated with k-fold cross-validation.

Main Results:

• The original MSKCC nomogram demonstrated an area under the ROC curve of 68%.
• The revised predictive model achieved a weighted mean area under the ROC curve of 84%, indicating significantly improved predictive accuracy.
• The refined model incorporating SLN metastasis size showed enhanced performance.

Conclusions:

• The modified predictive model significantly improves the prediction of further axillary lymph node disease in breast cancer patients.
• Incorporating the size of sentinel lymph node metastasis enhances predictive accuracy.
• Further validation of the refined model on an independent dataset is recommended to confirm its clinical utility.