New models and online calculator for predicting non-sentinel lymph node status in sentinel lymph node positive breast cancer patients

Table 3 Model comparisons for predicting NSLN metastasis among SLN+ patients (Bay Area SLN Database).

	For all SLN+ Pts (n = 285)	For pts with known angiolymphatic invasion status (n = 213)	For pts with known angiolymphatic invasion and ER status (n = 171)
RP-ROC with 10-fold cross validation
Sensitivity (%)		78.8	83.2
Specificity (%)		75.5	78.1
Diagnostic Accuracy by AUC (%)		76.7	80.2
Boosted CART with 10-fold cross validation
Sensitivity (%)	78.2	87.9	89.0
Specificity (%)	62.0	71.4	74.7
Diagnostic Accuracy by AUC (%)	67.7	77.5	80.3
Multivariable logistic regression with 10-fold cross validation
Sensitivity (%)		78.0	78.9
Specificity (%)		86.2	88.3
Diagnostic Accuracy by AUC (%)		83.3	84.9
MSKCC Breast Cancer Nomogram ^a
Diagnsotic Accuracy by AUC (%)			76.7

^a Memorial Sloan-Kettering Cancer Center Breast Cancer Nomogram for Prediction of Axillary Lymph Node Metastasis applied to SLN+ pts in Stanford dataset who have complete data for all Nomogram variables

ISSN: 1471-2407