ABSTRACT

INTRODUCTION:

Non sentinel axillary lymph node metastasis (NSLNM) occurs in 35-50% of breast cancer (BC) patients having positive sentinel lymph nodes (SLN). A nomogram which includes 8 variables was developed at Memorial Sloan Kettering Cancer Center (MSKCC) in 2003 and it has been validated at sources outside that institution. The Stanford University group recently reported their nomogram which evaluated 3 variables.

AIM:

The aim of this study is to evaluate the predictability of two diff erent scoring systems wherein 3 or 8 variables are used in the same patient groups.

MATERIALS and METHODS:

We identifi ed 201 patients who had a positive SLN biopsy and completion axillary lymph node dissection at Magee-Womens Hospital of UPMC over a 5 year period. The computerized BC nomograms developed by MSKCC and Stanford University were used to calculate the probability of non-sentinel lymph node metastases. Area Under (AUC) Receiver Operating Characteristics Curve (ROC) was calculated for each nomogram and the values more than 0.70 have been accepted that presents considerable discrimination.

RESULTS:

Sixty-six of 201 patients (32.8%) had positive axillary NSLNM. The mean predicted probability of positive NSLNM was 25.4% (3-93), and 66.3% (7-100) for the MSKCC and Stanford nomograms, respectively. The AUC values were 0.73 and 0.67 for MSKCC and Stanford nomograms, respectively.

DISCUSSION and CONCLUSION:

Nomograms for predicting the probability of NSLNM in BC patients have been in use for 5 years. It is clear there are discrepancies in the results of nomograms among the studies using the same scoring system. Notwithstanding the Stanford nomogram is easier to implement as it considers only 3 variables in our study, we found the MSKCC nomogram to be more predictive than the Stanford nomogram Nomograms developed at outside institutions should be used with caution when counseling patients regarding the risk of additional nodal disease.