Validating the MSKCC nomogram and a clinical decision rule in the prediction of non-sentinel node metastases in a Portuguese population of breast cancer patients.

BACKGROUND: In order to assess the risk of non-sentinel node involvement in breast cancer patients, some prediction tools have been developed and tested. However, a wide range of results are observed. We tested a simplified clinical decision rule, against the complex nomogram created from the MSKCC sentinel node database.
METHODS: Two single institutional consecutive series of patients with a positive SN, submitted to SN biopsy plus axillary dissection from June 1999 to October 2007, were evaluated. A receiver operating curve was drawn and the area under the curve was calculated as well as the negative predictive value for both tests, assuming discriminative values of 10 and 15%.
RESULTS: Considering the derivation series, our results showed an area under the curve of 0.69 for both our clinical decision rule and the MSKCC nomogram. The analysis of the validation series showed an area under the curve of 0.65 for our clinical decision rule and of 0.67 for the MSKCC nomogram. The nomogram results are inferior to those found in the original population and are similar to our clinical decision rule results.
CONCLUSIONS: Individual centres should develop and prospectively test their own clinical decision rules, based on their institutional Sentinel Node data.