Abstract P1-01-16: Prediction of non-sentinel lymph node status in breast cancer patients with sentinel lymph node metastases: Comparison of the MD Anderson, Cambridge and Helsinki nomograms in a cohort of 80 patients

Abstract

Abstract Introduction: Up to 60% of patients with one or more positive sentinel lymph nodes (SLNs) have no additional nodal involvement and do not benefit from completion axillary lymph node dissection (ALND). Several predictive nomograms have been developed to predict non-SLN status in patients with breast cancer and SLN metastases. The purpose of our retrospective investigation was to compare some of these nomograms and we chose 3, all of which incorporate the size of the metastases in the SLNs among their variables. Methods: Three different predictive nomograms (MD Anderson, Helsinki and Cambridge) were tested on a cohort of 80 SLN-positive patients who subsequently underwent completion ALND. The nomograms utilise 3 (Cambridge), 7 (MD Anderson) and 9 (Helsinki) variables. The predictive accuracy of the nomograms for our cohort was evaluated by assessing receiver operating characteristic (ROC) curves and areas under curves (AUC) for each nomogram. Descriptive characteristics of the patient populationTotal number of patients80 Tumor size (T)T1: 36 (45%)T2: 40 (50%)T3: 4 (5%)Tumor typeDuctal: 58 (72.5%)Lobular:17 (21.25%)Mixed:2 (2.5%) Other:3 (3.75%)Lympovascular invasionPresent: 49 (61.25%)Absent:31 (38.75%) Histological GradeG1: 9 (11.25%)G2: 54 (67.5%)G3: 17 (21.25%)Multifocality:Present:48 (60%)Absent:32 (40%) Average SLNs removed (range)3.7 (1-11) Average SLNs positive (range)1.8 (1-7) Average non-SLNs positive (range)3.5 (1-16) Max. size of metastasesAverage:5.2mmMicrometastases: 13 (16.25%)Macrometastases: 67 (83.75%)Her-2 statusPositive:10 (12.5%)Negative:70 (87.5%) Prevalence of non-sentinel node metastasis in patient series31 patients (39%) Extranodal extensionPresent:26 (32.5%)Absent:54 (67.5%) Results: In our cohort, the average number of SLNs removed was 3.7 (1-11) and the average number of positive SLNs was 1.8 (1-7). 13 of the 80 patients had micrometastases within the SLNs (16.2%). 31 of the 80 patients had at least one positive non-SLN (39%) and the average number of additional positive nodes found in the ALND was 3.5 (1-16). For our cohort, all of the nomograms had a similar predictive accuracy (MD Anderson AUC = 0.69, Helsinki AUC = 0.69, Cambridge AUC = 0.63). All nomograms fell short of an acceptable AUC value (at least 0.8). Variables included in the different predictive models testedVariablesMD AndersonHelsinkiCambridgeTumor size++ Tumor type+ Nuclear grade +Lymphovascular invasion++ Number of positive SLNs++ Number of negative SLNs + Number of SLNs+ Proportion of positive SLNs +Extranodal extension++ Max. size of SLN metastases+++Her-2 status + Prevalence of non-sentinel node metastasis in patient series + Multifocality + Conclusions: Our study found that none of the chosen nomograms could reliably predict the presence of positive non-SLNs in our cohort. One reason for this may be the fact they utilize different methods of pathological assessment of nodes and different surgical techniques to those used in our centre. In view of this it may be desirable to produce a more accurate nomogram for our population based on specific variables. Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr P1-01-16.