PD02-07: Models Predicting Non-Sentinel Node Involvement in Breast Cancer Also Predict for Regional Recurrence If the Axilla Is Not Treated.

Abstract

Abstract Background Series on breast cancer patients with sentinel node (SN) isolated tumor cells or micrometastases show low recurrence rates in the presence of SN isolated tumor cells, and varying recurrence rates in the presence of SN micrometastases. Non-SN prediction models are frequently used as a decision aid to identify patients that may not need axillary treatment, but this still needs to be validated in respect to regional recurrence. Patients and Methods We followed a cohort of 486 patients in the MIRROR-study of whom none received axillary treatment with favorable primary tumor characteristics and pN0(i+)(sn) or pN1(mi)(sn) for a median of 5 years. The patients were categorized into low or high-risk subgroups based on several published non-SN prediction models (cut-off levels between models varied from 3 to 10%), i.e. the Tenon Scoring system, MSKCC nomogram, Stanford nomogram and a Dutch model (Bolster et al.). The primary endpoint was 5-year regional recurrence-free survival. Results The overall 5-year regional recurrence rate was 3.0% (+/− 0.1%). Using the Tenon scoring system, we identified 438 patients with a low risk score of 3.5 or lower with a 5-year regional recurrence rate of 2.3% (+/− 0.8%), compared to a recurrence rate of 10.1% (+/− 0.4%) in 48 patients with a score above 3.5. For the MSKCC nomogram a low risk score of 0.10 or lower identified 300 patients with a 5-year recurrence rate of 2.8% (+/− 1.1%), and a score above 0.10 identified 166 patients with a recurrence rate of 3.4% (+/− 1.5%) (20 patients not assessable). By the Stanford nomogram a low risk cutoff level of 10% identified 21 patients without recurrence, whereas 465 patients had a 3.2% (+/−0.9%) recurrence rate. Using a Dutch model, a low risk cutoff score of 20 discriminated between 384 patients with a 5-year recurrence rate of 2.2% (+/− 0.8%) and 102 patients with a recurrence rate of 6.3% (+/− 2.9%). Further analyses with different cut-off values and subgroups will be presented at the conference. Conclusion Using several published non-SN prediction models for predicting regional recurrence, the Tenon scoring system outperformed the other models in that it identified the largest subgroup of patients with a low recurrence rate. We would recommend axillary treatment in patients classified as high risk according to the Tenon score. Funding: Netherlands Organization for Health Research and Development (ZonMw 945-06-509). Citation Information: Cancer Res 2011;71(24 Suppl):Abstract nr PD02-07.