Comparative prognostic value of lymph node metastases and primary tumor quantitative variables in breast cancer

Primary tumor features are quite able to predict prognosis of breast cancer patients, but yet fail to be 100% accurate. The aim of this study was to compare the prognostic value of mitosis counts and nuclear morphometric variables derived from lymph node metastases and primary tumor proliferation variables and to evaluate their additional prognostic value in a group of 52 lymph node positive breast cancer patients with long-term follow-up. All lymph nodes containing metastases were considered for all patients (in total 304). In all these lymph node metastases, mitoses counts and nuclear morphometry were performed, calculating mean, s.d., minimum, maximum and range of the different variables per patient. In addition, mitoses counts were performed in the primary tumor and the previously established Multivariate Prognostic Index (MPI) was calculated based on the primary tumor mitoses count, tumor size and lymph node status. In univariate survival analysis, the lymph node metastases features s.d. (p < 0.01, Mantel-Cox 9.1) and range (p < 0.02, Mantel-Cox 7.6) of the mitoses counts in the lymph node metastases showed prognostic value. The primary tumor mitoses counts (p < 0.03, Mantel-Cox 4.9) and MPI (p < 0.02, Mantel-Cox 5.9) were significant as well. None of the nuclear morphometric lymph node metastases variables showed prognostic value, neither were the classic variables number and percentage of positive lymph node metastases nor the capsular penetration prognostically significant. In multivariate survival analysis, the maximum of the lymph node metastases mitoses count appeared to have additional prognostic value to the MPI. In conclusion, mitoses counts derived from axillary lymph node metastases have additional prognostic value in lymph node positive breast cancer patients. They may therefore serve to understand the behavior of distant metastases which, at the time of diagnosis, are unknown.