PROGNOSTIC-SIGNIFICANCE OF TUMOR-CELL DET ECTION IN COMPARISON TO NODAL STATUS IN 1026 PATIENTS WITH PRIMARY BREAST-CANCER

Consensus about risk-adapted systemic therapy of node negative breast cancer patients remains difficult because of the lack of reliable prog nostic factors. Approximately 30% of nodal-negative breast cancer pati ents will relapse within 10 years and 10-20% of those with distant met astases will be lymph node-negative at the time of surgery. Evaluation of potentially new prognostic factors, therefore, have to concentrate on this issue. Tumor cell detection (TCD) in bone marrow indicates th e presence of tumor cell shedding in malignant disease. In a prospecti ve study we examined TCD immunocytologically in 1026 women who underwe nt primary breast cancer surgery at the Women's Hospital of the Univer sity of Heidelberg (1985-1995). Bone marrow aspiration at two sites on each anterior iliac crest was performed immediately after surgery und er general anesthesia. Most patients received some type of systemic ad juvant treatment. The monoclonal antibody 2 E11, directed against the breast-mucin TAG 12, was used to detect tumor cells in bone marrow sam ples. The statistical analysis of follow-up data included Kaplan-Meier estimation of survival curves, logrank tests and multivariate Cox reg ression analysis with stratification by adjuvant therapy. TCD was posi tive in 53% of 480 node positive and in 31% of 546 node-negative patie nts. TCD correlated significantly with larger tumors (p < 0.001), lymp h node involvement (p < 0.001), undifferentiated tumor grade (p < 0.00 6), and postmenopausal status (p = 0.018). After a median follow-up of 42 months tumor cell positive patients showed a reduced relapse free and overall survival (p(0.001). Multivariate Cox regression analysis s howed that TCD was an independent prognostic factor superior to lymph node status, tumor stage, and grading. Especially in tumors less than 2 cm diameter TCD was the most powerful predictor of outcome (TCD: p < 0.001, RR 12.3; Nodal status: p = 0.096, RR 2.34). Tumor cell detecti on in bone marrow of breast cancer patients is a highly significant pr ognostic factor with a negligible morbidity (in comparison to nodal ax illary dissection) and is the best marker of dissemination in node neg ative disease. Prospective randomized studies should be performed to d etermine whether axillary node dissection might be replaced by bone ma rrow aspiration and TCD in a defined subgroup of patients with small t umors.