3493

Background: Detection of disseminated tumor cells (DTC) in bone marrow (BM) has prognostic value in primary breast cancer (BC) and is currently under investigation for therapy stratification and monitoring. Anti-cytokeratin (CK) antibodies (Abs) have been established as markers for detection of DTC in patients with BC. Here we performed a direct comparison of the two mAbs most frequently used for DTC detection over the past 10 years.

Methods: We screened BM samples of 351 node-negative BC patients. All patients were free of any signs of overt metastases (stage M0, pT1-2). The population was characterized by smaller primary tumors and more favorable clinical outcome than in other, comparable studies. Tumor cells (TC) in BM were enriched by Ficoll-Hypaque and the mononuclear cell (MNC) fraction was spun on slides. For each patient 2*106 MNC were analyzed per test, using A45-B/B3 (A45) and AE1/AE3 (AE) Abs in parallel. A45 staining was, however, performed 5-9 years later than AE staining, on cytospins stored at -80°C. All samples were screened automatically (ACIS™, Chromavision) for the presence of CK-positive cells. By strict morphological evaluation only TC-compatible cells were scored positive. In addition, we performed double staining on 20 selected BM, using the combination of Cy3 directly coupled to A45/ AE visualized by Alexa 488.

Results: For both Abs, the number of DTC revealed was less than 5 per 2*106 BM cells. Of the 347 patients with evaluable A45-results, 19 patients (5.5%) were BM-positive, whereas of the 331 patients with evaluable AE-results 27 patients (8.2%) were positive; concordance obtained by cross-analysis for the two Abs was 86.9%. Immunocytochemical double staining revealed subpopulations of DTC labeled with both Abs and others exclusively stained for either A45 or AE. AE-positive BM were found more frequently in hormone receptor negative (HR-) compared to HR+ patients (13.1% (8/61) vs. 7.1% (18/252)), while an opposite trend was found for A45. The number of A45-positive BM was higher for pT2 tumors than for pT1 (6.3% (5/80) and 4.7% (11/235)). A45-positive patients suffered from systemic relapse significantly earlier than A45-negative patients (26.3% vs. 10.4%, p=0.044, log rank).

Conclusion: These data should be evaluated bearing in mind the overall low number of relapses/deaths in the population analyzed. Preliminary results indicate that two frequently used Abs for DTC detection do not give identical results. They appear to identify at least partially different populations of DTC in BC patients, and the correlation to clinical outcome and biological factors may vary.

98th AACR Annual Meeting-- Apr 14-18, 2007; Los Angeles, CA