Minimal tumor contamination of hematopoietic harvests from breast cancer patients can be easily detected by liquid culture assay.

Background: Recurrence after PBSC transplantation in breast cancer (BC) patients may be related to the reinfusion of tumor cells contaminating the graft. We have developed a liquid culture (LC) method for the identification of viable epithelial tumor cells in PBSC collections.

Methods: Mononuclear fraction from PBSC harvests of BC patients undergoing high dose chemotherapy (HDC) (adjuvant setting n = 60, metastatic disease n = 30) were seeded in petri dishes containing round cover slips. Cells were cultured for 3 weeks, then cover slips were stained with the pan-cytokeratin A45-B/B3 mAb and scored under a light microscope. Samples were considered positive when more than one adherent cell or a cluster of cells staining bright red was present. Results were compared with those obtained on cytospins prepared directly from the PBSC harvest. Specificity of the method was tested on lymphoma patients, collections: all were negative. The sensitivity, evaluated by serial dilutions of CG5 BC cell line, was 1 epithelial cell in 10(6) mononuclear cells.

Results: The percentage of positivity was superimposable in the two groups (adjuvant 25%, metastatic 24%). However, a significantly higher proportion of positive samples from metastatic vs adjuvant patients has shown the presence of tumor clusters (86% vs 33%, p = 0.063). In 21% of all samples a discrepancy with the results obtained by immunocytochemical analysis (ICC) was found, mostly due to liquid-culture-positive/ICC-negative PBSCs.

Discussion: Our data suggest that LC assay may enhance the identification of viable disseminated epithelial tumor cells in PBSC grafts and might provide insights about their growth capacity.