Cytokeratin-positive cells in PBSC collections from normal donors and patients with non-epithelial cell-derived tumors.

BACKGROUND: It is possible that post-transplant relapse in patients with breast cancer may result, in part, from residual tumor in the autologous PBSC product. It is unclear from the literature what effect residual breast tumor cells have on clinical outcome and whether purging tumor cells would be beneficial. We hypothesized that lack of standardization of assays for detection of residual breast tumor may be responsible for the inconclusive clinical data. METHODS: We compared two assays routinely for detection of cytokeratin (CK)-positive cells in stem-cell grafts, immunohistochemistry (IHC) and flow cytometry (FCM). The patient population consisted of individuals with breast cancer, non-epithelial cell-derived tumors and normal donors. A rigorous statistical model was developed for evaluation of the data. RESULTS: We found that the IHC assay out-performed the FCM assay. Importantly, both assays detected CK-positive cells in PBSC collections of patients with non-epithelial cell-derived tumors and in normal donors. No distinguishing morphological characteristics could be identified to differentiate potentially malignant from non-malignant CK-positive cells. Due to the inability to distinguish true positive from false positive results, we developed a statistical model to establish a quantitative threshold to discriminate positive from negative samples. Among the patients tested, no clinical outcome differences were detected, regardless of where the threshold of CK-positive cells was set. DISCUSSION: We conclude the more stringent criteria and more specific markers, rather than the presence or absence of CK-positive cells, need to be established to determine the clinical significance of minimal residual disease in autologous breast-cancer