AIMS

The aim of this study was to prospectively investigate metastatic pathways of spread to lymph node versus bone marrow and identify biological characteristics that determine these patterns in early invasive breast cancer.

PATIENTS AND METHODS

In all, 177 patients with early invasive breast cancer underwent surgical extirpation of the primary tumour with sentinel lymph node biopsy (SLNB). Bone marrow (BM) aspiration was performed to screen for cytokeratin-positive cells by immunocytochemistry. Lymphatic spread was assessed by histopathological examination of lymph nodes (LN). A representative subset of 87 tumours was analysed by tissue microarray (TMA) to evaluate expression of markers that potentially influence haematogenous vs. lymphatic spread. Patients were followed up for a median of 54.7 months.

RESULTS

Of the 177 patients, 114 (64%) were BM-/LN-, 38 (22%) BM-/LN+, 19 (11%) BM+/LN- and 6 (3%) BM+/LN+. Multivariate analysis of histopathological characteristics revealed that increasing tumour size was significantly associated with both LN positivity (p = 0.003) and BM positivity (p = 0.01), the presence of lymphovascular invasion significantly correlated with LN+ (p = 0.01), whereas lower histological grade was significantly associated with BM+ (p = 0.03). LN+ and BM+ were non-significantly negatively related to each other. Univariate analysis of the TMA data showed differential expression patterns for several factors; significant differences between effects on the two metastatic pathways (lymphatic vs. haematogenous) were found for expression of CD54 (p = 0.03), osteopontin (p = 0.04), bone sialoprotein (p = 0.04) and CXCR4 (p = 0.009). High expression of CD54, osteopontin and bone sialoprotein (BSP) was positively associated with BM + but was either not associated, or negatively associated, with LN+. High CXCR4 expression was positively associated with LN+ and negatively with BM+. High VEGF-C expression was associated with both LN+ and BM+, although this did not attain statistical significance. Due to the small number of clinical events during clinical follow-up, no associations were identified between metastatic spread patterns, recurrence and/or death.

CONCLUSION

These findings suggest that distinct lymphatic and haematogenous metastatic pathways exist in early breast cancer and that these pathways are governed by specific biological markers.