The presence of epithelial cells in bone marrow may be a prognostic factor
in breast cancer, and so we evaluated their evolution in treated and untrea
ted patients. A first bone marrow aspirate was obtained from 125 stage I/II
breast cancer patients at diagnosis and repeated every 6-8 months; the sam
ples were processed for leukocyte separation, used to prepare cytospin slid
es, stained with a pool of monoclonal antibodies (MoAb) recognising epithel
ial antigens, and immunocytochemically processed. The median follow-up was
48 months (range 15-82); 23 patients relapsed, and 14 died. MoAb positive c
ells were observed in 31.2% of first, 24.3% of second, and 27.8% of third a
spirates. In 68/100 pairs of successive aspirates, bone marrow status remai
ned unchanged; in 20 it became negative, and in 12 positive (not statistica
lly significant even after adjusting for adjuvant therapy). An analysis bas
ed on Mantel and Byar's approach to time-dependent covariates using all 225
aspirates found no statistically significant prognostic difference between
the patients with negative and positive bone marrow. Bone marrow status ch
anged over time in about 1/3 of the patients; adjuvant therapy did not affe
ct the probability of its becoming negative or positive. No significant ass
ociation was found between bone marrow evolution and relapse or death, but
the relatively high probability of a change in status over time cannot excl
ude the possibility that a positive aspirate during the course of breast ca
ncer may be a negative prognostic factor.