Mobilized peripheral blood has been shown to be a suitable source of hematopoietic progenitor cells for autologous transplantation in oncologic patients. However, tumor cell contamination can potentially occur, although to a lesser extent than in the bone marrow. CD34+ cell positive selection has been developed as a system for the ex vivo purging of remaining tumor cells when used in mobilized peripheral blood. This method has shown a lower purification potential than that obtained with bone marrow or cord blood. The reason for this is not clear, but different groups have tried to improve the purity and yield of the positive selection on mobilized peripheral blood by predepletion of nonlymphoid cell populations, since they can induce nonspecific binding. The present study was designed to test an indirect immunomagnetic CD34+ cell selection method to make it reproducible, feasible, and effective for purging mobilized peripheral blood. Twenty-nine samples from mobilized peripheral blood were tested. The median starting CD34+ percentage was 0.8 (0.3%-4.2%), and the median final purity was 87% (32.7%-99.7%), with a median yield of 44.8% (15%-83.5%). The highest purity was reproducibly achieved when the starting percentage of CD34+ cells was higher than 0.65% (median purity 93.7, range 81%-99.7%, CV 6%) on samples obtained from patients primed with chemotherapy alone or chemotherapy plus recombinant human granulocyte-colony stimulating factor. No relation was found between the content of contaminating nucleated cells and the final CD34+ cell purity. This method showed a depletion capacity, assessed by PCR on samples contaminated with K562 leukemic cells, of about 3 logs. These results indicate that this indirect immunomagnetic method can produce high purity CD34+ cell fractions from mobilized peripheral blood with a high efficiency of tumor depletion.