Peripheral blood is increasingly used instead of bone marrow as a source of hemopoietic precursor cells for transplantation. The optimal technique still needs to be defined. Selection of CD34+ cells in transplant material may be of benefit in allogeneic and autologous peripheral blood precursor cell transplantation (PBPCT), since it allows elimination of unwanted CD34-negative cells, such as T-cells and contaminating tumor cells. We have evaluated the feasibility of CD34 selection in PB transplants and studied hemopoietic reconstitution after autologous transplantation of CD34 selected precursor cells. Between August 1994 and June 1995 CD34 selection was performed on 12 transplants for 9 patients with malignant disease (non-Hodgkin lymphoma [n = 5]; Ewing sarcoma [n = 1]; chronic lymphocytic leukemia [n = 1]; breast cancer [n = 1]; multiple myeloma [n = 1]). PBPC were collected with a Fenwall CS 3000 harvester after stimulation with G-CSF. For selection of CD34+ cells the Ceprate LC34 system (CellPro) was used. A median CD34 purity of 73% (range 40-94%) was achieved. The median number of CD34 positive cells per transplant was 4.8 x 10(6)/kg body weight (range 0.7-15.8). The median number of colony forming cells per transplant was 31 x 10(4)/kg body weight (range 1.5-131.3). For autologous PBPCT the minimal number of CD34 positive cells required in the transplantate was arbitrarily set at 1.0 x 10(6)/kg body weight. This number was achieved in 10 of the 12 transplants. The median loss of CD34+ cells during selection was 1.5 x 10(6)/kg body weight (range 0.2-6.4). In 2 patients the total number was reduced to below the critical value of 1.0 x 10(6)/kg. 7 of the 9 patients received the CD34 selected transplant after intensive chemotherapy and irradiation. The median follow-up time after PBPCT was 196 days (range 62-278). All 7 patients are now alive and with normal hemopoietic function. A granulocyte count above 0.5 x 10(9)/l and a platelet count above 20 x 10(9)/l was achieved on day 14 (median), and on day 19 after PBPCT. We conclude that CD34 selection is technically feasible and that CD34 selected cells can be used for PBPCT. The procedure is time consuming and expensive; it requires complex organization at laboratory level, and the benefit of CD34 selection with regard to T-cell depletion and tumor purging still needs to be proven. However, CD34+ selection is likely to open new perspectives in transplantation medicine.