Correlation of hematopoietic progenitor cell count determined by the SE-9000тДв automated hematology analyzer with CD34+ cell count by flow cytometry in leukapheresis products.код для вставкиСкачать
American Journal of Hematology 67:42–47 (2001) Correlation of Hematopoietic Progenitor Cell Count Determined by the SE-9000™ Automated Hematology Analyzer With CD34+ Cell Count by Flow Cytometry in Leukapheresis Products Keon Uk Park,1* Sang Hee Kim,2 Cheolwon Suh,2 Shin Kim,2 Sun Jong Lee,2 Jung Sun Park,2 Hwa Jung Cho,2 Kang Wook Kim,2 Keehyun Lee,2 Hyo Jung Kim,2 Jinny Park,2 Young Joo Min,2 Jeong Gyoon Kim,2 Taewon Kim,2 Je Hwan Lee,2 Sung Bae Kim,2 Sang We Kim,2 Kyoo Hyung Lee,2 Jung Shin Lee,2 Woo Kun Kim,2 Chan Jeong Park,3 and Hyun Sook Chi3 1 Department of Medicine, Dongguk University College of Medicine, Kyongju, Korea Department of Medicine, University of Ulsan College of Medicine, Asan Medical Center, Seoul, Korea 3 Department of Clinical Pathology, University of Ulsan College of Medicine, Asan Medical Center, Seoul, Korea 2 The yield of stem cell collection after mobilization is crucial for autologous peripheral blood stem cell (PBSC) transplantation. Quantitative determinations of CD34+ cells using flow cytometry or stem cell culture have been used, but these methods require much time, technical experience, and expensive reagents. The automated hematology analyzer (Sysmex SE-9000™, TOA, Japan) equipped with the Immature Information (IMI) channel for immature myeloid cells can detect IMI+ cells within 90 sec. Detection is made possible by the combination of a special reagent system and direct current/radiofrequency biosensors. We studied the relation of IMI+ cells and variable cell counts with CD34+ cell yield in autologous stem cell harvest. In a series of 32 patients (median age, 44 years; M:F = 11:21), 184 leukaphereses were performed after mobilization regimens with chemotherapy and G-CSF or G-CSF alone. Full blood cell counts were enumerated on peripheral blood (PB) samples taken prior to each leukapheresis. Mononuclear cell (MNC) and IMI+ cell counts by automated hematology analyzer and flow cytometry based CD34+ cell yield were measured on the harvested product. The relationship among PB white blood cells (WBC), PB monocytes, IMI+ cells, MNC, and CD34+ cell yield in a single leukapheresis was estimated by Pearson correlation analysis. PB WBC count showed no correlation with CD34+ cell yield in a single leukapheresis (r = 0.02, P = 0.81). PB monocyte count showed a weak correlation (r = 0.21, P = 0.01) and MNC in harvest also showed a weak correlation (r = 0.36, P = 0.0001) with CD34+ cell yield. In contrast, CD34+ cell yield correlated well with IMI+ cell count (r = 0.68, P = 0.0001), and data could be fitted by a linear regression equation, y = 0.330 + 0.974x. IMI+ cell assay by the automated hematology analyzer correlated well with the CD34+ cell yield in a mobilized autologous stem cell harvest. The IMI+ cell count might be used as a simple and efficient indicator of blood stem cell mobilization and collection. Am. J. Hematol. 67:42–47, 2001. © 2001 Wiley-Liss, Inc. Key words: peripheral blood stem cell; flow cytometry; CD34+ cells; automated hematology analyzer INTRODUCTION Peripheral blood stem cell (PBSC) transplantation is now used extensively to provide rapid and durable hematopoietic reconstitution after high-dose chemotherapy for malignant disease [1,2]. Mobilization techniques include the administration of hematopoietic growth factor after chemotherapy . It is widely accepted that the CD34+ cell count is generally a good predictor of the rate of engraftment. A number of studies have demonstrated © 2001 Wiley-Liss, Inc. that reinfusion of a stem cell dose of ⱖ5 × 106 CD34+ cells/kg increases the probability of rapid engraftment, leading to reduced use of supportive measures, short hos*Correspondence to: Keon Uk Park, M.D., Department of Medicine, Dongguk University College of Medicine, 1090-1 Sukjang-Dong, Kyongju, Kyongbuk 780-350, South Korea. E-mail: firstname.lastname@example.org Received for publication 10 March 2000; Accepted 4 October 2000 Technique: Correlation of Hematopoietic Progenitor Cell Count pital stays, and decreased costs . Although the yields of PBSC after mobilization are crucial for engraftment, no conventional method to identify the presence of PBSC in the recovered mononuclear cell fraction has been available. Quantitative CD34+ analysis performed by flow cytometry is used as a guide to determine a target number of progenitor cells needed for timely hematopoietic recovery. The best established method to identify the quantity of PBSC is CFU-GM assay . However, these methods require time (flow cytometry, 2 hr; CFU-GM assay, 14 days), technical expertise, and expensive reagents. The automated hematology analyzer (Sysmex SE9000, TOA Medical Electronica Co. Ltd, Japan) is equipped with a special detection unit for immature white blood cells (WBC), called the Immature Information (IMI) channel. Detection of immature WBC is made possible by the combination of a special reagent system and direct current/radiofrequency biosensors. The reagent system contains detergents, which are capable of lysing more mature white blood cells because of their higher membrane lipid content while immature cells remain relatively intact. Because various types of immature WBC react differently to the reagent, they also occupy distinct areas on the bivariate matrix of the IMI scattergram. Thus it is possible to distinguish blast cells from less immature WBC and, in turn, immature granulocytes such as myelocytes and metamyelocytes from left-shifted neutrophils [6,7]. This technique, which does not require any pretreatment of the blood sample or specific operator skills, can be carried out with a 90-sec turnaround time to completion. We study here the correlation of IMI+ cell count determined by the SE-9000™ automated hematology analyzer with CD34+ cell by flow cytometry in PBSC harvesting products. Also, the relationships of WBC and monocyte counts in peripheral blood with CD34+ cells by flow cytometry were evaluated. MATERIALS AND METHODS Patients Between December 1998 and October 1999, 32 patients [breast cancer, 14; non-Hodgkin’s lymphoma (NHL), 11; multiple myeloma (MM), 5; ovarian cancer, 1; and medulloblastoma, 1] following at our institution underwent PBSCs harvesting after chemotherapy: either cyclophosphamide alone (2–4 g/m2) on Day 1 or other specific tumor-oriented chemotherapies. After chemotherapy, the first dose of G-CSF (5–10 g/kg/day) was given on Day 7 or 8. All patients received the dose of G-CSF administered subcutaneously until the last harvesting day of PBSC. PBSC collection was started when white cell (WBC) recovery reached 10 × 109/l or the monocyte count reached 1 × 109/l. Leukocyte subsets 43 were monitored frequently (every 1–3 days) from the start of mobilization until the end of the leukapheresis procedure. Methods Leukapheresis was performed with continuous flow blood cell separator (Fenwal CS-3000 plus, Baxter, USA). Venous access was established by central venous catheter. Anticoagulant, consisting of heparin at 10 U per ml of ACD-A, was infused at a ratio of 1 ml of anticoagulant to 30 ml of whole blood. Inlet flow rate was maintained at 50–80 ml per minute in the small-volume leukapheresis procedure and 80–100 ml per minute in the large-volume leukapheresis procedure. The collection rate was maintained at 1–2 ml per minute. A set volume of 14 l per leukapheresis was used. The criteria for adequate PBSC collection was a target number 3 × 106 CD34+ cells/kg. Leukapheresis was continued daily in an attempt to achieve that goal. The leukocyte count in the sample was determined with an automated hematology analyzer (Sysmex SE9000, TOA Medical Electronica Co. Ltd, Japan). Differential counts were done microscopically on Giemsastained cell smear. The mononuclear cell count was obtained by multiplying the number of leukocyte with the sum of the percent of lymphocytes and monocytes from the differential count. The CD34+ cells were enumerated by flow cytometry (FACScan威 Becton Dickinson, Fullerton, CA). Leukapheresis products were used without Ficoll-Hypaque centrifugation. The cell suspensions were stained with the following antibody combination5: CD34 FITC + CD14 PE and clgG1 FITC (negative isotope control) + CD14 PE. In addition, a sample was stained with CD45 FITC as a marker for leukocytes. For staining, 30 l of cell suspension was incubated with each 20 l of the monoclonal antibody combinations. After being washed, the remaining red blood cells were lysed by adding 1 ml of ammonium chloride lysis buffer for 6 min at room temperature in the dark. Then, the cells were washed, resuspended and examined with a FACScan flow cytometer (Becton Dickinson). They were then analyzed by the FACScan research software (Becton Dickinson). The detection of immature granulocyte is possible through the IMI channel with the automated hematology analyzer (Sysmex SE-9000, TOA Medical Electronica Co. Ltd, Japan). The lysis reagent used for the IMI channel (Stromatolyser-IM) contains a polyoxyethyleneseries nonionic surfactant and sulfur-containing amino acid, both for fixation of blood cell cytoplasm and membrane, and an anionic surfactant for reduction of erythrocyte ghosts by damaging the red cell membrane. In the IMI channel, the polyoxyethylene nonionic surfactant first damages blood cells, with different degrees of dam- 44 Technique: Park et al. age to different blood cell types. The normal mature granulocyte contains lipids with an amount about 2 times greater than that in the lymphocyte. This is because the lipid content is nearly proportional to the cell size. For both the normal mature granulocyte and lymphocyte, the phospholipid content is about 35% and the cholesterol count is about 10%. In contrast, the immature granulocyte has a lower cholesterol content than the mature granulocyte, and its phospholipid composition has a relatively higher ratio of phosphatidylcholine and a lower ratio of sphingomyelin. The difference in the degree of damage among the blood cell types is attributable to the difference in lipid contents and compositions. When exposed to polyoxyethylene nonionic surfactant, the mature granulocytes cell membrane becomes damaged, resulting in an exposed nucleus due to elution of its intracellular components. The juvenile granulocytes cell membrane also becomes damaged; however, before its intracellular components are eluted, the polyoxyethylene nonionic surfactant and sulfur-containing amino acid enters it through its damaged cell membrane site and will eventually fix its cell membrane and intracellular components. In this process, the sulfur-containing amino acid acts as a protector for the cell against the action of the surfactant. The juvenile granulocyte is thus fixed while retaining the intact cell membrane and cytoplasm. The cationic surfactant then reduces erythrocyte ghosts and mature leukocyte size, facilitating discrimination of juvenile granulocyte from erythrocyte ghosts and mature leukocytes with exposed nucleus, which are then differently measured. Statistical Analysis The clinical and laboratory datas were retrieved from the transplant data and analyzed using the SAS system (SAS Institute, Cary, NC). The relationship between the number of CD34+ cells and IMI+ cells was estimated by linear regression and correlation analysis. Also, the relationships with WBC and monocyte count in peripheral blood and harvest products were evaluated. The Pearson rank correlation was used to evaluate the relationships. A significant level of P < 0.05 was chosen. The receiver operating characteristic (ROC) curve was used to set the cutoff point of IMI+ cells when more than 1 × 106 CD34+ cells/kg were collected in leukapheresis product. RESULTS Patients’ Characteristics Table I presents patients characteristics. A total of 184 PBSC components were collected from 32 patients (male:female ⳱ 11:21). The median age was 44 years. The median number of aphereses was 5 (range, 3 to 9). TABLE I. Characteristics of Patients Median age (range) Sex (male/female) Diagnosis Breast cancer NHL MM Medulloblastoma Ovarian cancer Mobilization method Chemotherapy + G-CSF/GM-CSF G-CSF/GM-CSF only Numbers of leukaphereses (total) Breast cancer NHL MM Medulloblastoma Ovarian cancer 44 (18–65) 11/21 14 11 5 1 1 30 2 184 111 49 21 3 10 The median count of collected CD34+ cells was 4.04 × 106 cells/kg. The 11 non-Hodgkin’s lymphoma patients were classified in accordance to the International Lymphoma Study Group classification. Three patients had diffuse large B-cell lymphoma, 2 had follicular center lymphoma, 3 had lymphoblastic lymphoma, 1 had angioimmunoblastic lymphoma, 1 had peripheral T cell lymphoma, and 1 had Burkitt’s lymphoma. The median number of leukaphereses was 3.5 (range, 3 to 7). The median count of collected CD34+ cells was 3.88 × 106 CD34+ cells/kg. Five multiple myeloma patients received four cycles of the vincristine–doxorubicine–dexamethasone (VAD) regimen before mobilization. The median number of leukaphereses was 3 (range, 2 to 6). The median count of collected CD34+ cells was 6.18 × 106 cells/kg. One medulloblastoma and one ovarian cancer patient were included. Correlation Between Peripheral Blood WBC, Monocytes, MNC in Apheresis Products, and CD34+ Cell Yield Most patients in our study underwent serial procedures, and the median WBC count in the blood for patients treated with a chemotherapy-containing regimen was 16.0 × 109/l (range, [1.0–55.5] × 109/l). The median number of CD34+ cells in a leukapheresis product was 0.37 × 106 CD34+ cells/kg. The WBC counts in peripheral blood (PB) showed poor correlation with CD34+ cell yield (r ⳱ 0.02, P ⳱ 0.81) (Fig. 1). The median monocyte counts in the peripheral blood was 0.63 × 109/l (range, [0–4.2] × 109/l). The PB monocyte counts showed a weak correlation with CD34+ cell yield (r ⳱ 0.21, P ⳱ 0.01) (Fig. 1). The median MNC in a leukapheresis product was 1.37 × 108 cells/kg (range, [0.14– 3.63] × 108 cells/kg). There was also a weak correlation between MNC in leukapheresis product and CD34+ cell yield (r ⳱ 0.36, P ⳱ 0.0001) (Fig. 2). Technique: Correlation of Hematopoietic Progenitor Cell Count 45 Fig. 1. (A) Relationship between the peripheral blood WBC (X axis) and CD34+ cells × 106 per kg of patient in apheresis products (Y axis). (B) Relationship between the peripheral blood monocyte (X axis) and CD34+ cells × 106 per kg of patient in apheresis products (Y axis). Fig. 2. (A) Relationship between the mononuclear cell (MNC) × 108 per kg of patient (X axis) and CD34+ cells × 106 per kg of patient in apheresis products (Y axis). (B) Relationship between the IMI+ cell counts (X axis) and CD34+ cells × 106 per kg of patient in apheresis products (Y axis). The relationship is described by the regression equation y = 0.330 + 0.974x. Correlation Between IMI+ Cells and CD34+ Cell Yield The median IMI+ cell count was 0.32 × 109/l (range, [0–14.72] × 109/l). A significant correlation between the IMI+ cells and CD34+ cells in a leukapheresis product (r ⳱ 0.68, P ⳱ 0.0001) was found, and the data could be fitted into a linear regression equation, y ⳱ 0.330 + 0.974x (Fig. 2). To set the cutoff point, we used the receiver operating characteristic (ROC) curve. The ROC curve graphically portrays the trade-offs involved between either a test’s sensitivity or its specificity. An ideal test is one that reaches the upper left corner of the graph (100% sensitivity and 100% specificity). We graphed sensitivity as a function of [1 − specificity]; this latter quantity is sometimes called the false-positive rate. The ROC curve showed that the best cutoff point for high sensitivity and specificity was 0.66 × 109/l (Fig. 3). When we selected the cutoff value as 0.7 × 109/l, the sensitivity and specificity to obtain more than 1 × 106 CD34+ cells/kg were 83.3% and 83.1% (Table II), respectively. DISCUSSION To make PBSC transplantation a cost effective procedure, it is necessary to optimize the conditions for priming, collection, storage, and engraftment of the leukapheresis products. Cytotoxic chemotherapy with or without hematopoietic growth factors has been used to mobilize hematopoietic progenitors into the peripheral blood, and it is known that chemotherapy and growth factor act synergistically to increase the number of hematopoietic progenitors . The more advanced the mobilization schemes become, the more important are the precise determination of the optimal timing and frequency for harvesting. Through optimal harvesting time prediction, it is possible to spare the patient from poten- 46 Technique: Park et al. Fig. 3. Receiver operating characteristic (ROC) curve for IMI+ cell counts to obtain more CD34+ cells in apheresis products. TABLE II. Possibility of Obtaining More Than 1 × 106 CD34+ Cells/kg of Patient in Apheresis Products When Selected Cutoff Point Is 0.7 × 109/l. ⱕ0.7 × 109 IMI+ cell/l >0.7 × 109 IMI+ cell/l ⱕ1 × 106 CD34+ cells/kg >1 × 106 CD34+ cells/kg 103 (83.06%) 21 (16.94%) 9 (16.67%) 45 (83.33%) tial complications of leukapheresis and reduce treatment costs. The optimal timing and frequency for leukapheresis have not been established on a well-designed study base. It is difficult to make a decision on when leukapheresis should be started and how many times it will be done. Many collection centers use WBC or monocyte count in peripheral blood as a predictor for the timing of leukapheresis, delaying collection until the WBC count exceeds 3 × 109 or 5 × 109 or even 10.0 × 109/l [13–15]. As shown in this study, our center started leukapheresis when WBC recovery reached 10 × 109/l or the monocyte count reached 1 × 109/l. Based on our analysis, WBC or monocyte counts could not predict the number of CD34+ cells in the leukapheresis products. The minimum threshold of CD34+ cells is achieved in most patients, but an optimal collection usually requires at least three harvests. Prior chemotherapy might adversely affect mobilization. Optimal timing of collections is important, because as few as two leukapheresis procedures are enough to obtain an adequate progenitor cell dose. We retrospectively analyzed the datas from 184 PBSC collections to study the predictive value of IMI+ cell count in the peripheral blood. The pluripotent stem cell has yet to be identified and isolated positively. Indirect laboratory assays provides limited information about the hematopoietic potential of collected hematopoietic progenitor cells. Mononuclear cell numbers, although easily determined, are not reliable, and colony culture assays, although reliable, remain poorly standardized, time-consuming, and are difficult to interpret for clinical purposes. Because culture methods cannot be evaluated for 2 weeks or longer, these assays are generally most helpful for retrospective analysis or for quality control of peripheral blood stem cell collection. Quantitative CD34 analysis may be quickly performed by flow cytometry. A number of CD34+ cell counts and CFU-GM are closely correlated, and the number of CD34+ cells in the leukapheresis components is associated with engraftment kinetics [8–10]. Quantitative CD34+ cells analysis showed a significant positive correlation with CFU-GM, and the number of infused CD34+ cells correlates with hematopoietic recovery . CD34+ cell counts by flow cytometry are not standardized and usually are not available for the next 24 hrs. In addition, flow cytometry measurements have required at least 2 hr and expensive reagents. It has caused prolonged apheresis and placed economic burdens on patients. It is possible that hematopoietic progenitor cells were included in IMI+ cells that were simply detected by an automated hematology analyzer (Sysmex SE-9000, TOA Medical Electronica Co. Ltd, Japan) with a usual detection technique, and the total infused IMI+ cells, like the CD34+ cell count, also provides a useful indication of recovery of hematopoietic function in PBSC transplantation. The pluripotent stem cell, capable of differentiation as well as self-renewal and ultimately responsible for all hematopoietic function, is currently presumed to be present in small numbers primarily in the bone marrow. We studied the possibility of a simple and easy identification of the stem cells in leukapheresis products using a conventional automated blood cell counter with the function of white cell differentiation. The SE-9000 (TOA Medical Electronics Co. Ltd, Japan) analyzes the five normal WBC populations using radiofrequency (RF) and direct current (DC) detection methods, as well as separating of eosinophil and basophil channels. The IMI channel method reduces the waiting time for analysis results of stem cells: from 2 hr with flow cytometry down to 90 sec. The analysis can be performed on a routine hematology analyzer, no special operator skills are required, and the results are available anytime of the day. No special pretreatment of the sample is required except when the total leukocytes concentration exceeds 5 × 109/ l, and then a simple dilution step is required. The test may be used to determine the appropriate timing for leu- Technique: Correlation of Hematopoietic Progenitor Cell Count kapheresis and may also be used to monitor the effectiveness of hematopoietic progenitor cell harvest procedures while in progress. To improve the timing of leukapheresis, we are studying the correlation between the IMI+ cell count in peripheral blood and CD34+ cell yield in leukapheresis products. The information provided by the IMI channel appears to be useful in determining the optimal timing of PBSC harvest because it includes semiquantitative datas. REFERENCES 1. Kessinger A, Armitage JO, Landmark JD, Smith DM, Weisenburger DD. Autologous peripheral hematopoietic stem cell transplantation restores hematopoietic function following marrow ablative therapy. Blood 1988;71:723–727. 2. Nademanee A, Sniecinski I, Schmidt GM, Dagis AC, O’Donnell MR, Snyder DS, Parker PM, Stein AS, Smith EP, Molina A, Stepan DE, Solmo G, Margolin KA, Woo D, Niland JC, Forman SJ. High-dose therapy followed by autologous peripheral blood stem cell transplantation for patients with Hodgkin’s disease and non-Hodgkin’s lymphoma using unprimed and granulocyte colony stimulating factor mobilized peripheral blood stem cell. J Clin Oncol 1994;12:2176–2186. 3. Siena S, Bregni M, Brando M, Ravagnani F, Bonadonna G, Gianni AM. Circulation of CD34+ hematopoietic stem cells in the peripheral blood of high-dose cyclophosphamide treated patients: enhancement by intravenous recombinant human granulocyte-macrophage colonystimulating factor. Blood 1989;74:1905–1914. 4. Weaver CH, Hazelton B, Birch R, Palmer P, Allen C, Schwartzberg L, West W. An analysis of engraftment kinetics as a function of the CD34 content of peripheral blood progenitor cell collection in 692 patients after the administration of myeloablative chemotherapy. Blood 1995; 86:3961–3969. 5. Douay L, Gorin NC, Mary JY, Lemarie E, Lopez M, Naiman A, Stachowiak J, Giarratana MC, Bailluo C, Salmon C. Recovery of CFU-GM from cryopreserved marrow and in vivo evaluation after autologous bone marrow transplantation are predictive of engraftment. Exp Hematol 1986;14:358–365. 6. Takahisa Y, Katsushiro T, Noriyuki T. Possibility of identification of hematopoietic stem cells using a conventional blood cell counter. Eur J Haematol 1995;55:207–208. 47 7. Katsuhoro T, Takahisa Y, Kenichi S, Masayuki H, Noriyuki T. Identification of hematopoietic stem cells by the SE-9000™ automated hematology analyzer in peripheral blood stem cell harvest samples. Acta Hematol 1997;98:54–55. 8. Ho AD, Gluck S, Germond C, Sinoff C, Dietz G, Maruyama M, Corringham RE. Optimal timing for collection of blood progenitor cells following induction chemotherapy and granulocyte-macrophage colony-stimulating factor for autologous transplantation in advanced breast cancer. Leukemia 1993;7:1738–1746. 9. Bensinger WI, Appelbaum FR, Rowley S, Storb R, Snaders J, Lilleby K, Gooley T, Demirer T, Schiffman K, Weaver C. Factors that influence collection and engraftment of autologous peripheral-blood stem cells. J Clin Oncol 1995;13:2547–2555. 10. Passos-Coelho JL, Braine HG, Davis JM, Huelskamp AM, Schepers KG, Ohly K, Clarke B, Wright SK, Noga SJ, Davidson NE. Predictive factors for peripheral-blood progenitor-cell collections using a single large-volume leukapheresis after cyclophosphamide and granulocytemacrophage colony-stimulating factor mobilization. J Clin Oncol 1995;13:705–714. 11. Pettengell R, Testa NG, Swindell R, Crowther D, Dexter TM. Transplantation potential of hematopoietic progenitor cells released into the circulation during the routine chemotherapy for non-Hodgkin’s lymphoma. Blood 1993;82:2239–2248. 12. Zimmerman TM, Lee WJ, Bender JG, Mick R, Williams SF. Quantitative CD34 analysis may be used to guide peripheral blood stem cell harvest. Bone Marrow Transplant 1995;15:439–444. 13. Watts MJ, Sullivan AM, Jamieson E, Pearce R, Fielding A, Devereux S, Goldstone AH, Linch DC. Progenitor-cell mobilization after lowdose cyclophosphamide and granulocyte colony-stimulating factor: an analysis of progenitor cell quantity and quality and factors predicting for these parameters in 101 pretreated patients with malignant lymphoma. J Clin Oncol 1997;15:535–546. 14. Dreger P, Marquardt P, Haferlach T, Jacobs S, Mulverstedt T, Eckstein V, Suttorp M, Loffler H, Muller-Ruchhlotz W, Schmitz N. Effective mobilization of peripheral blood progenitor cells with “Dexa-BEAM” and G-CSF: timing of harvesting and composition of the leukapheresis product. Br J Cancer 1993;68:950–957. 15. Schots R, Riet IV, Damiaens S, Flament J, Lacor P, Staelens Y, Steenssens L, van Camp B, De Waele M. The absolute number of circulating CD34+ cells predicts the number of hematopoietic stem cells that can be collected by apheresis. Bone Marrow Transplant 1996;17:509–515.