Autologous peripheral blood stem cell transplantation for acute myelogenous leukemia

Transcription

1 Bone Marrow Transplantation, (1997) 20, Stockton Press All rights reserved /97 $12.00 Autologous peripheral blood stem cell transplantation for acute myelogenous leukemia H Gondo 1, M Harada 2, T Miyamoto 1, K Takenaka 1, K Tanimoto 1, S Mizuno 1, T Fujisaki 1, K Nagafuji 3, S Hayashi 3, T Eto 3, S Taniguchi 4, K Akashi 3, N Harada 4, K Yamasaki 4, T Shibuya 4, E Matsuishi 5, Y Ohno 6, S Makino 1, Y Takamatsu 1, M Murakawa 1, T Teshima 6, Y Hirota 7, T Okamura 1, N Kinukawa 8, S Inaba 9 and Y Niho 1 1 First Department of Internal Medicine, 8 Department of Medical Informatics, 9 Blood Transfusion Service, Kyushu University; 2 Second Department of Internal Medicine, Okayama University Medical School; 3 Department of Hematology, Harasanshin Hospital; 4 Department of Internal Medicine, Hamanomachi Hospital; 5 Department of Internal Medicine, Saga Koseikan Hospital; 6 Department of Internal Medicine, Kitakyushu Munincipal Hospital; and 7 Department of Internal Medicine, Matsuyama Red Cross Hospital, Japan Summary: with acute myelogenous leukemia (AML). 1,2 However, most will relapse despite consolidation and maintenance The safety and efficacy of myeloablative therapy followed chemotherapy. Therefore, post-remission therapy is one of by autologous peripheral blood stem cell trans- the most important issues in the treatment of AML. An plantation (ABSCT) for acute myelogenous leukemia increasing number of AML patients have been treated with (AML) were evaluated in 60 patients. Peripheral blood intensified consolidation chemotherapy, autologous bone stem cells (PBSC) were collected during recovery after marrow transplantation (BMT), or allogeneic BMT. 3,4 Allo- consolidation chemotherapy. High-dose chemotherapy geneic BMT is most successful for reducing leukemic consisting of busulfan (16 mg/kg), etoposide (40 mg/kg), relapse by virtue of the high-dose chemoradiotherapy con- and cytosine arabinoside (3 g/m 2 4) (BEA regimen) ditioning regimen and graft-versus-leukemia (GVL) was used for pretransplant conditioning in 13 patients. effects. 5 Nevertheless, in addition to the requirement of For the remaining 47 patients, granulocyte colony- HLA-identical marrow donors, transplant-related morbidity stimulating factor (G-CSF) was administered concurrently and mortality remain the major obstacles to improving with the BEA regimen during conditioning. clinical results of allogeneic BMT. In autologous BMT, Unpurged, cryopreserved PBSC containing a median regimen-related toxicities are fewer, although a GVL effect number of MNC/kg or CFU-GM/kg cannot be obtained. One randomized study has demonstrated were reinfused at transplantation. The median number that patients treated with autologous BMT have of days to granulocytes exceeding 500/ l and last plate- superior disease-free survivals (DFS) to those receiving let transfusion were 15 (8 44) and 24 (0 180), respectively. intensive chemotherapy alone. 3 The 3-year probabilities of disease-free survival Recently, autologous peripheral blood stem cell trans- (DFS) and relapse were 78.6 and 21.4% for patients plantation (ABSCT) has begun to replace autologous BMT transplanted in first remission, 29.6 and 64.4% for those because of rapid engraftment and low transplant-related in second or third remission, and 11.1 and 77.8% for mortality However, high relapse rates have been those in relapse, respectively. There were no transplant- reported after ABSCT for AML Moreover, Reiffers et related deaths within 100 days of transplantation. Age, al 12 have demonstrated no difference in DFS or survival disease status at transplantation, and number of induc- rate between ABSCT and chemotherapy alone. We now tion chemotherapies to first complete remission were report very encouraging results of ABSCT for AML in risk factors affecting the outcome of ABSCT. These first remission. results of ABSCT for AML in first remission warrant a prospective study of ABSCT as post-remission therapy. Keywords: autologous; peripheral blood stem cells; Materials and methods transplantation; acute myelogenous leukemia Patients Recent advances in intensive chemotherapy and supportive care can produce complete remission in 60 80% of patients Correspondence: H Gondo, The First Department of Internal Medicine, Faculty of Medicine, Kyushu University, 3-1-1, Maidashi, Higashi-ku, Fukuoka 812, Japan Received 28 February 1997; accepted 30 June 1997 Between May 1989 and July 1996, we evaluated 60 consecutive patients with de novo AML who received myeloablative therapy and ABSCT. There were 35 males and 25 females with a median age of 40.5 years and a range of years (Table 1). The distribution of morphologic types of AML according to the FAB classification was as follows; 11 patients were M1, 24 M2, seven M3, eight M4, seven M5 and three M6. Informed consent was obtained from all patients or responsible family members.

2 822 Table 1 Patient characteristics Disease status at transplant Total 1CR 2CR/3CR Relapse No. of patients Median age (range, years) 40 (17 66) 36 (13 54) 49 (23 65) 40.5 (13 66) Male/female 24/18 4/5 7/2 35/25 FAB subtype M M M M M M No. of induction courses to remission 1 course courses WBC at diagnosis / l / l Karyotype 46,XX; 46,XY t(8;21), t(15;17), inv(16) Others G-CSF administration on mobilization Yes No Pretransplant conditioning G-CSF combined BEA Non-G-CSF combined BEA Median interval from diagnosis to ABSCT (months, range) (4 20) (12 54) (4 22) (4 54) Median No. of MNC infused ( 10 8 /kg, range) ( ) ( ) ( ) ( ) Median No. of CFU-GM infused ( 10 4 /kg, range) ( ) (1.5 40) ( ) ( ) CR = complete remission; G-CSF = granulocyte colony-stimulating factor; BEA = busulfan 16 mg/kg orally, etoposide 40 mg/kg intravenously, and cytosine arabinoside 3 g/m 2 intravenously every 12 h for 2 days; ABSCT = autologous peripheral blood stem cell transplantation; MNC = mononuclear cells; CFU-GM = colony-forming unit granulocyte macrophage. Chemotherapy and collection of peripheral blood stem Consolidation chemotherapy consisting of Ara-C cells (PBSC) 500 mg/m 2 intravenously every 12 h for 6 days combined with mitoxantrone 7 mg/m 2 intravenously for 3 days in the Figure 1 shows the treatment protocol used in this study. first course and etoposide 100 mg/m 2 intravenously for 5 Induction chemotherapy for all but AML/M3 patients condays in the second course, was begun after achieving comsisted of daunorubicin (DNR) 45 mg/m 2 intravenously for plete remission. 14 Collection of PBSC was performed with 3 days and cytosine arabinoside (Ara-C) 100 mg/m 2 by cona continuous blood-cell separator during hemopoietic tinuous infusion for 7 days. Bone marrow aspiration was perrecovery after each round of consolidation chemotherapy formed on day 8, and if the marrow was not severely hypoplwith or without G-CSF (filgrastim, 200 g/m 2 intraastic, DNR was added on days 8 and 9, and Ara-C venously). Unpurged PBSC harvests were cryopreserved by administration was continued for 3 more days (days 8, 9 and a simplified method without rate-controlled freezing until 10). For AML/M3 patients, all-trans retinoic acid was admintransplantation. 15 Granulocyte macrophage progenitor cells istered according to the protocol described elsewhere. 13 (colony-forming unit granulocyte macrophage, CFU-GM) Induction CR Consolidation No.1 Consolidation No.2 DNR/Ara-C or ATRA PBSC harvest No.1 PBSC harvest No.2 Conditioning Ara-C/MIT Ara-C/VP16 BEA or BEA/G-CSF ABSCT Figure 1 Treatment protocol. DNR/Ara-C = daunorubicin 45 mg/m 2 i.v., days 1, 2, 3 (8, 9), and Ara-C 100 mg/m 2 continuous i.v., day 1 7 (8 10) for all but acute myelogenous leukemia (AML)/M3 patients. For AML/M3 patients, all-trans retinoic acid 45 mg/m 2 was administered orally. Ara-C/MIT = Ara-C 500 mg/m 2, i.v., every 12 h for 6 days and mitoxantrone 7 mg/m 2, i.v. for 3 days. Ara-C/VP16 = Ara-C 500 mg/m 2, i.v., every 12 h for 6 days and etoposide 100 mg/m 2, i.v. for 5 days. BEA or BEA/G-CSF = busulfan 16 mg/kg, etoposide 40 mg/kg, and Ara-C 3 g/m 2, every 12 h for 2 days or busulfan 16 mg/kg, etoposide 40 mg/kg, and Ara-C 3 g/m 2, every 12 h for 2 days combined with G-CSF during pretransplant conditioning.

3 in each PBSC harvest were assayed as described pre- Results viously Hemopoietic recovery after ABSCT No patient had engraftment failure. The median number of ABSCT days to granulocytes exceeding 500/ l was 15 (range 8 44), which was not affected by G-CSF administration after ABSCT (Figure 2). The speed of granulocyte recovery was The pretransplant conditioning regimen for 13 patients coninversely correlated with the number of CFU-GM infused sisted of busulfan 4 mg/kg orally on days 8 to 5, etopo- (r = 0.415). Granulocytes exceeded 500/ l within 14 days side 20 mg/kg intravenously on days 4 and 3, and Araafter transplantation in 21 of 37 patients (57%) who were C 3 g/m 2 intravenously every 12 h on days 3 and 2 (BEA autografted with more than CFU-GM/kg, and in regimen). For the remaining 47 AML patients, G-CSF was 18 of 27 patients (67%) with more than CFUcombined with the BEA regimen to increase the chemo- GM/kg. The median time to the last platelet transfusion was sensitivity of leukemic cells to the S-phase-specific anti- 24 days (range 0 180) (Figure 2). Prophylactic platelet leukemic agent, Ara-C. G-CSF was administered at a dose transfusions were given to maintain platelet counts above of 5 g/kg on days 14 to 8, 10 g/kg on days 7 and / l. Platelet recovery was also inversely correlated 6, and 20 g/kg on days 5 and 4 by 6-h intravenous with the number of CFU-GM infused (r = 0.409). Howinfusion in combination with continuous infusion of Ara- ever, 13 of 60 patients (22%) continued to be dependent C 100 mg/m 2 on days 12 to 6, in addition to the above on platelet transfusion for more than 3 months, and seven BEA regimen. 16,17 One day later, cryopreserved PBSC were of these 13 patients (54%) relapsed within 5 months of rapidly thawed in a 37 C waterbath and infused through a transplantation. central venous catheter without washing on day 0. If a sufficient number of CFU-GM had been obtained at the second PBSC collection (CFU-GM /kg), the second Treatment outcome PBSC harvest was used instead of the first harvest for trans- Of 60 patients with AML, 42 received ABSCT in first plantation. G-CSF 200 g/m 2 was administered intrarelapse. remission, nine in second or third remission, and nine in venously from day 1 until full recovery of granulocytes in Among the nine patients transplanted in second or 18 of 60 patients. Complications related to the pretransplant third remission, PBSC collection was done during the first conditioning were classified according to the criteria prothird remission in two, during the second in six, and during the posed by Bearman et al. 18 in one. In all relapsed patients receiving ABSCT, PBSCs were harvested during first remission. The 3-year and 5-year probabilities of DFS for patients transplanted in first remission were 78.6 and 70.7%, Statistical analysis respectively (Figure 3). The median survival time for these patients without relapse was 17 months (range 4 A linear regression analysis was employed for evaluating 73 months). Nine patients relapsed, eight (89%) within relationships between numbers of CFU-GM infused and the 12 months of transplantation. The median interval from speed of hemopoietic recovery after transplantation. 19 Disfor patients who had transplants in first remission. The 5- remission to ABSCT was 5 months (range 3 13 months) tributions of achieving granulocytes 500/ l, achieving platelet transfusion independency, DFS, and relapse were year DFS probability for 20 patients autografted more than estimated by the product-limit method. 20 Comparisons of 6 months after remission was 90.0%, compared with 43.5% these results were based on the log-rank test. DFS was condifference was statistically significant (P 0.05 by log- for 22 patients who had transplants before 6 months. This sidered as the time from transplant to relapse or death due to any cause. Actuarial curves were analyzed as of 1 Novrank test). Of 42 patients autografted in first remission, 17 (40.5%) had favorable cytogenetics. The 5-year DFS probember Cox s regression analysis in a stepwise manner was performed to determine prognostic factors with regard to DFS. 21 Factors used in this analysis included gender, 100 age ( 40 vs 40 years), WBC count at diagnosis ( / vs / l), FAB subtype (M2 or M3 vs M1, M4, M5 or M6), karyotype (favorable vs all others), number of induction chemotherapy cycles to remission (1 50 course vs 2 courses), G-CSF administration for PBSC mobilization (yes vs no), G-CSF administration after transplantation (yes vs no), disease status at transplantation (first remission vs beyond first remission), and pretransplant conditioning 0 (G-CSF combined BEA vs non-g-csf combined Days after transplantation BEA). Analysis was carried out using a statistical package, BMDP 2L, on an IBM system 3090 computer. Figure 2 Probability of achieving granulocytes 500/ l (, n = 60) and platelet transfusion independency (------, n = 60). A median MNC/kg or CFU-GM/kg were reinfused at transplantation. Probability (%) 823

4 824 Probability of DFS (%) G-CSF administration for PBSC mobilization or after transplantation, and pretransplant conditioning, did not affect the outcome of ABSCT for AML. Toxicities Oral mucositis and gastrointestinal symptoms such as nausea, vomiting, and diarrhea, were common toxicities related 0 to the BEA or G-CSF combined BEA regimen. The incidence of regimen-related toxicities greater than grade II Years after ABSCT was as follows: oral toxicity, 23%; gastrointestinal toxicity, Figure 3 Kaplan Meier estimates of DFS for AML patients transplanted 7%; hepatic toxicity, 3%; and cardiac toxicity, 7%. There in first remission (, n = 42), in second or third remission (-----, n = 9), was no early transplant-related mortality (within 100 days). and in relapse (, n = 9). The median follow-up time of AML patients transplanted in first remission was 17 months. First remission vs However, one patient transplanted in second remission second/third remission, P 0.01; first remission vs relapse, P developed severe veno-occlusive disease (VOD) on day 18 and died of hepatic failure on day 137. Prognostic factors Multivariate analyses revealed that age, disease status at transplantation, and the number of induction chemotherapies to first complete remission were significant risk factors influencing DFS (Table 2). AML patients who were younger than 40 years old, autografted in first remission, and achieved first remission with only one course of induc- tion chemotherapy, showed better DFS. Factors such as gender, WBC count at diagnosis, FAB subtype, karyotype, Probability of relapse (%) Years after ABSCT Figure 4 Kaplan Meier estimates of relapse for AML patients trans- planted in first remission (, n = 42), in second or third remission (-----, n = 9), and in relapse (, n = 9). First remission vs second/third remission, P 0.01; first remission vs relapse, P Discussion abilities for those with favorable cytogenetics and for the remaining 25 patients were 93.3 and 55.6%, respectively (P = 0.06 by log-rank test). The 3-year probabilities of DFS for patients autografted in second or third remission, and in relapse were 29.6 and 11.1%, respectively (Figure 3). Two of nine patients transplanted in second or third remission and one of nine patients in relapse are alive in complete remission. The 3-year probability of relapse for AML patients transplanted in first remission was 21.4% (Figure 4). In contrast, the 3-year probabilities of relapse for patients transplanted in second or third remission and in relapse were 64.4 and 77.8%, respectively. Differences in relapse rates were sig- nificant between patients autografted in first remission vs second or third remission, and between those in first remission vs relapse (P 0.01 for both). The present study confirmed rapid granulocyte recovery and low transplant-related mortality in patients with AML treated with myeloablative chemotherapy and ABSCT. Pla- telet recovery was slightly delayed; 13 of 60 patients (22%) required platelet transfusions for more than 3 months. Because the speed of hemopoietic recovery is inversely cor- related with the number of PBSC infused, 14 this delay was probably due to the relatively low numbers of CFU-GM infused as compared to other studies. 7 Early relapse was observed in seven of these 13 patients, another factor which may have delayed platelet recovery. The remaining six patients are alive without relapse and eventually became independent of platelet transfusions. There was no transplant-related mortality within 100 days after transplantation in our study, although early severe hepatic toxicity developed in one patient who was autografted in second remission. Fatal VOD in this patient may have been related to the repeated intensive chemo- therapy prior to transplantation. An increased incidence of VOD has been linked to pretransplant conditioning with busulfan in allogeneic BMT. 22,23 Sanz et al 9 similarly reported fatal VOD in two of 24 patients who received ABSCT for AML using a busulfan-containing regimen. The use of busulfan in these patients needs to be evaluated in prospective studies. ABSCT has several advantages over autologous BMT, including rapid engraftment and low transplant-related mortality. However, the collective data indicate that relapse rates are as high as 60% after ABSCT for AML, 6,9,10 and early relapse is frequent. 7,8 This may occur because the ratio of leukemic clonogenic progenitors to normal stem cells is higher in PBSC harvests than in marrow 8 or because a large number of malignant cells are reinfused since the cell dose is higher in ABSCT than in autologous BMT. 7 However, a retrospective study has demonstrated that the risk of relapse and the likelihood of disease-free survival after ABSCT for AML are similar to those seen after autologous unpurged BMT. 6,10 Reiffers et al 24 demonstrated that the source of stem cells did not affect the outcome of patients undergoing autologous transplantation for AML

5 Table 2 Autologous PBSCT for AML Prognostic factors influencing disease-free survival (DFS) selected by stepwise Cox s regression analysis a 825 Variables Coefficient Standard error Odds ratio P value Age b No. of induction courses to CR c Disease status at transplant d a Disease-free survival (DFS) was defined as the time from transplantation to relapse or death due to any cause. b 40 vs 40 years. c 1 course vs 2 courses; CR = complete remission. d First remission vs beyond first remission. and that the use of PBSC was not associated with an containing intermediate-dose Ara-C in the present study increased risk of relapse. may be responsible for in vivo purging. AML patients transplanted in first remission had a favor- Lastly, intensification of pretransplant conditioning by able DFS in our study. The 5-year probability of DFS was cytokine priming may have improved DFS in our study. G- more than 70%. Possible explanations include a low inci- CSF can stimulate the proliferation of myeloid leukemic dence of transplant-related mortality which may contribute cells because most express receptors for G-CSF. 29 An to DFS. Although we did not encounter engraftment failure increased susceptibility of leukemic cells to Ara-C in the or early transplant-related mortality, most studies have presence of G-CSF has been demonstrated in vitro. 17 More- reported the incidence of transplant-related mortality to be over, clinical studies support a potential role for G-CSF in 5 10%. 6,9,25 In addition, patient selection may have affected the pretransplant conditioning regimen to reduce leukemic the outcome of our study, since 22 of 42 patients (52.4%) relapses after ABSCT or allogeneic BMT. 16,17 A prospective autografted in first remission were classified as M2 or M3 randomized study will be required to assess further the and 17 of these (40.5%) had favorable cytogenetics such efficacy of G-CSF priming for pretransplant conditioning. as t(8;21), t(15;17), or inv (16); the proportion of patients There are few reports describing the prognostic factors with these subtypes is higher than in other studies. 6,9 How- for DFS and relapse after ABSCT for AML. Korbling et ever, multivariate analyses did not indicate that FAB subtype al 6 have suggested that FAB M5 is a prognostic factor. In or karyotype were significant factors for DFS. In a autologous BMT, on the other hand, a WBC / lat recent report by Mehta et al, 7 AML patients treated with diagnosis, the need for more than two courses of induction ABSCT with biphenotypic or dysplastic features often had chemotherapy to achieve remission, and certain cytogenetic early relapse; although such patients were not included in findings have been shown to be adverse prognostic fac- the present study, three other patients with biphenotypic tors. 3,25 In the present study, age, disease status at transplantation, leukemia who received ABSCT at our institution relapsed and the number of induction courses to achieve soon after transplantation (data not shown). Therefore, the remission affected DFS, indicating that younger patients biologic characteristics of the leukemic cells should be considered ( 40 years old) with AML who achieved complete as a potent prognostic factor for ABSCT. remission after one course of induction therapy may benefit The timing of PBSC collection may also be important for from ABSCT during the first remission. preventing relapse. Recently, we demonstrated that minimal In summary, the present clinical study has demonstrated residual disease (MRD) in marrow and PBSC harvests the safety and efficacy of myeloablative chemotherapy fol- could be reduced with repeated chemotherapy. 26,27 Based lowed by ABSCT for the treatment of AML. Although the upon these data, we collected PBSC not after induction number of patients was small and follow-up time was short, chemotherapy, but instead during the recovery phase after the encouraging results obtained with this study warrant consolidation. 9,10 Furthermore, when PBSC harvests after a additional prospective studies to evaluate the use of second consolidation course contained sufficient numbers ABSCT in the treatment of AML. of PBSC, we preferentially used that harvest because it potentially contained fewer leukemic cells than the harvest after the first consolidation course. This strategy of in vivo purging has been used to reduce the leukemic burden in Acknowledgements the setting of autologous BMT and has led to an improved This work was supported in part by grants-in-aid from the Minis- DFS. 25,28 Mehta et al 28 reported that adequate consolidation try of Health and Welfare, the Ministry of Education, Science and of remission before autologous BMT was the most Culture ( , Cancer Research No. 7 3), and the Uehara important factor associated with continuing remission after Memorial Foundation. autologous BMT. Stein et al 25 have conducted a prospective study of autologous BMT for AML in first remission, in which marrow was collected after consolidation chemotherapy containing high-dose Ara-C. They demonstrated References that the probabilities of DFS at 24 months were 49% in the 1 Gale RP, Foon KA. Therapy of acute myelogenous leukemia. intention-to-treat group and 61% in patients who actually Semin Hematol 1987; 24: underwent autologous BMT. 25 Consolidation chemotherapy 2 Kobayashi T, Miyawaki S, Tanimoto M et al. Randomized

6 826 trials between behenoyl cytarabine and cytarabine in combi- for cryopreservation of peripheral blood stem cells at 80 C nation induction and consolidation therapy, and with or without without rate-controlled freezing. Bone Marrow Transplant ubenimex after maintenance/intensification therapy in 1991; 8: adult acute myeloid leukemia. J Clin Oncol 1996; 14: Takamatsu Y, Teshima T, Akashi K et al. Successful second 213. autologous blood stem cell transplantation after G-CSFcombined 3 Zittoun RA, Mandelli F, Willemze R et al. Autologous or allogeneic conditioning for the treatment of high-risk acute bone marrow transplantation compared with intensive myelogenous leukemia. Bone Marrow Transplant 1994; 13: chemotherapy in acute myelogenous leukemia. New Engl J Med 1995; 332: Takahashi S, Okamoto S, Shirafuji N et al. Recombinant 4 Mayer RJ, Davis RB, Schiffer CA et al. Intensive postremission human glycosylated granulocyte colony-stimulating factor chemotherapy in adults with acute myeloid leuke- (rhg-csf)-combined regimen for allogeneic bone marrow mia. New Engl J Med 1994; 331: transplantation in refractory acute myeloid leukemia. Bone 5 Horowitz MM, Gale RP, Sondel PM et al. Graft-versus-leuke- Marrow Transplant 1994; 13: mia reactions after bone marrow transplantation. Blood 1990; 18 Bearman SI, Appelbaum FR, Buckner CD et al. Regimen- 75: related toxicity in patients undergoing bone marrow transplantation. 6 Korbling M, Fliedner TM, Holle R et al. Autologous blood J Clin Oncol 1988; 6: stem cell (ABSCT) versus purged bone marrow transplan- 19 Cox DR. Regression models and life-tables. J R Stat Soc [B] tation (pabmt) in standard risk AML: influence of source 1972; 34: and cell composition of the autograft on hemopoietic reconsti- 20 Kaplan EL, Meier P. Nonparametric estimation from incomplete tution and disease-free survival. Bone Marrow Transplant observations. J Am Stat Assoc 1958; 53: ; 7: Hopkins A. Survival analysis with covariate 2L. In: Dixon 7 Mehta J, Powles R, Singhal S, Treleaven J. Peripheral blood WJ (ed). BMDP Statistical Software Manual. University of stem cell transplantation may result in increased relapse of California Press: Berkley, 1990, pp acute myeloid leukaemia due to reinfusion of a higher number 22 Morgan M, Dodds A, Atkinson K et al. The toxicity of busul- of malignant cells. Bone Marrow Transplant 1995; 15: 652 phan and cyclophosphamide as the preparative regimen for 653. bone marrow transplantation. Br J Haematol 1991; 77: Laporte JP, Gorin NC, Feuchtenbaum J et al. Relapse after 534. autografting with peripheral blood stem cells. Lancet 1987; 23 Nevill TJ, Barnett MJ, Klingemann HG et al. Regimen-related 2: toxicity of a busulfan-cyclophosphamide conditioning regimen 9 Sanz MA, Rubia J, Sanz GF et al. Busulfan plus cyclophosphamide in 70 patients undergoing allogeneic bone marrow transplantation followed by autologous blood stem cell transplan- tation. J Clin Oncol 1991; 9: for patients with acute myeloblastic leukemia in first 24 Reiffers J. Incidence of relapse following blood stem cell complete remission: a report from a single institution. J Clin transplantation for acute myeloid leukemia in first remission. Oncol 1993; 11: Bone Marrow Transplant 1996; 17: Reiffers J, Korbling M, Labopin M et al. Autologous blood 25 Stein AS, O Donnell MR, Chai A et al. In vivo purging with stem cell transplantation versus autologous bone marrow high-dose cytarabine followed by high-dose chemoradiother- transplantation for acute myeloid leukemia in first complete apy and reinfusion of unpurged bone marrow for adult acute remission. Int J Cell Clon 1992; 10: myelogenous leukemia in first complete remission. J Clin 11 Demirer T, Petersen FB, Bensinger WI et al. Autologous Oncol 1996; 14: transplantation with peripheral blood stem cells collected after 26 Nagafuji K, Harada M, Takamatsu Y et al. Evaluation of leu- granulocyte colony-stimulating factor in patients with acute kaemic contamination in peripheral blood stem cell harvests myelogenous leukemia. Bone Marrow Transplant 1996; 18: by reverse transcriptase polymerase chain reaction. Br J Haematol 1993; 85: Reiffers J, Stoppa AM, Michallet M et al. Allogeneic vs autologous 27 Miyamoto T, Nagafuji K, Harada M et al. Quantitative analywith stem cell transplantation vs chemotherapy in patients sis of AML1/ETO transcripts in peripheral blood stem cell acute myeloid leukemia in first remission: the BGMT 87 harvests from patients with t(8;21) acute myelogenous leu- study. Leukemia 1996; 10: kaemia. Br J Haematol 1995; 91: Kanamaru A, Takemoto Y, Tanimoto M et al. All-trans retinoic 28 Mehta J, Powles R, Singhal S et al. Autologous bone marrow acid for the treatment of newly diagnosed acute promye- transplantation for acute myeloid leukemia in first remission: locytic leukemia. Blood 1995; 85: identification of modifiable prognostic factors. Bone Marrow 14 Teshima T, Harada M, Takamatsu Y et al. Cytotoxic drug and Transplant 1995; 16: cytotoxic drug/g-csf mobilization of peripheral blood stem 29 Shimoda K, Okamura S, Harada N et al. Granulocyte colony- cells and their use for autografting. Bone Marrow Transplant stimulating factor receptors on human acute leukemia: 1992; 10: biphenotypic leukemic cells possess granulocyte colony-sti- 15 Makino S, Harada M, Akashi K et al. A simplified method mulating factor receptors. Cancer Res 1992; 52: