Acute leukemia Bone marrow transplantation from partially HLA-mismatched family donors for acute leukemia: single-center experience of 201 patients

Transcription

1 (2004) 33, & 2004 Nature Publishing Group All rights reserved /04 $ Acute leukemia Bone marrow transplantation from partially HLA-mismatched family donors for acute leukemia: single-center experience of 201 patients J Mehta, S Singhal, AP Gee, K-Y Chiang, K Godder, F van Rhee, S DeRienzo, W O Neal, LLamb, and PJ Henslee-Downey Division of Transplantation Medicine, South Carolina Cancer Center, University of South Carolina, Columbia, SC, USA Summary: Between February 1993 and December 1999, 201 patients (1 59 years old, median 23) with acute leukemia (67% not in remission) underwent ex vivo T-cell-depleted (TCD) bone marrow transplants (BMT) from partially mismatched related donors (PMRD; 92% mismatched for 2 3 HLA A, B, DR antigens). Conditioning comprised total body irradiation, cyclophosphamide, cytarabine, etoposide, anti-thymocyte globulin (ATG), and methylprednisolone. Graft-versus-host disease (GVHD) prophylaxis comprised partial TCD with OKT3 (n ¼ 143) or T10B9 (n ¼ 58), steroids, ATG, and cyclosporine. The engraftment rate was 98%. The cumulative incidences of grades II IV acute GVHD and chronic GVHD were 13 and 15%, respectively. The 5-year cumulative incidences of relapse and transplant-related mortality (TRM) were 31 and 51%, respectively. The actuarial 5-year overall survival (OS) and disease-free survival (DFS) probabilities were 19 and 18%, respectively. Patient age 415 years, active disease at transplant, donor age 425 years, and 3-antigen donor mismatch (host-versus-graft) affected the outcome adversely. The actuarial 5-year OS of four groups of patients identified based upon these risk factors was 39, 20, 13, and 0%, respectively (Po001). We conclude that PMRD BMT is a potential treatment option for patients with high-risk acute leukemia who require an alternative donor transplant and fall into a group with a reasonable expected outcome. (2004) 33, doi: /sj.bmt Published online 12 January 2004 Keywords: acute leukemia; HLA mismatch; allogeneic transplantation High-dose chemoradiotherapy with hematopoietic stem cell support is therapy of choice in a significant proportion of Correspondence: Dr J Mehta, Division of Hematology/Oncology, Northwestern University Medical School, 676 North St. Clair Street, Suite 850, Chicago, IL , USA; j-mehta@northwestern.edu Received 03 December 2002; accepted 29 March 2003 Published online 12 January 2004 patients with acute leukemia at some stage of the disease. The ideal stem cell donor is an HLA-identical sibling. For patients without HLA-identical sibling donors, autotransplantation or an allograft from an alternative donor are options. 1 6 Alternative donors include HLA-mismatched family donors (partially mismatched related donors; PMRD), HLA-matched unrelated donors, and cord blood. The clinical application of PMRD transplantation (PMRDT) in practice, despite a relatively long history and the ready availability of donors for the majority of patients, 1 3,6 10 has not been widespread because of the intensive nature of the procedure. However, because of encouraging short-term results reported recently using large doses of T-cell-depleted blood-derived stem cells, 11,12 there has been a resurgence of interest in PMRDT as a therapeutic option for patients requiring alternative donor transplants. In this paper, we have analyzed the outcome of 201 consecutive acute leukemia patients undergoing PMRDT at a single center. Patients and methods All consecutive acute leukemia patients undergoing PMRDT at the Division of Transplantation Medicine of the South Carolina Cancer Center, Palmetto Richland Memorial Hospital and University of South Carolina between February 1993 and December 1999 who fulfilled the following criteria were included: no preceding transplant (patients autografted previously, for the same disease or another disease, were excluded), 13 de novo disease (patients with a known antecedent hematologic disorder such as myelodysplastic syndrome, chronic myeloid leukemia or other myeloproliferative disorders were excluded), and total-body irradiation (TBI)-containing conditioning regimen. Table 1 shows the characteristics of the 201 patients. All patients and donors provided written, informed consent. All research protocols (transplantation as well as supportive care) were approved by the institutional review board. HLA typing Class I and II typing was performed by standard serologic techniques. Additional typing was performed in a proportion of cases using high-resolution techniques (PCR-SSOP). The typing results reported here (Table 2 and the analyses)

2 390 Table 1 reflect serologic typing for HLA-A, B, and DR, because the choice of donors was usually based upon these results. Donor choice Patient and donor characteristics n 201 Gender Male 122 (61%) Female 79 (39%) Age 1 59 years (median 23) Diagnosis Acute myeloid leukemia 102 (51%) Acute lymphoblastic leukemia 99 (49%) Stage Primary refractory 19 (9%) First remission 18 (9%) Second remission 39 (19%) Advanced remission 9 (4%) Relapse 116 (58%) Patient cytomegalovirus serostatus Positive 109 (54%) Negative 92 (46%) Donor age in years (median) 4 67 (28) Donor relationship to patient Sibling 74 (37%) Parent 71 (35%) Offspring 38 (19%) Other 18 (9%) Donor patient gender mismatch No 118 (59%) Yes 83 (41%) Donor cytomegalovirus serostatus Positive 100 (50%) Negative 101 (50%) Donor patient ABO compatibility Match 131 (65%) Mismatch 70 (35%) Table 2 Number of mismatch loci in the host-versus-graft (HVG) (rejection) and graft-versus-host (GV) directions GVH Total HVG (32%) (49%) (16%) (2%) Total 76 (38%) 85 (42%) 34 (17%) 6 (3%) 201 Donors were prioritized on the basis of the greatest HLA matching, cytomegalovirus (CMV) seronegativity for seronegative recipients, younger age, same sex, and better health. 14 Conditioning. regimens and graft-versus-host disease (GVHD) prophylaxis All patients received TBI as part of the conditioning. Additional cytoreductive agents comprised cytarabine, cyclophosphamide, and etoposide. Pre-transplant methylprednisolone and anti-thymocyte globulin (ATG) contributed to immunosuppression to facilitate engraftment (Table 3). Table 4 shows the number of patients receiving each type of regimen. Tables 3 and 4 show the GVHD prophylaxis. The pharmacologic mainstay of GVHD prophylaxis was cyclosporine, methylprednisolone, and ATG (Atgam, Pharmacia, Kalamazoo, MI, USA). The marrow was depleted of T cells using the monoclonal antibodies T10B9 ( ) or OKT3 ( ) using techniques described previously. 6,15,16 Acute and chronic GVHD were diagnosed and staged based on standard criteria, and treated with combination immunosuppressive therapy In addition to corticosteroids, cyclosporine, and ATG, etanercept (soluble tumor necrosis factor-a), 20 daclizumab (anti-cd25 monoclonal antibody), thalidomide, and extracorporeal photopheresis were employed as needed. Supportive therapy Patients were treated in HEPA-filtered single rooms. CMVseronegative patients with CMV-seronegative donors received CMV-screened blood products. All blood products were irradiated. Intravenous immunoglobulin was administered once a week post transplant for 100 days, less frequently between 100 and 180 days, and longer in patients with chronic GVHD. Filgrastim (granulocyte colonystimulating factor; G-CSF) or sargramostim (granulocytemacrophage colony-stimulating factor; GM-CSF) were administered to accelerate myeloid recovery post transplant. All patients received systemic antimicrobial agents prophylactically: usually a fluoroquinolone (ciprofloxacin or levofloxacin), an antiviral (acyclovir or valaciclovir), and a triazole antifungal (fluconazole or itraconazole). 21 Patients remained on antifungal and antiviral agents, penicillin V, and pneumocystis prophylaxis for as long as they were on immunosuppression. Statistical analysis The w 2 test was used to compare categoric variables and the Wilcoxon rank-sum test was used to compare continuous variables. The probabilities of disease-free survival (DFS) and overall survival (OS) were estimated by the Kaplan Meier method, and compared using the log-rank test. Cumulative incidence of TRM and relapse was estimated using each type of event as a competing risk for the other. 22,23 Cumulative incidence of acute and chronic GVHD was estimated using TRM and relapse as competing risks for each. The significance of differences in TRM and relapse was calculated using the likelihood-ratio statistic for proportional-hazards regression models. All follow-up data are current as of August 2000.

3 Table 3 Conditioning regimens and GVHD prophylaxis 391 TBI Etoposide globulin Cytarabine Cyclophosphamide Methylprednisolone Anti-thymocyte globulin TBI-VP16- ARAC-Cy-MP TBI-VP16-ARAC- Cy-MP-ATG 235 cgy b.i.d. 20 mg/kg 2 g/m 2 b.i.d. 50 mg/kg 1 g/m 2 ; days 2 and 0 days 9 to 7 day 6 days 5 to 3 days 2 and 1 1 g/m 2 b.i.d.; day cgy BID 20 mg/kg 2 g/m 2 b.i.d. 50 mg/kg 1 g/m 2 ; days 2 and 0 10 mg/kg days 9 to 7 day 6 days 5 to 3 days 2 and 1 1 g/m 2 BID; day 1 days 5 to 3 GVHD prophylaxis Cyclosporine 3 mg/kg from day 1 Methylprednisolone 30 mg/m 2 days 5 16 Anti-thymocyte 10 mg/kg globulin days 5 16 Adjust dose to maintain level between 113 and 168 ng/ml. Tapers off by 6 12 months Taper by 10% every week thereafter Abbreviations: ARAC ¼ cytarabine; ATG ¼ anti-thymocyte globulin; Cy ¼ cyclophosphamide; MP ¼ methylprednisolone; TBI ¼ total-body irradiation; VP16 ¼ etoposide. Table 4 T-cell depletion technique, conditioning regimen, and cellular composition of the graft T-cell depletion technique Conditioning regimen Total nucleated cell dose 10 8 /kg (median) T cell dose 10 5 /kg (median) CD34+ cell dose 10 6 / kg (median) T10B9 OKT3 TBI-VP16-ARAC- Cy-MP TBI-VP16-ARAC- Cy-MP-ATG 58 (29%) 143 (71%) 58 (29%) 143 (71%) Eight missing values () OKT3 group (0.9) a - T10B9 group (1.5) a (3 missing values) (0.5) - OKT3 group () b - T10B9 group 16.1 (0.9) b (109 missing values; no values available in the T10B9 group) 7.6 (1.9) a OKT3 vs T10B9: Po001. b OKT3 vs T10B9: P ¼ 2. Abbreviations: ARAC ¼ cytarabine; ATG ¼ anti-thymocyte globulin; Cy¼ cyclophosphamide; MP ¼ methylprednisolone; TBI ¼ total-body irradiation; VP16 ¼ etoposide. The following factors were analyzed in Cox proportional-hazards regression models for effect on acute GVHD, chronic GVHD, TRM, relapse, DFS and OS: patient age (p15 vs 415), gender, diagnosis (AML vs ALL), disease status at transplant (remission vs active disease), donor (sibling vs other), donor age (p25 vs 425), donor recipient gender (match vs mismatch), donor recipient ABO compatibility (match vs mismatch), CMV serostatus of patient (positive vs negative), CMV serostatus of donor (positive vs negative), number of mismatched antigens in the host-versus-graft (HVG) direction (o3 vs 3), number of mismatched antigens in the graft-versus-host (GVH) direction (o2 vs X2), nucleated cell dose (p1.5 vs /kg vs unknown), T-cell dose ( /kg vs other vs unknown), and T-cell depletion technique (T10B9 vs OKT3). The choice of the cutoff limits for the T-cell dose requires explanation. Very low T-cell doses can compromise engraftment and immune reconstitution, whereas high doses can increase the incidence and severity of GVHD. Therefore, rather than choose high vs low, it was felt that choosing an optimum T-cell dose range would be more appropriate. It should be emphasized that this was only for the purposes of the analysis; in practice, there was no control over the number of T cells infused once the marrow was processed. The categories for the number of mismatched host (GVH) and donor (HVG) antigens were derived from exploratory analyses of the data (details not shown). CD34 þ cell doses were available in half the patients and thus were not included in the Cox model. Results Table 2 shows details of HLA mismatching. In total, 51 patient donor pairs were haploidentical with no shared antigens on the mismatched haplotype. At the other extreme, one nonsibling donor recipient pair was phenotypically identical at HLA-A, B and DR. In all, 92% of the donor recipient pairs was mismatched for at least two antigens in the GVH or HVG directions, and 70% in the GVH as well as HVG directions. Table 4 shows the cell numbers infused. T10B9-treated grafts contained significantly higher numbers of nucleated cells and T cells. CD34 þ cell numbers were not available

4 392 Cumulative incidence Cumulative incidence Figure Days Engraftment (time to a neutrophil count of /l) Years Figure 3 Cumulative incidence of chronic graft-versus-host disease. Cumulative incidence Cumulative incidence 100 of 201 died of treatment-related causes 61 of 201 relapsed Days Figure 2 Cumulative incidence of grades II IV acute graft-versus-host disease Years Figure 4 Cumulative incidence of TRM and relapse. for T10B9-treated grafts. As Figure 1 shows, neutrophil recovery to /l was prompt and 98% of evaluable patients engrafted. Acute GVHD The cumulative incidence of grades II IV acute GVHD was 13% (95% CI: 9 19) (Figure 2). The patients who developed grades II IV acute GVHD had received a significantly higher T-cell dose than the remaining patients (median 10 5 /kg vs /kg; P ¼ 4). However, no factor was found to influence the development of acute GVHD independently in Cox analysis. Chronic GVHD Chronic GVHD was seen in 33 of the patients at risk (surviving at least 80 days). As Figure 3 shows, the cumulative incidence of chronic GVHD was 15% (95% CI: 11 21). In Cox analysis, the use of OKT3 for T-cell depletion was associated with a lower risk of developing chronic GVHD (relative risk 0.1; 95% CI: 2 ; P ¼ 02). Transplant-related mortality The cumulative incidence of TRM at 5 years was 51% (95% CI: 44 58) (Figure 4); with 100 patients dying of treatment-related causes days (median 71) post transplant. The causes of death were as follows: viral infections (n ¼ 18), multiorgan failure (n ¼ 16), pulmonary toxicity (acute respiratory distress syndrome, noninfectious pneumonitis) (n ¼ 15), fungal infections (n ¼ 13), acute GVHD (n ¼ 8), bacterial infections (n ¼ 8), Epstein Barr virus (EBV) þ post transplant lymphoproliferative disease (PTLD) (n ¼ 6), chronic GVHD (n ¼ 5), graft rejection or failure (n ¼ 4), secondary malignancies (n ¼ 2), cardiotoxicity (n ¼ 2), thrombotic microangiopathy (n ¼ 2), and CNS toxicity (n ¼ 1). All except 10 deaths (four chronic GVHD,

5 three EBV, two bacterial infections, and one pulmonary toxicity) occurred within 1 year of transplantation. Relapse As Figure 4 shows, the cumulative incidence of relapse at 5 years was 31% (95% CI: 25 39), with 61 patients relapsing days (median 147) following the transplant. A total of 58 patients died of relapsed disease or consequences of further therapy. The remaining three patients are alive in CR 1071, 1253, and 1789 days after relapse; one each after chemotherapy, an infusion of donor leukocytes, and a second PMRDT. Donor leukocyte infusions These were administered to 62 patients at doses ranging from to CD3 þ cells/kg. The indications were treatment of recurrent disease (n ¼ 17), decreasing the likelihood of relapse (n ¼ 37), and treatment of PTLD (n ¼ 8). Among those treated for recurrent leukemia, one is alive in CR, whereas the rest died of progressive disease or GVHD. Among the 37 receiving donor cells prophylactically, 17 eventually relapsed and 20 did not. Of the latter, seven are alive and 13 experienced TRM (including six from GVHD). EBV þ PTLD PTLD was seen in 11 patients at months (median 4.5). Eight received donor cells as therapy (median /kg CD3 þ cells); two attained CR and one is alive. The other seven died of PTLD (n ¼ 5), acute GVHD (n ¼ 1), or relapse n ¼ 1). All three patients who did not receive donor leukocyte died; but not all from PTLD (one PTLD, one relapse, one CMV infection). Survival At the last follow-up, 43 patients (21%) were alive in remission at 4 83 months (median 58); 40 in continuous remission. The actuarial 5-year probabilities of DFS and OS were 18% and 19% respectively (Figure 5). The actuarial 5-year DFS and OS were 29 and 34% for patients in first or second remission, and 13 and 14% for the remaining patients, respectively (P ¼ 008 for DFS and P ¼ 009 for OS). The actuarial 5-year DFS and OS were 26 and 30% for patients in remission, and 13 and 14% for patients with active disease, respectively (P ¼ 02 for both comparisons). Among the 92 patients in whom infused CD34 þ cell numbers were available, higher numbers were associated with a trend towards superior OS (5-year probability 9% for o /kg vs 19% for X /kg; P ¼ 8) and DFS (5-year probability 9% for o /kg vs 18% for X /kg; P ¼ 8). The OS of the 51 patients who were bidirectionally threeantigen mismatched with their donors was significantly inferior to the remaining patients (13 vs 21% at 5 years; P ¼ 3); a difference that stemmed exclusively from patients with active disease at transplant (3 vs 18% at 5 years; P ¼ 007). None of the 11 patients aged 50 years or Survival Figure 5 more survived; 10 dying of TRM and one of recurrent leukemia. Cox analysis Table 5 shows the results of the Cox analysis. Remission at the time of the transplant, younger donor age, and o3 antigen donor (HVG) mismatch emerged as the three most important favorable variables. Younger patient age, o2 antigen host (GVH) mismatch, and a sibling donor were also associated with superior outcome. The three major factors were combined with patient age to produce four risk groups as follows: A (zero adverse factor or one adverse factor in patients aged p15 years; n ¼ 49), B (one adverse factor in patients aged 415 years; n ¼ 48), C (two adverse factors; n ¼ 82), and D (three adverse factors; n ¼ 22). The outcome of these four groups of patients was significantly disparate with actuarial 5-year survival of 39, 20, 13,and 0%, respectively (Po001). (Figure 6). The disparity in outcome was largely the result of differing TRM (5-year cumulative incidence 28, 68, 49, and 68%); relapse rates were not significantly different (5- year cumulative incidence 38, 15, 37, and 32%). While the infusion of an optimum number of T cells and TCD with T10B9 also affected some outcome measures significantly (Table 5), these factors were not included in deriving risk groups because these were factors over which there was no control: OKT3 was used because T10B9 was not available, and the number of T cells infused depended upon the extent of TCD on an individual donor s marrow. Discussion 43 of 201 alive (OS) 40 of 201 alive and event-free (DFS) Years Overall and disease-free survival. Our data show that PMRDT results in long-term survival of a proportion of patients with acute leukemia. Depending upon the number of risk factors present, the outcome varies substantially (Figure 6). To the best of our knowledge, this is the largest and the most mature reported series of PMRDT up for any disease. 393

6 394 Table 5 Multivariable analysis of prognostic factors Covariate Favorable Adverse Relative risk (95% CI) P TRM Patient age o15 years 415 years 2.1 ( ) 1 Donor Sibling Other 1.7 ( ) 2 Number of donor (HVG) mismatch vectors ( ) 3 Relapse Disease status at transplant Remission Not in remission 4.1 ( ) 001 Number of host (GVH) mismatch vectors 2 3 o2 2.2 ( ) 1 Method of T-cell depletion T10B9 OKT3 2.5 ( ) 2 Donor age o25 years 425 years 1.9 ( 3.6) 5 DFS Disease status at transplant Remission Not in remission 1.7 ( ) 04 Number of donor (HVG) mismatch vectors ( ) 1 T-cell dose Optimum Other 1.6 ( 2.6) 3 Donor age o25 years 425 years 1.4 ( 2.0) 4 OS Disease status at transplant Remission Not in remission 1.7 ( ) 06 Number of donor (HVG) mismatch vectors ( ) 05 Donor age o25 years 425 years 1.5 ( ) 2 Abbreviations: TRM ¼ transplant-related mortality; OS ¼ overall survival; DFS ¼ disease-free survival. Overall survival 0 (21 of 49 alive) 1 (10 of 48 alive) 2 (12 of 82 alive) 3 (0 of 22 alive) Years Figure 6 The effect of the number of risk factors on overall survival (Po001). The adverse factors are active disease at the time of transplant, donor age 425 years, and three-antigen donor (HVG) mismatch. Owing to the favorable effect of young patient age, patients aged p15 years who have one adverse risk factor are included with the no risk factor group. It is critical to separate modifiable (treatment-related) variables from nonmodifiable (biological) variables in order to refine clinical practice and improve outcome. 24 Of the factors identified to be significant here, patient age is clearly a factor over which there is no control. The others however may be modifiable to a variable extent. Based on these data, we would suggest choosing a younger 25 sibling donor with o3 antigen donor (HVG) mismatch, if there was a choice. While a 2 3 antigen host (GVH) mismatch was associated with a lower risk of relapse, this did not impact survival beneficially because of increased TRM in univariate analysis (data not shown). The increase in TRM was not significant in Cox analysis. Our data therefore do not provide any guidance on the number of host mismatches to choose. Based on retrospective analysis of PMRDT data from the International Blood and Marrow Transplant Registry (IBMTR), it may be advisable to choose a sibling mismatched for noninherited maternal antigens when possible, and a mother when only parents are available as donors. 26 We would also suggest identifying appropriate patients without a suitable HLA-identical sibling or wellmatched unrelated 27,28 donor early (ie while the disease is in remission). The number of adverse variables present may be used to determine if PMRDT is indeed appropriate for a given patient or not. For patients with no adverse features (and younger patients with 1 adverse factor), the probability of long-term survival was 39%. This dropped to 20% with 1 adverse risk feature in older patients. Clearly, PMRDT is a reasonable treatment option in patients with zero to one adverse features who need an alternative donor transplant. This group comprised 97 of the 201 patients in this series just under 50%. The long-term survival decreased to 13% in those with two adverse features and to 0% with three adverse features. This suggests that the use of PMRDT should be selective in patients with two adverse features and it should be avoided in patients with three adverse features. All 11 patients aged 50 years and over died; 10 of TRM and one of relapse. This suggests that patients over the age of 50 years should also probably not undergo PMRDT using standard (myeloablative) conditioning regimens; they could be candidates for investigational/nonmyeloablative transplant approaches possibly combined with more refined methods of graft preparation. 29 For patients requiring an allograft, an HLA-identical sibling is the best donor. If a matched sibling is not available, what should be the next preferred option? The question of PMRDT vs unrelated donor transplantation is

7 not an easy one to answer. In addition to HLA-A, B and DRB1, there are data that suggest that matching for HLA C and DQB1 also affects outcome. 27,28 Thus, the definition of a good match in the unrelated donor setting is constantly changing. A patient with a perfectly matched unrelated donor (matched at all 10 alleles) is likely to be best served by an allograft from that donor. Drobyski et al 30 showed recently that the outcome of patients allografted from wellmatched unrelated donors was superior to those transplanted from HLA-mismatched relatives or mismatched unrelated donors. This contrasts with the report by Handgretinger et al, 31 who observed superior outcome after transplantation of large doses of blood-derived CD34 þ cells from haploidentical related donors compared to matched unrelated donors. Taken together, these reports suggest that for patients with less than perfectly matched unrelated donors, PMRDT may be an option if an allograft is essential. Purely from a practical standpoint, family donors are far more readily accessible, for the primary transplant procedure as well as for subsequent cell donations for adoptive immunotherapy. This may be especially important when dealing with a patient suffering from a disease with a rapid tempo where time is of the essence, and it may not be possible to organize a transplant from an unrelated donor. For non-caucasian patients, the chances of finding an unrelated match are still low; making the option of PMRDT a more realistic prospect. Autotransplantation may also be a viable alternative is some patients, 32 because the high relapse rates associated with it are compensated for by its favorable safety profile. This results in the eventual outcome being equal to or better than with PMRDT. 32 Similar data exist showing the equivalence or superiority of autotransplantation to unrelated donor transplantation in first and second remission AML 33 and ALL. 34 The key is identifying patients at the maximum risk of relapse with autotransplantation and offering them alternative-donor allografts, and identifying those at the maximum risk of TRM with alternative-donor allografts and offering them an autograft (if feasible) or nontransplant therapeutic options. There is a suggestion in these data that higher CD34 þ cell doses are better, but this is an issue that requires further elucidation in the context of techniques employed here. CD34 þ cell doses that are considerably higher than this have been employed using T-cell-depleted blood (with or without marrow) as the source of stem cells in order to attain alloengraftment. 11 With limited follow-up, the results of the 43 patients with acute leukemia reported by Aversa et al 11 are comparable to the results we have obtained here. The issue of growth factor administration after alternative-donor transplantation has received some attention, with all the data indicating a negative effect on outcome. This issue cannot be addressed adequately in our population of patients because almost all patients received growth factors post transplant. We conclude that PMRDT results in long-term survival of a proportion of patients with acute leukemia, and may be a reasonable treatment option in selected patients. Acknowledgements This paper is dedicated to the nursing, supportive, and scientific staff of the Division of Transplantation Medicine of the South Carolina Cancer Center, and our patients and their families. This work would not have been possible without their indefatigable enthusiasm and support. In particular, we acknowledge S Abhyankar, S Adams, J Benton, K Bridges, LBryant, S Carpenter, N Christiansen, J Fitzpatrick, A Fontana, B Foster, S Geier, K Goon-Johnson, S Heiney, K Higgins-Smith, C Lee, J McGuirk, R McKnight, W Neglia, M Paoli, A Pati, P Patterson, K Westfall, G Wingo, and D Worthington-White for their efforts and dedication to clinical care, laboratory, and research management during significant portions of the work. References 1 Clift RA, Hansen JA, Thomas ED et al. Marrow transplantation from donors other than HLA-identical siblings. Transplantation 1979; 28: Powles RL, Morgenstern GR, Kay HE et al. Mismatched family donors for bone-marrow transplantation as treatment for acute leukaemia. Lancet 1983; 1: Anasetti C, Amos D, Beatty PG et al. Effect of HLA compatibility on engraftment of bone marrow transplants in patients with leukemia or lymphoma. N Engl J Med 1989; 320: Kernan NA, Bartsch G, Ash RC et al. Analysis of 462 transplantations from unrelated donors facilitated by the National Marrow Donor Program. N Engl J Med 1993; 328: Gluckman E, Rocha V, Boyer-Chammard A et al. Outcome of cord-blood transplantation from related and unrelated donors. Eurocord Transplant Group and the European Blood and Marrow Transplantation Group. N Engl J Med 1997; 337: Henslee-Downey PJ, Abhyankar SH, Parrish RS et al. Use of partially mismatched related donors extends access to allogeneic marrow transplant. Blood 1997; 89: Henslee-Downey PJ. Mismatched bone marrow transplantation. Curr Opin Oncol 1995; 7: Henslee-Downey PJ, Parrish RS, MacDonald JS et al. Combined in vitro and in vivo T lymphocyte depletion for the control of graft-versus-host disease following haploidentical marrow transplant. Transplantation 1996; 61: Henslee-Downey PJ. Allogeneic related partially mismatched transplantation. In: Barrett J, Treleaven JG (eds.) The Clinical Practice of Stem Cell Transplantation. Isis Medical Media: Oxford, 1998, pp Henslee-Downey PJ. Allogeneic transplantation across major HLA barriers. Baillieres Clin Haematol 2002; 14: Aversa F, Tabilio A, Velardi A et al. Treatment of high-risk acute leukemia with T-cell-depleted stem cells from related donors with one fully mismatched HLA haplotype. N Engl J Med 1998; 339: Handgretinger R, Schumm M, Lang P et al. Transplantation of megadoses of purified haploidentical stem cells. Ann NY Acad Sci 1999; 872: Godder KT, Mehta J, Chiang KY et al. Partially mismatched related donor bone marrow transplantation as salvage for patients with AMLwho failed autologous stem cell transplant. Bone Marrow Transplant 2001; 28: Henslee-Downey PJ. Choosing an alternative bone marrow donor among available family members. Am J Pediatr Hematol Oncol 1993; 15:

8 Lee C, Brouillette M, Lamb L et al. Use of a closed system for Vab-positive T cell depletion of marrow for use in partially mismatched related donor (PMRD) transplantation. Prog Clin Biol Res 1994; 389: Collins NH, Gee AP, Henslee-Downey PJ. T-cell depletion of allogeneic bone marrow transplants by immunologic and physical techniques. In: Lasky L, Warkentin P (eds). Stem Cell and Marrow Processing for Transplantation. American Association of Blood Banks: Bethesda, Maryland, Glucksberg H, Storb R, Fefer A et al. Clinical manifestations of graft-versus-host disease in human recipients of marrow from HL-A-matched sibling donors. Transplantation 1974; 18: Martin PJ, Schoch G, Fisher L et al. A retrospective analysis of therapy for acute graft-versus-host disease: initial treatment. Blood 1990; 76: Sullivan KM, Shulman HM, Storb R et al. Chronic graftversus-host disease in 52 patients: adverse natural course and successful treatment with combination immunosuppression. Blood 1981; 57: Chiang KY, Abhyankar S, Bridges K, Godder K, Henslee- Downey PJ. Recombinant human tumor necrosis factor fusion protein (Enbrel) as complementary treatment for chronic graftversus-host disease. Transplantation 2002; 73: Safdar A, van Rhee F, Henslee-Downey JP et al. Candia glabrata and Candida krusei fungemia after high-risk allogeneic marrow transplantation: no adverse effect of low-dose fluconazole prophylaxis on incidence and outcome. Bone Marrow Transplant 2001; 28: Gooley TA, Leisenring W, Crowley J et al. Estimation of failure probabilities in the presence of competing risks: new representations of old estimators. Stat Med 1999; 18: Klein JP, Rizzo JD, Zhang MJ et al. Statistical methods for the analysis and presentation of the results of bone marrow transplants. Part I: Unadjusted analysis. Bone Marrow Transplant 2001; 28: Mehta J, Powles R, Singhal S et al. Autologous bone marrow transplantation for acute myeloid leukemia in first remission: identification of modifiable prognostic factors. Bone Marrow Transplant 1995; 16: Godder KT, Hazlett LJ, Abhyankar SH et al. Partially mismatched related-donor bone marrow transplantation for pediatric patients with acute leukemia: younger donors and absence of peripheral blasts improve outcome. J Clin Oncol 2000; 18: Van Rood JJ, Loberiza Jr FR, Zhang MJ et al. Effect of tolerance to noninherited maternal antigens on the occurrence of graft-versus-host disease after bone marrow transplantation from a parent or an HLA-haploidentical sibling. Blood 2002; 99: Petersdorf EW, Gooley TA, Anasetti C et al. Optimizing outcome after unrelated marrow transplantation by comprehensive matching of HLA class I and II alleles in the donor and recipient. Blood 1998; 92: Hansen JA, Gooley TA, Martin PJ et al. Bone marrow transplants from unrelated donors for patients with chronic myeloid leukemia. N Engl J Med 1998; 338: Andre-Schmutz I, Le Deist F, Hacein-Bey-Abina S et al. Immune reconstitution without graft-versus-host disease after haemopoietic stem-cell transplantation: a phase 1/2 study. Lancet 2002; 360: Drobyski WR, Klein J, Flomenberg N et al. Superior survival associated with transplantation of matched unrelated versus one-antigen-mismatched unrelated or highly human leukocyte antigen-disparate haploidentical family donor marrow grafts for the treatment of hematologic malignancies: establishing a treatment algorithm for recipients of alternative donor grafts. Blood 2002; 99: Handgretinger R, Lang P, Schumm M et al. Immunological aspects of haploidentical stem cell transplantation in children. Ann NY Acad Sci 2001; 938: Singhal S, Henslee-Downey PJ, Powles R et al. Haploidentical versus autologous hematopoietic stem cell transplantation in patients with acute leukemia beyond first remission. Bone Marrow Transplant 2003; 31: Lazarus HM, Perez WS, Weisdorf D et al. Autologous versus unrelated donor (URD) transplantation for acute myeloid leukemia (AML) in first or second remission (CR1 or CR2). Blood 2000; 96 Abstract # Weisdorf D, Bishop M, Dharan B et al. Autologous versus allogeneic unrelated donor transplantation for acute lymphoblastic leukemia: comparative toxicity and outcomes. Biol Blood Marrow Transplant 2002; 8: Anasetti C, Anderson G, Appelbaum FR et al. Phase III study of rhgm-csf in allogeneic marrow transplantation from unrelated donors. Blood 1993; 82 (Suppl 1): 454a. 36 Schriber JR, Chao NJ, Long GD et al. Granulocyte colonystimulating factor after allogeneic bone marrow transplantation. Blood 1994; 84: Volpi I, Perruccio K, Tosti A et al. Postgrafting administration of granulocyte colony-stimulating factor impairs functional immune recovery in recipients of human leukocyte antigen haplotype-mismatched hematopoietic transplants. Blood 2001; 97: