Peripheral blood hematopoietic stem cells for transplantation of hematological diseases from related, haploidentical donors after reduced-intensity conditioning

Abstract

In a multi-center collaboration, we carried out T-replete, peripheral blood stem cell (PBSC) transplants from related, HLA-haploidentical donors with reduced-intensity conditioning (RIC) and post-transplant cyclophosphamide (Cy) as GvHD prophylaxis in 55 patients with high-risk hematologic disorders. Patients received 2 doses of Cy 50mg/kg IV on days 3 and 4 post-infusion of PBSC (mean 6.4×10⁶/kg CD34+ cells, mean 2.0 ×10⁸/kg CD3+ cells). The median times to neutrophil (500/mcL) and platelet (>20,000/mcL) recovery were 17 and 21 days respectively. All but two of the patients achieved full engraftment. The one-year cumulative incidences of grade II and grade III acute GVHD were 53% and 8% respectively. There were no cases of grade IV GvHD. The two-year cumulative incidence of chronic GHVD was 18%. With a median follow up of 509 days, overall survival (OS) and event free survival (EFS) at two years were 48% and 51%, respectively. The two-year cumulative incidences of non-relapse mortality (NRM) and relapse were 23% and 28% respectively. Our results suggest that PBSC can be substituted safely and effectively for bone marrow (BM) as the graft source for haploidentical transplantation following reduced intensity conditioning.

INTRODUCTION

Allogeneic hemopoietic stem cell transplantation from HLA-matched donors is curative in a proportion of patients with hematologic malignancies as well as inherited diseases such as hemoglobinopathies and bone marrow failure syndromes. A suitable HLA-identical sibling donor will be available for about 30–35% of patients. For patients without a matched sibling donor, the likelihood of identifying a volunteer unrelated donor that is suitably matched at HLA-A, -B, -C, and -DRB1 is population-specific ranging from about 79% for Caucasian patients of European descent to 30–50% for patients of other ethnic backgrounds ¹. Even if a matched, unrelated donor is identified, the likelihood of proceeding to transplant is less than 50% largely due to progression of disease during the search process which renders the patient ineligible for transplant². For this substantial number of patients, alternative donors such as haploidentical family donors, unrelated umbilical cord blood or mismatched unrelated donors can bridge this gap, enabling transplantation. Over the past decade, single-center and multi-center, cooperative group trials have shown that the administration of high doses of cyclophosphamide 60–72 hours after the infusion of bone marrow cells from related, haploidentical donors (haplo-BM) enables engraftment with low rates of rejection, acute and chronic GvHD and non-NRM^3–5. In these studies, disease relapse remained the major cause of treatment failure. Currently, PBSC are the preferred source of allografts because of ease of collection, higher yields of CD34-positive progenitor cells, faster engraftment and, in the case of matched sibling donor transplants, improved short and long term survival^6–8. However, PBSC products contain approximately 5–10 times higher numbers of CD3-positive T-cells than BM harvests ⁹ which correlate with higher rates of acute and chronic GHVD compared to BM ^{7, 10–13}. In this report, we describe outcomes after transplantation of patients with high-risk hematologic disorders following RIC in which haplo-PBSC was substituted for haplo-BM. Prophylaxis of graft-versus-host disease included post-transplant cyclophosphamide (PTCy) to eliminate alloreactive T-cell populations, a calcineurin-inhibitor and mycophenolate mofetil. Although the rate of grade II acute GVHD was higher, engraftment kinetics, grade III acute GVHD, chronic GVHD and survival outcomes were similar to those reported for haplo-BM.

PATIENTS AND METHODS

Eligibility Criteria

Consecutive, eligible patients from 4 centers, Guy’s and St Thomas’ Hospital, London, King’s College Hospital, London, Westmead Hospital, Sydney and the Fred Hutchinson Cancer Research Center, Seattle who underwent a stem cell transplant with PBSC from related, HLA-haploidentical donors between March 2009 and February 2013 were included (Table 1). All patients had at least 160 d of follow-up. Results were analysed as of August 2013. All patients signed consent forms approved by their local institutional review boards. Sharing of de-identified transplant data was approved by the institutional review boards of each of the participating centers. Patients were ≤ 70 years of age with a high risk hematologic disorder but lacked a suitably matched related or unrelated donor defined as a donor with an 9–10/10 locus HLA match at HLA-A, HLA-B, HLA-C, HLA-DRB1 and HLA-DQB1. An unrelated donor search was not required for a patient to be eligible for this protocol, or a search could be abandoned if the clinical situation dictated an urgent transplant. Clinical urgency was defined as 6–8 weeks from referral to transplant or a low likelihood of finding a matched, unrelated donor. In this study, the only times that an unrelated donor search was not performed was in the cases of 4 patients who had rejected allografts from 10/10 matched donors and there was urgency in performing a salvage transplant. Patients with acute leukemia were required to be in morphologic complete remission (CR). Patients with primary or secondary graft failure in a prior allogeneic transplant were also eligible. Patients were required to have adequate organ function defined as left ventricular ejection fraction ≥ 35%, Forced expiratory volume in the first second, functional vital capacity or carbon monoxide corrected diffusion lung capacity >50% of predicted; total bilirubin ≤ 2.5mg/dL and aspartate and alanine aminotransferases and alkaline phosphatase <5 times the upper limit of normal; serum creatinine within the normal range for age or creatinine clearance or calculated glomerular filtration rate >40ml/min/1.73 m². A Karnofsky performance score of ≥60 was required. Patients who had undergone a prior autologous transplant were eligible provided 3 months had elapsed since the procedure.

Table 1.

Patient and Disease characteristics

Characteristics	HLA-Haploidentical Related
N	55
Seattle FHCRC	20
London Guy’s and St. Thomas’	7
London Kings College Hospital	18
Westmead Hospital	10
Age
Recipient Median in years (Range)	49 (14–69)
Donor Median in years (Range)	40 (15–73)
Sex, n (%)
Male	35 (64)
Female	20 (36)
Recipient Ethnicity, n (%)
Caucasian	37 (67)
Afro-Caribbean	9 (16)
Asian	9 (16)
Donor Relationship, n (%)
Mother	8 (15)
Father	5 ( 9)
Brother	8 (15)
Sister	12 (22)
Son	14 (25)
Daughter	8 (15)
CMV serostatus, n (%)
Recipient −/Donor −	14 (25)
Recipient +/Donor −	8 (15)
Recipient −/Donor +	7 (13)
Recipient +/Donor +	26 (47)
ABO Compatibility, n (%)
Compatible	27 (49)
Minor Mismatch	12 (22)
Major Mismatch	15 (27)
Bi-directional Mismatch	1 ( 2)
Time diagnosis-Allogeneic HCT
Median Months (range)	23 (2–215)
Prior chemotherapy/radiotherapy
Median Lines n (range)	3 (1–15)
Prior Transplantation, n (%)
Autologous	12 (22)
Median months from Auto	23 (6–106)
Allogeneic	7 (13)
Median months from Allo	6 (2–39)
Disease, n (%)
B-NHL	7 (13)
T-NHL	5 ( 9)
AML	16 (29)
HL	9 (16)
ALL	2 ( 4)
MDS	5 ( 9)
SAA	4 ( 7)
CLL	4 ( 7)
CML	3 ( 5)
Myeloid	24 (44)
Lymphoid	27 (49)
Aplastic Anemia	4 ( 7)
Disease status at HCT, n (%)
CR1	20 (36)
CR>1	11 (20)
PR	9 (16)
SD	9 (16)
RD	2 ( 4)
PD	4 ( 7)
ASBMT/CIBMTR disease risk classification, n(%)	51 (excludes AA)
Low Risk	12 (24)
Intermediate Risk	25 (49)
High Risk	14 (25)
Cell dose/kg, mean (SD)
CD34+cells (x106)	6.4 (1.6)
CD3+ cells (x108) (n=19)	2.0 (1.4)
HLA Mismatches No (%)*	HvG	GvH
0	0	0
1	0	0
2	2 ( 4)	3 ( 7)
3	5 (12)	8 (18)
4	13 (29)	10 (22)
5	25 (56)	24 (53)
Median (range)	5 (2–5)	5 (2–5)

AML, Acute Myeloid Leukaemia; MDS, myelodysplastic syndrome; NHL, Non Hodgkin’s lymphoma; CLL, chronic lymphocytic leukaemia; HL, Hodgkin lymphoma; ATLL, Acute T leukaemia lymphoma; TNHL, T-non-Hodgkin’s lymphoma; SAA, severe aplastic anaemia; CML, chronic myeloid leukaemia; CR1, first complete remission; CR>1, subsequent complete remission; PR, partial remission; SD, stable disease; RD, refractory/progressive disease;

^*based on N=45.

Characteristics at HCT

Details regarding patient and donor characteristics are outlined in Table 1. Of the five MDS patients, prior chemotherapy was not administered in 2 patients with IPSS low risk one of whom had rejected 2 prior allografts from the same HLA-matched, unrelated donor and the other patient had rejected a double allograft from 2 unrelated, umbilical cord blood donors. Two of 3 IPSS high risk patients were treated with 4 cycles of 5-azacytidine (75 mg/m² for 7 d every 28 d). One of these 2 patients had cytogenetic progression and was treated subsequently with induction chemotherapy (daunorubicin 50mg/m2 for 3 days and cytarabine 100mg/m2 bd for 10 days (DA 3+10)) prior to proceeding to haplo PBSC. The third high risk patient had previously received 10 cycles of lenalidomide and on disease progression to RAEB-2 was treated with 2 cycles of induction chemotherapy (first with DA 3+10 and then DA 3+8 [8 days of cytarabine at the same dose as above]).

All four patients with severe aplastic anemia (SAA) had failed prior immunosuppressive therapy and 2 of the 4 patients had subsequently rejected stem cell transplants from HLA-matched, unrelated donors. None of the patients received further chemotherapy within 90 d of haplo-PBSC transplantation. One patient with SAA/PNH was treated with eculizumab prior to and during the transplant. MMF was stopped at day +35; however, tacrolimus was maintained at levels of 10–15 ng/mL for 9–12 mo post-transplant and then tapered over a 3 mo period.

HLA Matching and Donor Selection

Haplo-PBSC donors were required to be first degree relatives of the patient, defined as biologic parents, siblings, children or half siblings. Donor and recipients were typed at HLA-A, HLA-B, HLA-C, HLA-DRB1, and HLA-DQB1 at the allelic level for the 45 patients enrolled at the London and Seattle transplant centers. At the Westmead center, 3 of 10 donor-recipient pairs were typed at HLA-A, HLAB and HLA-DRB1 and 7 of 10 donor-recipient pairs were typed at HLA-A, HLA-B, HLA-C and HLA-DRB1 as per the preference of the local Red Cross laboratory. Donor-recipient pairs were considered HLA haploidenticaL if they were genotypically identical for one allele at each of the typed loci. Donors were required to be ≥ 16 years of age and were screened as per the American Association of Blood banks and FACT/JACIE guidelines. Donors were excluded if the recipient’s serum contained anti-donor HLA antibodies. If more than one possible donor was identified, donor selection hierarchy was 1) donor-recipient matching for cytomegalovirus (CMV) serology and 2) donor-recipient red blood cell compatibility. A previous report showed that increasing donor-recipient HLA -mismatch did not have a detrimental effect after non myeloablative HLA haploidentical bone marrow transplant with post-transplant Cy¹⁴. Therefore, higher degrees of HLA matching were not prioritised in donor selection.

Conditioning Regimen and Immunosuppressive Therapies

Recipients were conditioned with fludarabine 30mg/m²/d IV daily from days -6 to -2 (total dose of 150mg/m²), Cy 14.5mg/kg IV on days -6 and -5 and 2Gy total body irradiation in a single dose on day -1 as previously described ^4,5. The dose of fludarabine was adjusted for creatinine clearance as clinically indicated. For patients with an actual body weight of >125% ideal body weight (IBW), Cy was dosed based on adjusted ideal body weight (AIBW). AIBW was computed as the sum of the ideal body weight and 25% of the difference between the actual and ideal body weight. MESNA and IV hydration were administered for uro-protection. GVHD prophylaxis consisted of Cy 50mg/kg by IV infusion over 1–2 hours on days +3 (between 60–72 hours after PBSC infusion) and +4 after transplantation. Patients received tacrolimus and MMF beginning day +5 after infusion of haplo-PBSC. MMF was given at a dose of 15mg/kg every 8 hours with the maximum daily dose not exceeding 3g. MMF prophylaxis was discontinued on day +35 or continued at the discretion of the treating center if active GVHD was present. Tacrolimus was administered to achieve a target trough level of 5–10 ng/mL with the goal of discontinuing at day +180 after transplantation. Filgrastim was initiated at day +5 at a dose of 5μg/kg/d and continued until the neutrophil count was ≥ 1000/μL for 3 consecutive days.

Collection of Hemopoietic Stem Cells and Supportive care

Filgrastim at a dose of 10 mcg/kg (London and Sydney) or 16 mcg/kg (Seattle) actual body weight was administered subcutaneously once daily or equally divided twice daily from day -5 to day -1 followed by collection of PBSC by apheresis on day -1. The target dose of PBSC for infusion was 5–6 × 10⁶ CD34+ cells/kg. If the target dose was met after the first apheresis procedure, PBSC were stored overnight at 4°C prior to infusion on day 0. If the target dose was not met, a second apheresis was performed the next day. PBSC in excess of the target dose were cryopreserved.

Antimicrobial prophylaxis was administered as per the institutional protocols. All patients received prophylaxis for Pneumocystis jirovici pneumonia, herpes simplex/zoster and Candida albicans. Prophylaxis for mold was as per institutional protocols (secondary prophylaxis in Seattle and Sydney and primary prophylaxis in London). Neutropenic patients received prophylaxis with quinolones. Blood products were irradiated to 25Gy prior to infusion. Transfusions for blood and platelets followed institutional protocols. All patients including CMV- negative patients received leucodepleted blood products (“CMV-safe”). CMV DNA viral load was monitored at least weekly by PCR of serum until day 100. Pre-emptive therapy with either ganciclovir (5mg/kg/IV twice daily) or foscarnet (90mg/kg twice daily) was initiated as per the institutional guidelines. Patients in London were treated as in-patients from conditioning to engraftment of neutrophils whereas conditioning and transplantation in Seattle and Sydney were outpatient procedures. Patients in Seattle and Sydney were admitted as indicated for regimen-related toxicity or infection.

GVHD Grading and Treatment

Acute GVHD was graded according to the consensus criteria ¹⁵. Chronic GVHD was assessed as per the NIH criteria ¹⁶. Each institution treated GVHD according to their local protocols. Typically, progressive or grade III/IV acute GVHD was treated initially with 1–2mg/kg of methylprednisolone parenterally with optimisation of the dose of tacrolimus or the addition of MMF to tacrolimus.

Analyses of Donor Chimerism

Peripheral blood donor chimerism was tested at least on day +28 post-transplant and then as per institutional protocols. Chimerism was studied by PCR analysis of short or variable nucleotide tandem repeats unique to the donor and recipient. Chimerism was performed on at least peripheral blood, CD3⁺ cells and CD33⁺ or CD15⁺ cells. Patients were considered fully donor chimeric if their un-fractionated, CD3⁺ and CD33⁺/CD15⁺ fractions were ≥ 95% donor.

Statistical Methods

The outcomes are reported as of August 2013. The main outcomes of interest were engraftment of neutrophils and platelets, incidence and severity of both acute and chronic GVHD, NRM, OS and EFS. Neutrophil engraftment was defined as the time from infusion of donor stem cells and the first of three consecutive days with an absolute neutrophil count of ≥ 500/mcL. Platelet engraftment was similarly defined as the interval between donor stem cell infusion and the first day of a platelet count >20,000/mcL without a platelet transfusion in the preceding 7 days. Donor engraftment was defined as a donor chimerism ≥ 95%. Graft failure was defined as ≤ 5% donor cells after transplantation not due to progressive disease. NRM was defined as death in the absence of detectable disease relapse or progression. Probabilities of OS and EFS were estimated using the Kaplan Meier method ¹⁷ on an intent-to-treat basis. Probabilities of acute GVHD, chronic GVHD, relapse and NRM were summarised using cumulative incidence estimates ¹⁸. Death without engraftment was considered a competing risk for engraftment; death without relapse was a competing risk for relapse; relapse was a competing risk for NRM; graft failure, relapse or death, without GVHD, were considered competing risks for GVHD.

RESULTS

Patient and Graft Characteristics

By ASBMT/CIBMTR disease risk criteria, 24% of patients were characterized as low-risk, 49% as intermediate-risk and 25% as high-risk (Table 1). Nineteen patients had received prior stem cell transplants: 12 (22%) had relapsed following a previous autograft whereas 2 (4%) had relapsed following previous allografts (matched sibling × 1, n=1; double umbilical cord blood × 1, n=1) and 5 (9%) had failed to engraft following prior allografts (matched sibling × 2, n=1; matched unrelated donor × 1, n=1; matched unrelated donor × 2, n=2; double umbilical cord blood ×1, n=1). Patients with hematologic malignancies were divided approximately equally between myeloid (44%) and lymphoid (49%) disorders. Two-thirds of patients were Caucasian whereas 32% were from an ethnic minority population (Afro-Caribbean, n=9, Asian, n=9).

Donor characteristics are also listed in Table 1. Donors were children of recipients in 22 (40%) cases, siblings in 20 (37%) cases and parents in 13 (24%) cases. In the 45 cases where HLA-typing was performed at 5 loci, there were a median of 5 mismatches (range 2–5) in the host-versus-graft direction and a median of 5 mismatches (range 2–5) in the graft-versus-host direction.

Engraftment

The median time to neutrophil recovery was 17 days (range: 12–29 days). The median time to platelet recovery was 21 days (range: 11–48 days); 6 patients did not reach a platelet nadir <20,000/mcL. Primary graft failure occurred in two patients (4%), a patient with MDS/AML who underwent a subsequent salvage transplant from an HLA-mismatched, unrelated donor but ultimately died of relapsed disease and a patient with CML in CP2 who had autologous neutrophil recovery at day +23 and who is alive on tyrosine kinase inhibitor therapy at day +622. Full donor chimerism was detected in un-fractionated peripheral blood or CD3⁺ and CD33⁺ or CD15⁺ fractions of peripheral blood in the remaining 53 patients (96%) by D28 and was sustained.

Infections

No invasive mold infections or EBV reactivations were observed in any patients on study. CMV reactivation occurred in 27 of 34 patients (79%) who were at high risk for reactivation (donor seropositive/seronegative, recipient seropositive) at a median of 33 days post-transplant. Patients received pre-emptive therapy and no case of CMV disease occurred. There were no cases of primary CMV infection.

Acute and Chronic GVHD

The cumulative incidences of grades II and grade III acute GVHD at one-year were 53% (95% confidence interval [CI]: 40–66%) and 8% (95% CI: 0.4–15%), respectively (Figure 1A). No grade IV acute GvHD was observed. The median time to onset of acute GVHD was 33 days. Two patients developed late acute GvHD of the skin and gut at 113 and 261 days, respectively. Twenty-nine patients developed grade II acute GvHD most frequently involving skin and lower gut in 15 patients. Seven of the 29 patients developed mild stage I GvHD of the upper GI tract confirmed by endoscopic biopsy¹⁹ that resolved completely after a short course of prednisone and beclomethasone²⁰. Six of these patients were diagnosed in the Seattle cohort. Four patients developed grade III acute GvHD; skin and gut were involved in one patient, skin, gut and liver in two patients and gut and liver in one patient. Twenty-three of 33 patients developing acute GvHD were treated with systemic steroids. All patients responded to initial therapy; 5 patients responded to 0.5 mg/kg of steroids, 9 patients responded to 1 mg/kg of steroids and 9 patients responded to 2 mg/kg of steroids including the 4 patients with grade III GvHD.

Figure 1.

Cumulative Incidence of A. Grade II and Grade III. B NIH Chronic GVHD

The cumulative incidence of chronic GHVD was 16% (95% CI 6–26%) at one-year and 18% (95% CI 7–29%) at two-years (Figure 1B). Nine cases of chronic GvHD were observed. Four cases were mild in severity, involving eye, gut or liver. Three cases were moderate with one involving the skin, one involving the lungs and one involving serosal surfaces. Two cases were severe with one involving the lungs and one involving the eye, skin and liver. One patient experienced immune myopathy and neuropathy at day+40 consistent with the diagnosis of Guillain-Barre Syndrome that resolved after treatment with steroids and IV immunoglobulin.

NRM and Relapse

The incidence of NRM was 17% (95% CI: 7–27%) at one-year and 23% (95% CI: 10–35%) at two years post- transplant (Figure 2A). There were 12 deaths from non-relapse causes. One patient died from complications associated with GvHD and one patient died from diffuse alveolar haemorrhage. There were 10 deaths from infections including 4 viral infections (parainfluenza, n=2 and adenovirus, n=2), 4 bacterial infections (P. aerugenosa, n=2, combined E. coli and K. pneumonia, n=1 and R. planticola, n=1), one yeast infection (S. cerevisae) and one likely infection of unclear etiology. As shown in Figure 2A, the cumulative incidence of relapse (excluding the 4 patients with SAA) at two years posttransplant was 28% (95% CI: 14–42%). Relapse occurred in 12 patients (ATLL n=3, AML n=3, HIV-associated NHL n=1, ALCL, n=1, CLL n=1, HL n=2, CML-CP2, n=1).

Figure 2.

A Cumulative incidence of relapse and non-relapse mortality. B Overall and event-free survival at 2 years post-transplant.

Overall and Event-Free Survival

The median follow-up of surviving patients was 509 (range: 160–1203) days. OS and EFS at one-year post-transplant were 78% (95% CI: 64–87%) and 66% (95% CI: 51–77%), respectively (Figure 2B). At two-years post-transplant, OS and EFS were similar at 48% (95% CI 30–64%) and 51% (95% CI 35–65%), respectively. The OS and EFS curves appear to cross at 2 years due to small sample variability (n=11). There were no significant differences in EFS or OS between patients with myeloid malignancies compared to patients with lymphoid malignancies (data not shown).

DISCUSSION

We have shown that substituting PBSC for BM as the graft source in the Hopkins protocol for haploidentical transplantation following RIC is feasible and does not affect outcomes adversely. Our results suggest that BM or PBSC could be used interchangeably as allograft sources for haploidentical transplantation using this regimen.

The fixed dose of CD34⁺ cells in the PBSC allograft (5–6 × 10⁶/kg) was chosen for two reasons. Firstly, it approximated the median CD34⁺ cell dose in BM allografts reported previously for this particular haploidentical protocol ⁴. Secondly, it standardized the T-cell dose since an increased number of T-cells in PBSC products compared to BM products has been associated with increased rates of acute and chronic GvHD in the setting of HLA-matched related ^7,10,11 or unrelated donors^12,13. In this study, the mean number of CD3⁺ T-cells in PBSC allografts (2.0 × 10⁸ CD3+ cells/kg) was about 5-fold higher than that reported previously for BM allografts ⁴. The most striking effect of increased T-cell dose was an almost 2-fold increase in the rate of grade II GvHD (53% with haplo-PBSC compared to 28–32% with haplo-BM^4,5) which, however, responded completely to steroid therapy in all treated cases. The incidences of grade III acute GvHD and chronic GvHD were low and similar to those reported previously for haplo-BM^3,4. Similar low rates of acute and chronic GvHD were also observed in 2 small studies of haplo-PBSC transplantation following myeloablative conditioning. The first involved a 2-step approach to haploidentical transplantation in which patients were conditioned with high-dose TBI followed by donor lymphocyte infusion of 2 × 10⁸ CD3⁺ cells/kg (the same dose as in this study) and then high-dose Cy prior to infusion of CD34⁺-selected cells from a filgrastim-mobilized PBSC donor²¹. In the second study, patients were conditioned with busulfan, fludarabine and Cy followed by a PBSC allograft targeted at 5 × 10⁶ CD34 cells/kg and post-transplant Cy, tacrolimus and MMF as in this study²².

When compared with the haplo-BM protocol of RIC, rates of non-engraftment (4% for haplo-PBSC, 2–16% for haplo-BM^4,5,14) and of neutrophil and platelet recovery (median 17 and 21 days, respectively, for haplo-PBSC; 16 and 24 days, respectively, for haplo-BM^4,5) were similar. Sustained engraftment was achieved in the 5 patients where haplo-PBSC was used as salvage for graft failure after prior allogeneic transplants from matched sibling, matched unrelated or double umbilical cord blood donors. Sustained engraftment also was achieved in 5 patients with MDS and in 4 patients with severe aplastic anemia which has not been reported previously for haplo-BM using this regimen.

With a median follow up for surviving patients of 17 mo, the two-year cumulative incidence of NRM was 23% for haplo-PBSC compared to 16% for haplo-BM⁴; the two-year cumulative incidence of relapse was 28% for haplo-PBSC compared to 58% for haplo-BM⁴; the two-year probability of EFS was 51% for haplo-PBSC compared to 26% for haplo-BM⁴ and the two-year probability of OS was also 48% for haplo-PBSC compared to 36% for haplo-BM⁴. For a comparable number of patients transplanted in 4 separate centers using haplo-PBSC compared to the multi-center study conducted by the BMT CTN using haplo-BM⁵, it appeared that transplant outcomes of patients with high-risk hematologic disorders after transplantation with haplo-PBSC were comparable to those after transplantation with haplo-BM.