1. INTRODUCTION

Allogenic hematopoietic stem cell transplant (HSCT) is a curative therapy for hematologic malignancies. However, posttransplantation relapse is a major cause of treatment failure. Donor lymphocyte infusion (DLI) can effectively prevent or treat post-HSCT relapse by reinforcing graft-vs-leukemia (GVL) effects. Although DLI-based therapeutic strategies were originally used for chronic myeloid leukemia,1 a growing body of evidence supports their application for acute leukemia and lymphoma.2,3

DLI can be used at several time points and for various purposes. Prophylactic DLI (proDLI) is typically started several months after HSCT in patients with a high risk of relapse, provided that such patients are not taking immunosuppressants and do not have graft-vs-host disease (GVHD).4 Preemptive DLI (preDLI) is a salvage therapy for minimal residual disease (MRD) positivity or decreased donor chimerism before morphological relapse.2 In the case of overt hematologic relapse, therapeutic DLI (tDLI) is started and is typically followed by cytoreductive chemotherapy.5,6

Steady-state (SS) DLI has frequently been performed using lymphocytes collected through apheresis without premedication.7 During the early era, a fixed dose of SS-DLI was advocated for both adult and pediatric patients.1 In 2016, the Acute Leukemia Working Party of the European Society for Blood and Marrow Transplantation conducted a pilot study and recommended initial prophylactic or preemptive DLI doses of 1 × 104, 1 × 105, and 1 × 106 cells/kg following haploidentical HSCT, matched unrelated donor allogeneic HSCT, and matched sibling donor allogeneic HSCT, respectively.8 Repeat DLI with dose escalation at intervals of 4 to 6 weeks was also recommended. The consolidation of a high dose of DLI (1 × 107 cells/kg) after chemotherapy was recommended for adult patients with acute myeloid leukemia (AML).5

Another type of DLI, granulocyte colony-stimulating factor (G-CSF)-primed DLI, is performed using cells derived through apheresis after G-CSF mobilization.9 Because of the cell collection procedure, the amounts of CD34+ stem cells, mononuclear cells, and lymphocytes infused are higher for G-CSF-primed DLI than for SS-DLI.10 G-CSF stimulates various immune cell populations (eg, plasmacytoid T cell–derived CD11c− dendritic cells, HLA-DR+ TCR gamma-delta cells, and CD56dim subpopulation of NK cells).11,12 Whether the polarization of immune cell subpopulations and nonlymphocytic components increases the GVL effect, reduces the GVHD risk, and provides survival advantages in a clinical setting is a highly debated topic.

In a real world setting, a wide range of SS-DLI doses (1 × 105 to 5 × 107 cells/kg) is administered for preemptive purposes in pediatric patients.13–19 Some scholars have advocated a modified protocol for G-CSF-primed DLI, which involves the administration of total mononuclear cells (TMNCs) at a dose of 1 × 108 cells/kg and the short-term use of the immunosuppressants cyclosporine and methotrexate.20 However, in the context of pediatric patients with hematologic malignancies, the literature on the optimal implementation of DLI with respect to cell dose, DLI type, and clinical outcomes is limited.

This study was conducted to analyze the outcomes of DLI after allogeneic HSCT in pediatric patients with hematologic malignancies. We further identified the factors influencing the survival outcomes and compared the effects of various types and doses of DLI on the development of GVHD and duration of survival. Our findings may facilitate the development of effective DLI strategies for pediatric patients.

2. METHODS 2.1. Patients

We retrospectively included every pediatric patient with hematologic malignancies (diagnosed before the age of 18 years) who had received ≥1 dose of DLI at Taipei Veteran General Hospital (between June 1998 and December 2022) after allogeneic HSCT. In our clinical practice, prophylactic DLI was considered in patients with high risk of relapse based on unfavorable genetics or advanced disease at HSCT. However, the final decision was based on attending physician’s clinical judgement and shared decision-making with patient and families. Preemptive DLI was indicated in patients with decreased short tandem repeat (STR) chimerism or molecular relapse. Therapeutic DLI was indicated for overt relapse.

Through a retrospective review of medical charts, data pertaining to their characteristics, comorbidities, stem cell sources, DLI intentions, infusion characteristics (ie, type and cell dose), and clinical outcomes were collected. The study was approved by the institutional review board (IRB) of the study hospital (TPVGH-IRB 2023-06-008AC). A waiver of documentation for signed informed consent was also granted by the IRB.

2.2. Donor chimerism and MRD monitoring

The donor chimerism of the patients’ bone marrow (BM) was monitored at intervals of 1 to 2 weeks during the first 2 months and subsequently at intervals of every 1 to 2 months in the first year following the transplant. Mixed chimerism was evaluated through STR analyses and defined as the presence of host DNA >1%. BM MRD was also monitored every 1 to 2 months through real-time quantitative polymerase chain reaction or leukemia-associated immunophenotyping in the first year following transplant. Molecular relapse was defined as the presence of ≥0.01% but <5% leukemia-related cells in the BM. Positron emission tomography-computed tomography was performed to evaluate the disease status of the patients with lymphoma. After each DLI infusion, disease status and STR chimerism in BM would be followed in 1 to 2 months. Short-term response was defined as positive if full donor chimerism and/or MRD negativity has been achieved in the first follow-up examination after DLI.

2.3. DLI

When DLI was initially introduced, a fixed dose of SS-DLI was administered as a single infusion or multiple infusions. With the accumulation of clinical DLI experience and the changes in the relevant guidelines, 2009 onward, the policy for serial DLI shifted toward dose escalation. Considering the potential advantages of G-CSF-primed DLI indicated in several immunologic studies,9,11,12 since 2017, we have been routinely cryopreserving an excessive amount of peripheral blood stem cell (PBSC) product obtained on day 0 of HSCT for future use in G-CSF-primed DLI. In studies focusing on SS-DLI, starting DLI doses ranging from 1 × 105 to 1 × 106 CD3+ cells/kg were used for prophylactic and preemptive purposes after matched sibling or unrelated donor HSCT and a higher initial dose (1 × 107 CD3+ cells/kg) was used for frank relapses.5,8 In a study on G-CSF-primed DLI, a mononuclear cell dose of 1 × 108 cells/kg was administered as a prophylactic, preemptive, or therapeutic modality,20 and the administration of DLI was discontinued upon achieving molecular remission, full donor chimerism, or GVHD (any grade) development.

2.4. Statistical analysis

The patients’ 5-year overall survival (OS) and event-free survival (EFS) were estimated through Kaplan-Meier survival analyses and univariate Cox regression. OS was defined as the interval between the date of the first DLI to that of death or the most recent follow-up visit. Disease relapse and the occurrence of death were regarded as events. GVHD was graded in accordance with the guidelines of the National Institutes of Health.21 Cumulative incidence was estimated for relapse and nonrelapse mortality (NRM) to account for competing risks; for this, we used a method developed by Fine and Gray in 1999. A p value of <0.05 was regarded as significant.

Because of a 25% data gap for CD3+ cell count pertaining to the early era of DLI treatment, the Pearson correlation test was performed to evaluate the linear relationship between total lymphocyte count (TLC) and CD3+ cell count. If a strong correlation (r > 0.8) was detected, the missing CD3+ cell count data were estimated using the regression model for TLC. All statistical analyses were performed using SPSS (version 25.0; SPSS Inc., Chicago, IL).

3. RESULTS 3.1. Patient demographics

Table 1 presents patient characteristics, including the details of their DLI and clinical outcomes. This study included 23 patients who had received their first DLI dose at median age of 9. The hematologic malignancy included acute lymphoblastic leukemia (ALL), AML, chronic leukemia, and lymphoma. All patients in this study received allogeneic transplant due to relapse/refractory disease status (n = 18), high-risk features (n = 3, ALL in VHR group in 1, AML with MRD positivity in 1, T-ALL in 1) or underlying diseases (n = 2, systemic EBV-positive T-cell lymphoma of childhood, and juvenile myelomonocytic leukemia [JMML]). The median follow-up period after first DLI was 10.5 years for the 10 living patients.

Table 1 - Characteristics in pediatric patients receiving DLI for prophylaxis or treatment of relapse after allogeneic HSCT Characteristics n (%) Age at DLI (y)  Median (range) 9 (1-21) Gender  Male 12 (52.2%)  Female 11 (47.8%) Diagnosis  ALL 11 (47.8%)  AML 6 (26.1%)  CML 2 (8.7%)  JMML 1 (4.3%)  Lymphoma 3 (13.0%) Stem cell source  Sibling-HSCT 12 (52.2%)   BM 9 (39.1%)   PBSC 3 (13.0%)  Unrelated-HSCT 11 (47.8%)   BM 3 (13.0%)   PBSC 8 (34.8%) HLA match in allelic level  6/6 match 20 (87%)  5/6 or 4/6 match 3 (13%) Disease status before HSCT  Remission 13 (56.5%)  Nonremission 9 (39.1%)  Missing data 1 (4.3%) Treatment period  1998-2010 14 (60.9%)  2011-2022 9 (39.1%) Disease status before DLI  Remission 8 (34.8%)  Mixed chimerism 5 (21.7%)  Molecular relapse 2 (8.7%)  Overt relapse 8 (34.8%) DLI indication  Prophylactic 7 (30.4%)  Preemptive 7 (30.4%)  Therapeutic 9 (39.2%) Interval between HSCT and DLI (mo)  Median (range) 3 (0-51) DLI type  Steady-state DLI 19 (82.6%)  G-CSF-primed DLI 4 (17.4%) DLI times  Median (range) 2 (1-6) Cell dose escalation in serial DLI  One dose only 6 (27.3%)  Fixed dose 11 (47.8%)  Escalating dose 5 (21.7%) Cell dose (×107 cells/kg)  Median initial TLC (range) 7.5 (0.6-23.8)  Median initial CD3 (range) 4.6 (0.5-12.6)  Median cumulative TLC (range) 18.7 (1.2-104.1)  Median cumulative CD3 (range) 13.2 (0.7-54.8) Follow-up duration in living patients (y)  Median (range) 10.5 (0.1-22.2) Clinical outcomes  Alive with leukemia free 10 (43.5%)  Died of leukemia 10 (43.5%)  Died of NRM 3 (13.0%)

ALL = acute lymphoblastic leukemia; AML = acute myeloid leukemia; BM = bone marrow; CML = chronic myeloid leukemia; DLI = donor lymphocyte infusion; G-CSF = granulocyte colony stimulating factor; HLA = human leukocyte antigens; HSCT = hematopoietic stem cell transplantation; JMML = juvenile myelomonocytic leukemia; NRM = nonrelapse mortality; PBSC = peripheral blood stem cell; TLC = total lymphocyte count.

3.2. DLI and its outcomes

In total, 70 DLIs were performed for the 23 patients. Before DLI, 8 patients (34.8%) achieved complete remission, 5 (21.7%) had mixed chimerism, 2 (8.7%) had molecular relapse, and 8 (34.8%) experienced overt relapse. Furthermore, 16 patients (69.6%) received multiple doses; a dose-escalating regimen was followed for 5 patients (21.7%).

Fifteen patients (65.2%) exhibited short-term clinical responses to DLI; these responses included achieving persistent remission within 6 months after proDLI (n = 6), full donor chimerism (n = 4), and MRD negativity (n = 1) after preDLI or remission in overt cases (n = 4) in 1 to 2 months after tDLI. The patients’ 5-year EFS and OS were 43.5% ± 10.9% and 38.3% ± 10.3% (Fig. 1A). Ten (43.5%) patients died of their underlying hematologic malignancies, and another three patients (13.0%) died of severe GVHD and infection (NRM). The other 10 patients (43.5%) maintained their disease-free status. The cumulative incidence of relapse (CIR) and NRM were 47.8% ± 10.5% and 13.3% ± 7.1% (Fig. 1B). The 5-year EFS of the patients who had received prophylactic, preemptive, and therapeutic DLIs was 42.9% ± 18.7%, 57.1% ± 18.7%, and 22.2% ± 13.9%, respectively (p = 0.195; Fig. 1E). Due to different disease burden for DLI in preemptive or therapeutic use, we separately compared our treatment results with available literatures regarding DLI use in two different indications in pediatric patients in Table 3. Outcome of prophylactic DLI was not listed due to lack of any indicator to reflect short-term effect in disease-free patients prior to DLI. In addition, there was scarce literature of prophylactic DLI in pediatric patients. Supplementary Table 1 (https://links.lww.com/JCMA/A209) demonstrates the treatment details and outcomes in patients receiving preemptive or therapeutic DLI.

Fig. 1:

(A) Overall, event-free survival (EFS), and (B) cumulative incidence of relapse (CIR) and nonrelapse mortality (NRM) in pediatric patients of hematologic malignancies receiving donor lymphocyte infusion (DLI) after allogeneic hematopoietic stem cell transplantation (HSCT). EFSs were compared according to (C) pre-HSCT disease status, (D) preemptive DLI (preDLI) disease status, (E) DLI intention, and (F) DLI type. OS = overall survival; proDLI = prophylactic DLI.

3.3. Risk factors for survival

The potential risk factors for reduced survival were identified through Cox regression and summarized in Table 2. The patients’ disease status before HSCT (complete remission vs active disease, p = 0.024) and DLI (complete remission + MRD(+) vs overt relapse, p = 0.039) significantly influenced their EFS (Fig. 1C, D). However, both factors exhibited nonsignificant correlations with OS (p > 0.05 for both factors).

Table 2 - Univariate analyses of risk factors for survivals in pediatric patients of hematologic malignancies receiving DLI after allogeneic HSCT Risk factors Event-free survival Overall survival HR (95% CI) p HR (95% CI) p Age at DLI (<10 vs ≥ 10 y old)a 0.650 (0.227-1.865) 0.423 0.812 (0.272-2.425) 0.709 Gender (male vs female)a 1.419 (0.490-4.105) 0.518 1.489 (0.488-4.542) 0.484 Diagnosis (myeloid vs lymphoid)a 2.202 (0.759-6.385) 0.146 1.969 (0.661-5.872) 0.224 Donor sources (sibling vs unrelated)a 0.505 (0.173-1.474) 0.211 0.528 (0.174-1.601) 0.259 Stem cell sources (BM vs PBSC)a 1.175 (0.407-3.392) 0.765 1.252 (0.404-3.880) 0.697 HLA match in allelic level (6/6 vs < 6/6)a 0.710 (0.158-3.196) 0.655 1.483 (0.323-6.800) 0.612 Disease status before HSCT (CR vs active disease)a 0.278 (0.091-0.844) 0.024b 0.400 (0.133-1.198) 0.102 Treatment period (1998-2010 vs 2011-2022)a 1.123 (0.376-3.358) 0.836 0.880 (0.278-2.278) 0.827 Disease status before DLI (CR vs overt relapse)a 0.306 (0.099-0.950) 0.040b 0.346 (0.110-1.083) 0.068 Disease status before DLI (CR + MRD(+) vs overt relapse)a 0.318 (0.107-0.944) 0.039b 0.354 (0.118-1.065) 0.065 DLI intention (prophylactic vs therapeutic)a 0.432 (0.122-1.534) 0.194 0.452 (0.125-1.627) 0.224 DLI intention (prophylactic + preemptive vs therapeutic)a 0.411 (0.138-1.218) 0.109 0.408 (0.135-1.238) 0.113 Time from HSCT to DLI (< 6 vs ≥ 6 months)a 1.577 (0.4395.657) 0.485 1.614 (0.444-5.873) 0.468 DLI type (steady-state vs G-CSF-primed)a 1.345 (0.301-6.017) 0.698 1.509 (0.329-6.929) 0.597 DLI times (< 2 vs ≥ 2)a 0.786 (0.219-2.824) 0.712 0.868 (0.234-3.229) 0.833 Cell dose escalation (fixed doses vs single dose)a 0.933 (0.233-3.739) 0.922 0.820 (0.196-3.435) 0.785 Cell dose escalation (escalating doses vs single dose)a 2.087 (0.448-9.720) 0.349 2.566 (0.456-14.42) 0.285 Initial DLI cell dose (CD3+ < vs ≥ 5 × 107 cells/kg)a 0.743 (0.243-2.274) 0.603 0.636 (0.189-2.141) 0.465 Final DLI cell dose (CD3+ < vs ≥ 5 × 107 cells/kg)a 0.585 (0.195-0.754) 0.339 0.377 (0.177-0.210) 0.099 Cumulative DLI cell dose (CD3+ < vs ≥ 13 × 107 cells/kg)a 0.621 (0.206-1.874) 0.398 0.528 (0.160-1.742) 0.294 Acute GVHD (Gr. <3 vs ≥ 3)a 1.219 (0.408-3.639) 0.723 0.951 (0.310-2.914) 0.930 Chronic GVHD (no or mild vs moderate or severe)a 1.332 (0.295-6.010) 0.709 1.149 (0.251-5.257) 0.858

aReference group.

BM = bone marrow; CI = confidence interval; CR = complete remission; DLI = donor lymphocyte infusion; G-CSF = granulocyte colony stimulating factor; GVHD = graft vs host disease; HLA = human leukocyte antigens; HR = hazard ratio; HSCT = hematopoietic stem cell transplantation; MRD = minimal residual disease; PBSC = peripheral blood stem cell

3.4. GVHD

The incidence of acute GVHD (aGVHD) was 61% (grade I, 4%; grade II, 17%; grade III, 26%; grade IV, 13%) and that of cGVHD was 47% (mild, 17%; moderate to severe, 30%). We found no correlation between the development of aGVHD and the duration of survival (Fig. 2A, B). The patients with mild cGVHD had more favorable EFS and OS than did individuals without cGVHD or patients with moderate to severe cGVHD; however, this difference was nonsignificant (Fig. 2C, D).

Fig. 2:

Overall survival (A) (C) and event-free survival (B) (D) in pediatric patients of hematologic malignancies receiving donor lymphocyte infusion (DLI) after allogeneic hematopoietic stem cell transplantation according to severity of acute or chronic graft-vs-host-disease (cGVHD). Distribution of acute graft-vs-host-disease (aGVHD) (E) (G) and cGVHD (F) (H) grading according to DLI type and DLI dosing escalation strategies. NRM = nonrelapse mortality.

3.5. DLI strategies (type, fixed/escalating dose, and cell dose)

DLI characteristics (ie, DLI type [Fig. 1F], DLI times, fixed or escalating dose, interval from HSCT or between DLIs) did not significantly influence the patients’ EFS or OS (p > 0.05 for all; Table 2). No significant difference was observed between the types of DLI (SS-DLI vs G-CSF-primed DLI) in the severity of aGVHD, cGVHD (Fig. 2E, F), CIR or NRM (Supplementary Fig 1, https://links.lww.com/JCMA/A207). Only the patients who had received dose-escalated DLI exhibited a lower degree of cGVHD severity than did the other patients with borderline significance (p = 0.051; Fig. 2H).

The missing data of CD3+ cell count in the early era were estimated using the regression model for TLC (R2 = 0.82; Supplementary Fig 2A, https://links.lww.com/JCMA/A208). The initial, final, and cumulative doses of DLI did not affect the patients’ EFS (p > 0.05 for all). However, the patients who had received a final CD3+ dose of ≥5 × 107 cells/kg had borderline significantly poorer 5-year OS than did those who had received a CD3+ dose of <5 × 107 cells/kg (30% ± 14.5% vs 58.3% ± 16.1%; p = 0.089; Fig. 3A). The differences between the final DLI doses (<5 × 107 vs ≥5 × 107 cells/kg) were nonsignificant for relapse and NRM (p > 0.05 for both; Fig. 3C, D). The distributions of aGVHD and cGVHD grades were significantly different between the final DLI doses (<5 × 107 vs ≥5 × 107 cells/kg; aGVHD, p = 0.039; cGVHD, p = 0.002; Fig. 3E, F). Furthermore, patients who had received higher CD3+ doses for their final infusion developed higher grades of aGVHD or cGVHD (p = 0.057 or 0.017; Fig. 3G, H).

Fig. 3:

Impact of CD3+ cell dose (< vs ≥ 5 × 107 cells/kg) in final donor lymphocyte infusion (DLI) on (A) overall survival, (B) event-free survival, (C) cumulative incidence of relapse, (D) cumulative incidence of nonrelapse mortality (NRM), and distribution of acute graft-vs-host-disease (GVHD) (E) and chronic GVHD (F) grading. Final CD3+ cell dose among different grading of acute GVHD (G) or chronic GVHD (H) in pediatric patients of hematologic malignancies receiving DLI after allogeneic hematopoietic stem cell transplantation.

4. DISCUSSION

Among the patients who had received DLI at our hospital, the overall short-term response rate was 65.2%, 5-year EFS rate was 43.5%, and 5-year OS rate was 38.3%. In compared to the results with other preDLI cohorts, similar short-term response rate was achieved in our study, accompanied with higher incidences of severe GVHD and NRM, which probably resulted from the extremely high initial cell dose (8 × 107 cells/kg) used in the early era in our cohort. Therefore, initiation with lower cell dose (1 × 105 ~ 1 × 106 cells/kg) with gradual escalation would be advocated to achieve the goal in a safer way because the disease burden is minimal in the preDLI group. For the patients with overt relapse in the tDLI cohorts, we used higher initial DLI dose (1 × 107 cells/kg) regardless of matched sibling or unrelated donor. Despite a fair short-term response (44% in our cohort vs 21% to 39%) was achieved, 7 patients (77.8%) still died of leukemia. Combing other more effective systemic therapies with therapeutic DLI should be considered in the current era when targeted therapy with anti-CD19 monoclonal antibody or anti-CD22 drug conjugate is available with a good safety profile.

The patients who had achieved complete remission or MRD(+) exhibited more favorable EFS outcomes than did those had relapse, indicating that those with a lower disease burden respond more favorably to DLI. With the use of STR analyses combined with widely available and precise MRD methodology for pediatric ALL in recent years, early DLI initiating might benefit further in obtaining a better clinical response.24 However, the positive effects of disease status-related factors on EFS did not extend to OS, as evidenced by the finding that several patients who had received successful DLIs developed severe GVHD-related complication and died because of prolonged immunosuppression. Nevertheless, part of the patients who had experienced DLI failure still underwent further salvage therapy, including a second transplant, to maintain their survival.

In our study, the patients with mild cGVHD had a more favorable OS than did individuals without cGVHD or patients with moderate to severe cGVHD; however, this difference was nonsignificant. GVHD has been reported to exert a time-dependent effect on patient survival following DLI. Grade II to IV aGVHD was associated with a high rate of mortality, whereas cGVHD was associated with a favorable outcome in patients who survived for >100 days after DLI.6 However, Schmid et al2 reported that GVHD was associated not with a reduced relapse rate but with a high NRM rate.

A study reported that an initial dose of ≥1 × 107 CD3+ cells/kg was associated with an increased risk of GVHD; further titration of the CD3+ cell dose up to ≥10 × 107 cells/kg failed to reduce the risk of relapse and did not improve OS.25 In the present study, a final DLI dose of ≥ 5 × 107 CD3+ cells/kg was borderline-significantly associated with poor OS and significantly associated with a high risk of moderate to severe cGVHD. Besides, in the present study, the patients following a cell dose escalation regimen exhibited a reduced risk of severe cGVHD, though this reduction in cGVHD risk did not translate into favorable OS or EFS or reduced NRM. This finding may be attributed to the small sample size the present study. Although directly comparison between different cell dose escalating schedule was not feasible in our study due to limited case numbers, the different cell dose escalating rule, 2- to 5-fold13 or 5- to 10-fold,15 did not make significant impacts on clinical outcome by literature review shown in Table 3.

Table 3 - Comparisons of treatment outcomes after preemptive DLI and therapeutic DLI in pediatric patients of hematologic malignancies Disease and case number Fixed or escalating CD3+ cell dose Survival outcome GVHD and NRM Preemptive DLI  This study (Taipei Veteran General Hospital, Taiwan) N = 7
ALL (n = 2)
AML (n = 2)
CML (n = 1)
Lymphoma (n = 2) Single dose (n = 4)
Fixed dose (n = 2)
Missing data (n = 1) Initial:
SS-DLI (n = 5)
8.6 × 107/kg
G-CSF-primed (n = 2)
TMNC 1 × 108/kg Short-term response: 71%
5 y EFS: 57%
5 y OS: 54% aGVHD (III-IV): 43%
cGVHD (moderate to severe): 14%
NRM: 14%  Liou et al13 (Benioff Children’s Hospital, USA) BMT (2017) N = 35
ALL (n = 17)
AML (n = 18) Escalating (2- to 5-fold) Initial:
MSD: 1 × 106/kg
MUD: 1 × 105/kg Short-term response: NA
5 y EFS: 71%
5 y OS: NA aGVHD (III-IV): 6%
cGVHD (moderate to severe): 20%
NRM: 6%  Goździk et al14 (Poland pediatric group for HSCT centers) BMT (2015) n = 3
Diverse hematologic malignancy
(NA in this subgroup) Escalating 2.8 × 107/kg Short-term response: 100%
5 y EFS: NA
5 y OS: NA NA in this subgroup
But for preDLI + tDLI
aGVHD (III-IV): 14%
cGHVD (moderate to severe): 8%
NRM: 5%  Horn et al15 (Benioff Children’s Hospital and All Children’s Hospital, USA) BBMT (2015) N = 14
Lymphoid (n = 5)
Myeloid (n = 9) Single (n = 3)
Escalating (5- to 10-fold) (n = 11) Initial:
MSD: 1 × 106/kg
MUD: 1 × 105/kg Short-term response: 79%
5 y EFS: 65%
5 y OS: 65% aGVHD (III-IV): 0%
cGVHD (mo