One of the consistent features in development of hematopoietic stem cell transplant (HCT) for Acute Lymphoblastic Leukemia (ALL) is the rapidity with which discoveries in the laboratory are translated into innovations in clinical care. In 1957, just a few years after murine studies demonstrated that rescue from radiation induced marrow failure was by cellular not humoral factors [1,2], E. Donnall Thomas reported on the transfer of bone marrow cells into irradiated leukemia patients [3] and in 1965 George Mathe first described a patient with ALL whose leukemia was cured by total body irradiation (TBI) and an HCT(4). While this patient succumbed to Graft versus Host Disease (GvHD) the report is credited with being the first description of a graft versus leukemia (GvL) effect. Despite the pivotal nature of these findings, early human transplants were uniformly unsuccessful [5] highlighting the issues that continue to thwart transplanters today – leukemic relapse, regimen related toxicity, and GvHD. The next step in the clinical application of transplantation came in understanding the role of self-antigens in mediating tolerance [6]. Understanding the inheritance pattern of human self-antigens [Human Leukocyte Antigens (HLA)] enabled development of methods to identify first related and ultimately unrelated individuals who were HLA matched with each other. The first successful transplant from an HLA matched unrelated donor was carried out in 1979 for a patient with acute lymphoblastic leukemia [7]. This patient carried two common HLA haplotypes, and finding a match required typing of only 5 individuals identified through blood donation, but more sophisticated typing techniques led to the ability to identify matched and partially matched unrelated adult and cord blood donors. Additionally, this early success was part of the impetus for the development of worldwide unrelated donor registries. Since the first unrelated donor transplant facilitated by the National Marrow Donor Program in 1987, patients identified as potential candidates for transplant have had a growing number of options for that transplant. Here we review current considerations and future directions for both adult and pediatric patients around transplant for acute lymphoblastic leukemia, including issues of indication for transplant, donor selection, cytoreductive regimens, and outcomes.

Approximately 80–90% of adults with acute lymphoblastic leukemia (ALL) achieve an initial complete remission (CR1) in response to induction chemotherapy [8], but these remissions are frequently not durable and until recently few patients were rescued after relapse [9] and five-year leukemia-free survival (LFS) was just 30%–40% [8,[10], [11], [12]] with even more dire results in the elderly [10] and those with poor prognostic features at diagnosis. Those features include high WBC count (>30x10e9/L in B-ALL and >100x10e9/L in T-ALL), age >35 years, adverse cytogenetics (Philadelphia Chromosome Positive – Ph+) immunophenotype (B-cell inferior to T cell ALL), and presence of minimal residual disease (MRD) after induction and consolidation [9] are prognostic of treatment failure [8] with five-year LFS for adults with any of these adverse factors ranging from 11% to 33% [13,14].

In a series of early retrospective and prospective studies published between 1987 and 1994 and summarized by Finiewicz et al. [15]; in the absence of high risk disease, neither matched sibling donor (MSD) allogeneic or autologous hematopoietic stem cell transplant (HCT) offered a survival advantage over intensive chemotherapy [16,17] with increased treatment related mortality (TRM) of 20–40% masking any improvement in disease control.

An international UKALL study (ECOG-2993) assessed this prospectively [10]. A total of 1929 patients 15–59 years old with Ph- ALL in remission after multiagent induction therapy were stratified based on high risk features of high WBC at diagnosis, age, and time to achieve remission and randomized to undergo allogeneic HCT (if a MSD was available), autologous HCT or maintenance chemotherapy. Standard-risk patients with versus without a MSD had a 5-year OS rate of 53% compared with 45% (P = .01) while for patients with high-risk disease the difference in OS was 41% versus 35% (P = .2). Relapse rates were significantly lower (P < .00005) for both standard- and high-risk patients with HLA-matched donors. This study also found that autologous HCT was less effective than maintenance chemotherapy as post-remission treatment (5-year OS rate, 46% for chemotherapy vs. 37% for autologous HCT; P = .03). While this study supported the use of MSD HCT as consolidation in CR1 with patients who had a donor, concerns remained around the high TRM.

Response to second line induction therapy, and treatment of relapse, have improved dramatically due to novel therapeutic agents such as tyrosine kinase inhibitors (TKIs), blinatumomab, inotuzumab ozogamicin, and anti-CD19 directed CAR T-cell therapy [[18], [19], [20], [21], [22], [23], [24], [25]]. As a result, transplant for adult patients in CR1 is now reserved for patients with high risk disease.

The strongest rational for allogeneic HCT across indications remains the potential for a GvL effect that could synergize with cytotoxic radio-chemotherapy. Identification of a GvL effect in ALL has been more elusive than for some other malignancies of hematopoietic origin. One way to identify a GvL effect is assess for a survival advantage in HCT recipients who develop GvHD. Yeshurun and colleagues performed a CIBMTR registry analysis of 5215 HCTs performed for ALL(26). Patients in CR1/CR2, with acute GVHD or chronic GVHD had a lower risk of relapse (hazard ratio [HR], 0.49–0.69) with those with grades I and II acute GVHD without chronic GVHD experiencing the best OS compared with no GVHD (HR, 0.83–0.76) showing that GvHD was associated with an increased GVL effect in ALL. Most studies of pediatric and young adult patients that address this issue suggest an effect of both acute and chronic GVHD in decreasing relapse [[26], [27], [28], [29], [30], [31]].

One of the most important questions facing the field is whether the GvL effect can overcome other risk features – especially that of minimal residual disease (MRD). MRD detected by multichannel flow cytometry (MFC) prior to HCT identifies patients at high risk for relapse, but many pre-HCT MFC-MRD negative patients also relapse. MRD can also be assessed by more sensitive next-generation sequencing (NGS) assays. In a subset analysis of 56 patients in the ASCT0431 COG trial [31] NGS-MRD was evaluated and predicted relapse and survival more accurately than MFC-MRD (P < .0001), especially in the MRD negative cohort where there were no relapses (0% vs 16%; P = .02). Additionally, the benefit of GvL linked to GvHD was demonstrated in the 19 patients who were MRD positive prior to HCT. In this cohort 9 of 12 patients who did not develop GvHD relapsed compared 1 of the 7 who did develop GvHD. In a multivariate model, pretransplant MRD and acute GVHD were independent predictors of relapse [32].

A number of approaches to enhance the GvL effect post-HCT have been explored [33]. Trials in Europe and the United States have shown that patients defined as having a high risk of relapse based on increasing recipient chimerism or detection of MRD can tolerate withdrawal of immune suppression (RIS) or even infusion of donor lymphocytes (DLI) [34,35]. One study evaluating dropping donor chimerism demonstrated 3-year Event Free Survival (EFS) of 37% in 31 HCT recipients who had RIS and/or DLI compared to 0% (p < .001) [36]. Similarly, in a small single study RIS [37] and in a larger international study, development of acute GvHD (HR, 0.29; P < .001) decreased relapse and improved EFS in both MRD-positive and MRD-negative patients [38].

However, caution in intervening for changes in donor chimerism is required [39] as with some platforms, recipient T cell chimerism is frequently observed and is related to control of viral infections, not as a harbinger of relapse.

While initial transplants for acute leukemia were performed in patients with active disease, it was understood in the mid-1970s that superior outcomes could be achieved when patients were transplanted in remission. This paradigm shift led to improvement in survival.

Transplant is now indicated in CR1 for adult patients with ALL with high-risk features (age >40 years; WBC>30x 10e9/L, high-risk cytogenetics*, and poor response to therapy*). High risk cytogenetic designations differ by study and include: Ph+; t(4:11)(q21; q23); t [8,14](q24.1; q32); complex karyotype, low hypodiploidy; and IKZF1 deletion. High risk features in patients with T ALL are less well defined but include early T cell progenitor phenotype (ETP). Adult patients without a contraindication to transplant, and especially those with a MSD can be considered for transplant in CR1. The advantage of matched sibling donor transplant for adults with standard risk ALL was most definitively demonstrated with the MRC UKALL XII/ECOG E2993 trial [10]. However, In caution is warranted in transplant for ALL in CR1 [40] with a recent report finding superior outcomes for those who continued on conventional pediatric-style chemotherapy.

Patients with ALL who experience a relapse following chemotherapy and maintenance therapy are unlikely to be cured by further chemotherapy alone. These patients should be considered for reinduction chemotherapy followed by allogeneic BMT in CR2. Reinduction can be with chemotherapy, blinatumomab, inotuzumab, or TKI based. Investigators have recognized an increased risk of sinusoidal obstruction syndrome in those patients proceeding to HCT after inotuzumab reinduction [23].

With appropriate induction chemotherapy, not all patients with T-ALL require transplant in CR1. However, those with high-risk features including a slow response to induction therapy, should be considered for transplant. There is not agreement on how to define high risk disease, but several studies are starting to aide in this. Two Dana-Farber Cancer Institute protocols enrolled 123 patients with T-ALL. 5-year EFS and OS was 81% (95% CI, 73–87%) and 90% (95% CI, 83–94%), respectively. ETP phenotype did not confer inferior OS. Low end-induction MRD (<10−4) was associated with superior DFS. Pathogenic mutations of the PI3K pathway with inferior 5-year DFS and OS([41]). On two ALL-Berlin-Frankfurt-Münster (BFM) 90 and ALL-BFM 95 trials, 36 of 191 high risk T-ALL patients underwent HCT in CR1. The 5-year DFS and OS was 67% ± 8% and 67% ± 8% for HCT recipients and 42% ± 5% and 47 ± 5% for non-HCT cohort. (P = .01) demonstrating superior results for HCT in CR1 for those children with high-risk T-ALL([42]).

Adults and children with relapsed T-ALL should be considered for HCT. Survival for children with relapsed T cell acute lymphoblastic leukemia (T-ALL) is poor when treated with chemotherapy alone and HCT after myeloablative condition from umbilical cord blood (UCB) MSD or MUD achieved a 3-year OS ad DFS 48% (95% CI, 41%–55%) and 46% (95% CI, 39%–52%) such that HCT for this patient population is warranted [43].

Patients who do not achieve remission with appropriate induction therapy are eligible for novel agents, and consideration should be made for enrollment on clinical trials. In a large retrospective series of patients with initial induction failure, the 10-year OS rate for patients with persistent leukemia was 32% [44]. A trend for superior outcome with allogeneic HSCT, compared with chemotherapy alone, was observed in patients with T-cell phenotype (any age) and with B-ALL who were older than 6 years.

In contrast to adults, the intensity of treatment required for cure pediatric patients with ALL varies substantially. Risk-based treatment assignment minimizes the toxicity and late effects for those who can be cured with less intensive therapy [45]. Recommendations have been generated and reviewed by experts in the field [46]. Approximately 10%–20% of patients with ALL are classified as very high risk and include [47,48] 1) Infants younger than 1 year, 2) those with adverse cytogenetics (BCR; ABL, TCF3; HLF, KMT2A and extreme hypodiploidy).3) slow response to therapy (either to a steroid prophase or at end of induction (week 4) or consolidation (week 12) and 4) induction failure.

Evidence for transplant of these highest risk patients in CR1 has been generated over the past 30 years [49,50]. An early European study assigned high risk patients with MSD to HCT and demonstrated superior 5-year DFS (57% ± 7% versus 41% ± 3%, P = .02) but no OS benefit [49]. The NOPHO trial assigned patients to HCT in CR1 based on end induction MRD([51]) For patients who underwent HCT, better DFS was obtained in those who were MRD negative prior to HCT. On the DCOG ALL-10 and ALL-11 trials [52], the 5-year EFS rate was 72.8% for all patients regardless of HCT, and in a landmark analysis of EFS from the end of the third high-dose chemotherapy block, no difference was observed in the outcomes of patients who received HCT versus those who received chemotherapy only. Two retrospective analyses investigated the role of HCT in CR1 for hypodiploid ALL and did not demonstrate a benefit of HCT, but also did not assess the status of MRD prior to HCT([53]). In a study of 306 hypodiploid patients from 16 ALL cooperative groups treated between 1997 and 2013, a subset of 228 patients was analyzed and underscored the importance of MRD. Patients with low end of induction (EOI) MRD who underwent HCT achieved DFS of 73.6%, compared with 70% for those who did not (P = .81). In patients with higher EOI MRD HCT was associated with DFS of 55.9%, compared to 40.3% (P = .29). The COG analyzed 113 patients with hypodiploid ALL 61 of whom underwent HCT in CR1([54]). The 5-year EFS was 57.4% for patients who underwent HCT versus 47.8% for those who did not (P = .49) and OS was 66.2% versus 53.8% (P = .34).Patients with high MRD after induction (≥0.01%) had a very poor EFS rate of 26.7% at 5 years, with no protection by HCT.

The role of allogeneic HCT during first remission in infants withKMT2A gene rearrangements remains controversial [[55], [56], [57], [58]]. This led to a randomized study of infants considered to be high risk (KMT2A rearranged, age <6 months, and WBC ≥300,000/μL) were eligible for HCT in CR1([59]). About one-half of the eligible infants did not proceed to HCT - primarily because of early relapse. The 6-year EFS rate of the entire high-risk group was 21%. The population who received HCT had a 4-year DFS of 44%, but the benefit remains inconclusive.

The critical role for MRD in defining pre-HCT risk has been highlighted over the past several years with patients with high MRD at EOI and end of consolidation benefiting from transplant and underscoring the importance of using NGS-based assessment tools [38].

An analysis from the CIBMTR examined pretransplant variables to create a model for predicting LFS posttransplant in pediatric patients (aged <18 years). All patients were first transplant recipients who had myeloablative conditioning, and all stem cells sources were included. For patients with ALL, the predictors associated with lower LFS included age younger than 2 years, second CR or higher, MRD positivity (only in second CR, not in first CR), and presence of morphologically detectable disease at time of transplant. A scale was established to stratify patients on the basis of risk factors to predict survival. The 5-year LFS rate was 68% for the low-risk group, 51% for the intermediate-risk group, and 33% for the high-risk group [60].

Remission status at the time of transplantation has long been known to be an important predictor of outcome, with patients not in CR at HSCT having very poor survival rates [61]. Several studies including a meta-analysis [62] have also demonstrated that the level of MRD at the time of transplant is a key risk factor in children and adults with ALL in CR undergoing allogeneic HSCT [[63], [64], [65], [66]] with survival rates of patients who are MRD positive pretransplant have been reported between 20% and 47%, compared with 60%–88% in patients who are MRD negative.

When patients received two to three cycles of chemotherapy in an attempt to achieve an MRD-negative remission, the benefit of further intensive therapy for achieving MRD negativity versus the potential for significant toxicity.

Currently agreed upon indications for HCT in CR1 include 1) Not attaining remission at EOI 2) Cytogenetic abnormalities with t [4,11], iMP21, severe Hypodiploidy, MRD at Time Point Two (12 weeks) ≥ 10-3.

Currently agreed upon indications for HCT in CR2 are those with very early or early relapse. After induction treatment, MRD-based treatment stratification resulted in excellent survival in patients with late relapsed B-ALL. Prognosis could be further improved in very poor responders by intensifying treatment directly after induction. TP53 alterations can be defined as a novel genetic high-risk marker in all MRD response groups in late relapsed B-ALL.

Patients who do not proceed to transplant in CR2 but relapse, should be offered transplant in CR3([67]).

COG-AALL1721 (NCT03876769) (Study of Efficacy and Safety of Tisagenlecleucel in High-Risk B-ALL End-of-Consolidation MRD-Positive Patients): Patients in morphological CR at end of induction and have end-consolidation MRD of ≥0.01%. will receive Tisagenlecleucel (a CD19-directed chimeric antigen receptor [CAR] T cell) as definitive therapy to determine whether the 5-year DFS rate with Tisagenlecleucel therapy exceeds 55%.

AALL1631 (NCT03007147) (Imatinib Mesylate and Combination Chemotherapy in Treating Patients with Newly Diagnosed BCR::ABL1 [Ph+] ALL):High-risk patients will proceed to HCT after completion of three consolidation blocks of chemotherapy. Treatment with imatinib will restart after HCT and be administered from day 56 until day 365. The aim is to test the feasibility of post-HCT administration of imatinib and describe the outcomes of these patients.

AALL1631 (NCT03007147) (Imatinib Mesylate and Combination Chemotherapy in Treating Patients with Newly Diagnosed BCR::ABL1 ALL): This study will assess MRD after the induction IB phase (weeks 10–12) and assign high-risk patients (∼15%–20%) to HCT. Imatinib will be restarted after HCT and administered from day +56 until day +365 to test the feasibility of post-HCT administration of this agent.

CAR CURE (NCT0562129) (A Study to Evaluate Next-Generation Sequencing (NGS) Testing and Monitoring of B-cell Recovery to Guide Management Following CART Induced Remission in Children and Young Adults with B Lineage Acute Lymphoblastic Leukemia). This study will use NGS monitoring of remission after CART therapy to inform decisions regarding post-CART HCT.

Improving LFS rates in the 40–50% range support consideration of second transplant in those relapsing after an initial HCT. While these patients are eligible for novel therapies including CAR T cell therapy, conclusions about the need to consolidate remissions with HCT have not yet been made. In 251 children and young adults undergoing second transplant for relapse of leukemia 75% (n = 187) were in remission, 63% received a myeloablative conditioning regimen (n = 157), and 92% underwent unrelated donor HCT (n = 230). The 2-year LFS was 33% after transplantation in patients in remission, compared with 19% after transplantation in patients not in remission (P = .02) [68].