Hematopoietic Stem Cell Transplantation in Pediatric Oncology
Stephen Gottschalk
Swati Naik
Meenakshi Hegde
Catherine M. Bollard
Robert A. Krance
Helen E. Heslop
INTRODUCTION
Hematopoietic stem cell transplantation (HSCT) is an established treatment approach for many malignant and nonmalignant diseases that affect the hematopoietic and immune systems. The initial human transplants for hematologic malignancy took place in the 1950s and showed transient engraftment only. As increasing information became available in the 1960s about the human leukocyte antigen (HLA) system and as tissue typing methods were developed, successful transplants were described for children with immunodeficiency. Patients with advanced leukemia also underwent marrow transplantation from matched sibling donors, and a small percentage became long-term survivors.1 As transplant has become safer, it has become used earlier in the course of malignant diseases, resulting in improved outcomes.1 Over the past 15 years, indications for HSCT and cellular therapies have broadened, and the future promises even wider applications, especially in delivering novel treatments for malignancies.
A number of key advances have contributed to making HSCT a more commonly available and successful treatment modality.1 These include an improved understanding of the critical role of histocompatibility in allogeneic HSCT and the development of high-resolution molecular methods to more accurately type donors and recipients. Furthermore, additional sources of stem cells were also employed so that bone marrow (BM), peripheral blood (PB), and umbilical cord blood (UCB) are all now widely used in clinical practice to provide long-term hematopoietic reconstitution. These advances, along with the increasing numbers of donors in large registries of unrelated donors, and the increased availability of cord blood units, expanded access to transplantation and allowed recipients to find more closely matched donors. More recently, approaches to either deplete T cells ex vivo or to deplete alloreactive cells in vivo have resulted in improved outcomes for haploidentical transplants, further increasing donor options.2 Finally, there have been improvements in graft-versus-host disease (GVHD) prophylaxis and supportive care during the period of hematopoietic and immune suppression post-transplant. HSCT should therefore be considered for patients in whom this procedure is likely to result in superior long-term disease-free survival (DFS) compared with other therapeutic modalities. Potential candidates must have a suitable source of hematopoietic stem cells (HSCs) available at an appropriate time in the course of the treatment for their disease.
In this chapter, we review the current status of HSCT in pediatric oncology and indications for its use in patients with acute lymphoblastic leukemia (ALL), acute myelogenous leukemia (AML), chronic myelogenous leukemia (CML), myelodysplasia (MDS) or myeloproliferative syndromes, non-Hodgkin lymphoma (NHL), Hodgkin disease (HD), and neuroblastoma as well as other solid tumors. In addition, we provide an overview of various HSCT procedures, including conditioning regimens, selection of donor and HSC source, and common early- and late-onset posttransplant complications.
ALLOGENEIC TRANSPLANTATION
In allogeneic transplantation, the recipient receives HSC from a closely matched donor, and alloantigens that differ between donor and recipient are targets for T-cell recognition. The most important criterion for choosing an allogeneic donor is the degree of histocompatibility with the recipient. With increasing genetic differences, there is an increased risk of both graft rejection and of graft-versus-host disease, although there may also be an increased graft-versus-tumor effect. The most important determinant of alloreactivity is matching at HLA loci, but even when major histocompatibility complex (MHC) antigens are identical, minor histocompatibility antigens, which are naturally processed peptides derived from normal cellular proteins, may evoke a strong MHC-restricted response when different polymorphisms are present in donor and recipient. Natural killer (NK) cells may also contribute to alloreactivity, particularly in the setting of haploidentical transplantation.
HLA Matching
Identification of HLA antigens and the human MHC was a prerequisite for transplantation of BM between family members. HLA antigens are cell surface molecules encoded by class I (HLA-A, HLA-B, and HLA-C) and class II (HLA-DR, HLA-DQ, and HLA-DP) genes, a series of closely linked loci known as the MHC on chromosome 6. These antigens present peptides to T lymphocytes and are highly polymorphic. Historically, HLA molecules were typed by alloantisera in complement-dependent cytotoxicity assays. However, serologically identical donors and recipients can have major genotypic differences not detected by this methodology that will be readily detected by alloreactive T cells. Analysis by gene sequencing has revealed multiple alleles for most serologically defined specificities, and more than 3,000 alleles have now been recognized. HLA terminology is designated by the World Health Organization Nomenclature Committee for Factors of the HLA system and is updated at regular intervals.3 The nomenclature distinguishes the technique used to determine the HLA type and its level of resolution. The broadest designation is based on serologic typing, whereas the highest resolution is based on actual DNA sequence. For example, “HLA A2” designates a specific class I serotype (i.e., determined by reactivity to serologic reagents); antigens (“HLA A2”) recognized by these reagents can be subdivided into HLA A*02 by additional serologic reagents or by low-resolution molecular techniques. In turn, HLA A*02 can be further classified by intermediate molecular analysis as A*0201/A*0205 or by high-resolution molecular analysis as HLA A*0201. This is incorporated into the nomenclature, resulting in a report confirming the allele as HLA A*0201 on high-resolution typing. The type of assay used and its sensitivity and specificity are important considerations in determining potential histocompatibility, particularly for mismatched family or for any unrelated donor HSCT.4
The genetic unit of HLA class I and II regions on one chromosome is referred to as an HLA haplotype, and the two HLA
haplotypes in one person are called the HLA genotype. Class I and class II antigens are codominant and are transmitted as dominantly inherited Mendelian traits. Each child expresses one set of paternal and one set of maternal HLA antigens corresponding to the HLA genes inherited as one paternal and one maternal HLA haplotype. The probability that a child will inherit any one of the four possible HLA genotypes is 0.25. When, by chance, an individual inherits a phenotypically identical HLA allele from each parent, the result is a person homozygous for that locus; a person may be homozygous for one allele at a single HLA locus or for the entire haplotype. Genetic recombination between HLA class I and II regions also occurs infrequently. Thus, although a careful examination of family haplotypes is essential to determining donor suitability, the HLA complex can generally be considered as a single genetic unit that is most often inherited as a block.
haplotypes in one person are called the HLA genotype. Class I and class II antigens are codominant and are transmitted as dominantly inherited Mendelian traits. Each child expresses one set of paternal and one set of maternal HLA antigens corresponding to the HLA genes inherited as one paternal and one maternal HLA haplotype. The probability that a child will inherit any one of the four possible HLA genotypes is 0.25. When, by chance, an individual inherits a phenotypically identical HLA allele from each parent, the result is a person homozygous for that locus; a person may be homozygous for one allele at a single HLA locus or for the entire haplotype. Genetic recombination between HLA class I and II regions also occurs infrequently. Thus, although a careful examination of family haplotypes is essential to determining donor suitability, the HLA complex can generally be considered as a single genetic unit that is most often inherited as a block.
The initial BM donors were siblings who shared the patient’s HLA genotype. Historically, only approximately 30% of patients have an HLA-matched sibling donor, so that the remaining 70% have to consider alternative donors such as an HLA-mismatched family member, a closely matched unrelated donor, or a cord blood unit. Development of more precise tissue typing methods using molecular techniques and establishment of large donor registries have facilitated transplants from closely HLA-matched unrelated donors (MUDs).
Minor Antigens and Genetic Loci Outside the MHC
The degree of alloreactivity between donor and recipient is also influenced by differences at minor histocompatibility loci. These genes encode polymorphisms of normal cellular genes and have only been characterized in humans over the last 10 to 15 years.5 Increasing numbers of minor antigens have been identified, and molecular typing for some of these antigens is becoming available. Although less GVHD has been reported in matched sibling transplants when minor antigens are matched, other studies have shown that T cells recognizing minor antigens differentially expressed between donor and recipient mediate antileukemia activity. There is also increasing evidence that genetic loci outside of the MHC may influence the risk of transplant complications such as infection or bronchiolitis obliterans and several groups are undertaking genome-wide association assays to define genetic variants that might predict these complications.5
AUTOLOGOUS TRANSPLANTATION
The rationale for autologous transplantation is that dose intensification will increase the response rate of chemosensitive tumors. Hematopoietic toxicity is a limiting factor for dose intensification, which can be overcome by harvesting HSC and then cryopreserving and reinfusing them after doses of chemotherapy and radiotherapy that would otherwise be lethal or require a prolonged period of recovery. While randomized trials or evidence-based reviews have shown improved DFS for adult patients with NHL and multiple myeloma, no such benefit was observed for adult patients with solid tumors such as breast cancer.6,7
SOURCE OF HSCs
Initial studies used HSCs derived from BM for transplantation. Over the past 20 years, with the availability of cytokines or chemokine inhibitors that can mobilize HSC into the bloodstream, there has been increasing use of mobilized PB as a source of stem cells.
Autologous Donors
In autologous transplantation in adults, mobilized PB is now the most commonly used product, as it results in faster engraftment. Its use has also increased in pediatrics, although logistical challenges with pheresis in children weighing less than 15 to 20 kg mean that marrow is still more frequently used. Cytokine-mobilized peripheral blood stem cells (PBSCs) can be harvested either after treatment with cytokines alone (most commonly, recombinant human granulocyte colony-stimulating factor, 10 to 16 µg/kg/day for 3 to 7 days) or with cytokines given during recovery from chemotherapy. An alternative strategy is to use the chemokine receptor 4 antagonist plerixafor and granulocyte colony-stimulating factor in combination.8
Allogeneic Donors
When an allogeneic transplant is indicated, the family is initially typed to determine whether there is an HLA-matched sibling. If one of the siblings matches, this is almost always the preferred donor. Rarely, patients will have an identical twin who can serve as the donor. In this setting, there is no alloreactivity, and the patients do not require posttransplant immunosuppression, but they have a higher risk of relapse, particularly if they have a myeloid malignancy. The issue of consenting children by proxy has been debated, and the American Academy of Pediatrics has issued a policy statement that calls for an independent advocate,9 but parental consent for the sibling donor is generally considered to be in the donor’s best interests.10 If there is no HLA match in the family, options include the use of a closely matched unrelated donor, a mismatched family member, or a cord blood unit. All potential allogeneic donors undergo an extensive medical evaluation; in the case of UCB donation, the mother serves as a “surrogate,” and the evaluation is adjusted appropriately. Physical examination and screening laboratory tests with complete blood cell count, biochemistry profile, hepatitis screen, and other testing for transmissible infectious agents, including cytomegalovirus and human immunodeficiency virus, should be completed. Many donor candidates will have preexisting medical problems that require further evaluation. BM donors are usually admitted to the hospital the morning of the harvest. The aspiration procedure is conducted in an operating room under sterile conditions and with appropriate anesthesia. Marrow is usually harvested only from the posterior iliac crests, but when the recipient is significantly larger than the donor or when large cell volumes are needed, BM may also be harvested from the anterior iliac crests. The total volume of marrow usually collected amounts to 10 to 20 mL per kg of recipient weight to obtain sufficient HSCs for engraftment. BM from children, especially infants, has a higher concentration of nucleated cells and probably a higher proportion of marrow-repopulating cells than marrow from older donors.
As with autologous transplant, cytokine-mobilized allogeneic PBSC harvest has become an alternative to marrow as a source of hemopoietic stem cells (HSCs). Early phase II studies showed that this source of HSCs resulted in faster engraftment, no increase in acute GVHD (perhaps due to a granulocyte colony-stimulating factor (G-CSF)-mediated shift to Th2 helper cells), but an increased incidence of chronic GVHD. In adult patients, both a single-center randomized study and International Bone Marrow Transplant (IBMT) registry data have shown that treatment-related mortality rates were lower and leukemia-free survival rates were higher with use of blood SC transplants in patients with advanced leukemia, although a difference was not seen for patients with better risk disease.11,12 However, a retrospective registry review suggested that in pediatric transplants for acute leukemia the outcome may be worse after PBSC transplant after adjusting for relevant risk factors.13 Recently, the initial report from a large national study from the BMT Clinical Trials Network (CTN) showed no significant difference in outcome when blood or marrow from unrelated donors was used, but a higher risk of chronic GVHD with mobilized PB.14
Donor safety is obviously an important issue for pediatric PBSC donors. A review from the Pediatric Bone Marrow Consortium of more than 200 donor collections found that PBSC collection was
safe in normal pediatric donors, that target CD34 cell yields were easily achieved, and that children weighing less than 20 kg usually require a single blood product exposure.15 More recently, a European study in 453 donors confirmed that PBSC and BM collection are safe procedures in children.16 Although this approach has not shown short-term toxicity, longitudinal studies are needed to establish whether there are long-term toxicities that might raise ethical issues for donor safety. It is reassuring that a recent report from a prospective study through the National Marrow Donor Program showed that the incidence of cancer, autoimmune illness, and thrombosis after donation was similar in BM versus PBSC donor receiving G-CSF.17
safe in normal pediatric donors, that target CD34 cell yields were easily achieved, and that children weighing less than 20 kg usually require a single blood product exposure.15 More recently, a European study in 453 donors confirmed that PBSC and BM collection are safe procedures in children.16 Although this approach has not shown short-term toxicity, longitudinal studies are needed to establish whether there are long-term toxicities that might raise ethical issues for donor safety. It is reassuring that a recent report from a prospective study through the National Marrow Donor Program showed that the incidence of cancer, autoimmune illness, and thrombosis after donation was similar in BM versus PBSC donor receiving G-CSF.17
Unrelated Donors
Volunteer unrelated donors include healthy persons between 18 and 60 years of age who fulfill health requirements similar to those applied to blood donors. As discussed previously, the outcome after unrelated transplant correlates with the degree of matching and mismatching at HLA class 1, class II, and HLA-C loci are all associated with poorer outcome.18 With increasing registry size, the chance of finding a donor has increased, so that more than 70% of patients undergo unrelated donor transplantation using an HLA-A, B, C, and DRB1 allele-matched unrelated donor.19 The likelihood of finding a donor matching at these 8 loci (or 10 loci if matching at DQB1 is also included) varies for different ethnic groups and is less for groups with more polymorphism of HLA antigens.
Historically, the outcome after transplantation from unrelated donors has been inferior to that observed after matched sibling transplantation because of an increased incidence of graft rejection and of GVHD resulting from increased alloreactivity in this setting. Over the past few years, improved results have been reported from several single-center and multicenter studies in defined patient populations, reflecting improvements in donor-recipient matching, GVHD prophylaxis, supportive care, and the timing of transplantation.20
Umbilical Cord Blood
Another alternative source of SCs is cord blood. There are several large cord banks where cord blood is collected, cryopreserved, and tested for infectious agents in accordance with standards developed by governmental and specialty oversight organizations. The immediate availability of cryopreserved cord blood units eliminates the usual delay in HSCT when unrelated donor marrow is used. After several studies demonstrated the feasibility of transplants with cord blood from unrelated donors, cord blood became an increasingly common source of HSCs for pediatric patients requiring transplants.21 Such transplants have slower engraftment, but they may also induce less GVHD owing to the relative naivety of cord T cells. A registry study comparing the outcome of cord and marrow transplants in children with acute leukemia found that 5-year leukemia-free survival was similar for recipients of allele-matched marrow and cord blood mismatched for either one or two antigens.22 The speed of myeloid engraftment is associated with the mononuclear cell count of the graft, whereas transplantation-related events were associated with the patient’s underlying disease and age, the number of leukocytes in the graft, and the degree of HLA disparity. While the cell count can be limiting for older children, the use of double cord blood transplantation or expansion techniques offers alternative options in individuals weighing more than 45 kg.21
Haploidentical Family Donors
The genetic sharing of one chromosome of the chromosome 6 pair, containing the complete DNA code for the MHC, makes a haploidentical donor in essence a “half-matched” donor. There is also a greater likelihood for identity between minor histocompatibility antigens expressed from other chromosomes than could be expected between unrelated individuals. This opens another potential source of donors that increases the access to allogeneic HSCT. In addition, it is more likely that a haploidentical family donor will be readily available compared with the longer task of identifying an unrelated donor. Use of a mismatched family member donor, however, is associated with an increased risk of GVHD due to increased alloreactivity, and this risk increases with the degree of mismatch. Most studies therefore show that transplants from donors mismatched in a single antigen produce results equivalent to those achieved with matched sibling donors. For greater degrees of mismatch, methods have been developed to manipulate or engineer hematopoietic progenitor cells (HPC) ex vivo or in vivo, to eliminate the cells that are thought to mediate alloreactivity. One strategy is to use G-CSF-mobilized, large-volume apheresis and CD34 selection with or without additional T-cell depletion. In a series of studies in children, Handgretinger and colleagues have shown that GVHD could be effectively reduced and primary engraftment attained in 83% to 100% of patients after transplantation of high stem cell doses.23 Another approach is to selectively deplete T-cell populations allowing transfer of NK cells and other CD3-negative cells present in the infused product.24 In several recently published pediatric series, selective T-cell depletion has resulted in 2-year event-free survival of 26% to 88% depending on the regimen and risk of the underlying hematologic malignancy.25 Investigators at Johns Hopkins have pioneered a strategy of infusing unmanipulated marrow followed by high-dose, posttransplantation cyclophosphamide (50 mg/kg on days 3 and 4) to eliminate alloreactive cells in vivo.2 In a phase II BMT CTN study in patients with leukemia or lymphoma, the 1-year probabilities of overall and progression-free survival were 62% and 48%, respectively. Other groups have administered unmanipulated marrow and relied on in vivo agents such antithymocyte globulin (ATG) to eliminate alloreactive cells. Several groups in China have administered unmanipulated marrow or mobilized PB and relied on ATG to eliminate alloreactive cells.26 In a report of 820 patients with hematologic malignancies receiving G-CSF-primed marrow or G-CSF-mobilized PB from haploidentical donors, the 3-year leukemia-free survival rates were 67.9% in standard-risk and 48.8% in high-risk patients.26 It is difficult, however, to compare these different strategies for haploidentical regimens as they have targeted different populations and diseases.
TRANSPLANT FOR LEUKEMIA
Acute Lymphoblastic Leukemia
Cure rates for children with newly diagnosed ALL are projected to reach nearly 90% following current primary chemotherapy treatments.27 With improvements in treatment for children with ALL, the role for stem cell transplantation as primary therapy has been reduced. The indications for transplant continue to evolve as the risks for treatment failure are refined.
ALL—First Remission
Indications for transplant in patients in initial complete remission (CR1) are generally assigned on the basis of risk stratification. Definitions of high risk and therefore indications for transplant vary considerably; however, most agree that patients who have poor outcomes with chemotherapy alone and are at high risk of treatment failure or relapse with 5-year event-free survival (EFS) of less than 50% should be considered candidates for transplant. These very high-risk ALL patients include those who have failed induction, infants with MLL rearrangements, those with T-cell ALL who have a poor early response to therapy, and those with high-risk cytogenetic features such as hypodiploidy (chromosome number <44) or Philadelphia chromosome positive (Ph+) ALL.28
Induction failure is associated with a very poor prognosis, with a reported survival of 30% or less.29 Patients with induction failure have been considered candidates for transplant as early studies showed significant improvement in 5-year DFS for transplant compared with chemotherapy.30 It is possible that not all patients with induction failure require transplant to achieve long-term remission. A recent analysis of 1,041 pediatric patients with induction failure treated between 1985 and 2000 showed that allogeneic transplant was beneficial over chemotherapy for patients with T-cell ALL and for older patients with precursor (pre-)B-cell ALL, while younger patients with pre-B-cell ALL fared better with chemotherapy.29
Induction failure has conventionally been defined by morphological assessment of disease. More sensitive methods, such as polymerase chain reaction (PCR) or flow cytometric-based assays can detect minimal residual disease (MRD) in as few as 1 in 100,000 malignant cells. These assays have become increasingly important as indicators for treatment failure.31 End induction MRD has been shown to be the single most important prognostic factor able to discriminate outcomes for patients with pre-B-cell ALL and T-cell ALL.32,33,34 Recent studies have incorporated MRD for risk stratification and assignment to transplant. This remains an active area of study.35
Despite the use of maximally intensified chemotherapy regimens, infants with ALL, in particular those <6 months of age with mutations involving 11q23, the MLL gene, and poor response to induction therapy, have very poor outcomes.36 The role for upfront transplant has been evaluated, but several trials have failed to show benefit of transplant over chemotherapy (Table 16.1).37,38 The lack of benefit has been largely due to high transplant regimen-related mortality. For chemotherapy, relapse remains the principal cause of failure. A recent study from the Interfant-99 group showed a significant difference in DFS (adjusted for waiting time to transplant) between those who received transplant versus those who received chemotherapy (5-year DFS of 59% vs. 22% p = 0.01). However, this advantage was restricted to the subgroup of infants with MLL gene rearrangements who were <6 months of age with poor response to steroids and/or presenting WBC > 300 × 109/L. For other infants with the MLL gene rearrangements but no other additional risk factors, there was no benefit to transplant over chemotherapy (5-year DFS of 60.9% vs. 53.8%, p = 0.09).39 One major drawback to transplant of infants is the concern for radiation-related growth failure and neurocognitive development. Reports have been mixed regarding the impact of radiation in these areas, but these concerns must be satisfactorily addressed.40
Hypodiploid ALL (<44 chromosomes) has a very poor outcome, and more than two-thirds of patients are likely to relapse. In a large multicenter study of patients with hypodiploid ALL, the 8-year EFS was significantly worse for patients with fewer than 44 chromosomes compared with patients with 44 chromosomes at 30.1% versus 52.2% (p = 0.01).41 In a retrospective review of the Center for International Blood and Marrow Transplant Research (CIBMTR) experience, the 5-year DFS was 51% in 78 patients who underwent transplant for hypodiploidy, suggesting that transplant may improve outcomes in these patients.42
Prior to the use of tyrosine kinase inhibitors (TKIs), for patients with Ph+ ALL, there was a clear benefit for transplant.43 With the incorporation of TKIs into upfront therapies for Ph+ ALL, a recent report from the Children’s Oncology Group (COG) showed long-term leukemia-free survival (LFS) was comparable between patients receiving imatinib-based chemotherapy versus transplant (70% vs. 65% for matched related donor and 59% for unrelated donors).44 While further studies are needed, current clinical trials do not utilize transplant for patients in CR1 unless there is evidence of persistent MRD.
Similarly, outcomes for T-cell ALL have improved with current therapies.45 However, patients who have a poor response to initial therapy may achieve superior outcomes with transplant (5-year DFS 67%), compared with chemotherapy (5-year DFS 42%).46 Therefore, transplant is recommended for T-cell ALL in CR1 when there is poor initial response to therapy. The recently described early T-cell precursor-ALL (ETP-ALL) is characterized by a distinct genetic and immunophenotypic profile, poor response to chemotherapy, and a very high risk of relapse. The overall survival for this subtype is significantly worse than for other patients with T-cell ALL.47 Results from a recent study show that transplant was able to salvage patients with ETP-ALL who relapsed, suggesting there may be a role for upfront transplant for this resistant T-cell ALL.48 Further studies will be needed to address this question.
Consortium studies have evaluated the role of transplant for patients with high-risk ALL in first remission, and these are summarized in Table 16.1.29,30,35,44,46,49,50 It is important to note that most studies comparing transplant with chemotherapy are not randomized controlled studies. To compare treatments, statistical analyses are used to adjust for treatment variables such as time to transplant. Outcome can be biased by the time and the logistics associated with finding a suitable allogeneic donor, with the result that children with high-risk ALL may relapse before transplantation can be performed. It is unclear whether this is a failure of chemotherapy or transplantation. To adjust for such biases, statisticians utilize “intent-to-treat” analysis, in which all transplant-eligible patients are considered as transplant, regardless of whether transplant is performed. “Time-adjusted” analysis excludes from the analysis any relapses that occur during the time interval to transplant. These statistical manipulations are never as satisfactory as randomized controlled studies.
Because of its diverse patient makeup and large patient numbers, registry data provide valuable perspective to transplant outcomes. For patients younger than 20 years with ALL in CR1, the CIBMTR reports a 65% overall probability of survival at 3 years after matched related donor transplant. After unrelated donor transplant, the overall probability of survival of patients aged 0 to 11 is 68% and of those aged 11 to 20 is lower at 48%.
The transplant conditioning regimen was originally intended to overcome leukemia cell resistance. Although graft versus leukemia (GVL) may be operative for some patients, much of the curative potential of transplantation for ALL comes from the conditioning therapy. It is reasonably well established that for patients with ALL, the ablative therapy, which includes total body irradiation (TBI), is superior to chemotherapy-only regimens.51 Most centers use a backbone of fractionated TBI to which either cyclophosphamide or etoposide (in European centers) are added. Recent multicenter trials have evaluated the role of reduced intensity conditioning (RIC) regimens for patients who cannot undergo myeloablative conditioning, and while results appear promising, long-term studies are required before this strategy can be accepted.52,53
ALL—Relapse
The principal role of stem cell transplantation in pediatric ALL therapy has been salvage after relapse. The decision to recommend transplant varies among investigative groups, but is generally based upon prognostic criteria such as length of remission, site of relapse, and leukemia immunophenotype. A number of studies have shown that the length of remission is the most important predictor of survival following relapse.54,55,56
Data from a collaborative study of the COG and CIBMTR showed that for children with early first relapse (<36 months), the risk of a second relapse was significantly lower after TBI-containing transplant regimens versus conventional chemotherapy (relative risk [RR] 0.49, p < 0.001). For children with a late first relapse (≥36 months), the risks of second relapse were similar after TBI-containing regimens and conventional chemotherapy (RR 0.92, p = 0.78).51 The Dutch COG study found that for patients with early relapse (defined as <30 months after attaining remission), transplant led to superior 5-year EFS compared with conventional chemotherapy (25% vs. 0%), whereas for patients with late relapse (defined as relapse >30 months after attaining remission) the
reverse was true, with 5-year EFS of 16% for transplant versus 65% for conventional chemotherapy.57
reverse was true, with 5-year EFS of 16% for transplant versus 65% for conventional chemotherapy.57
TABLE 16.1 Hematopoietic Stem Cell Transplantation for Acute Lymphoblastic Leukemia | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Consensus is less as to the prognostic importance of the site of relapse. Several studies have reported that outcomes after isolated extramedullary relapse were superior to BM relapse or combined BM and CNS relapse,55,58 whereas other reports showed no difference in outcome.59 Leukemia immunophenotype also affects outcomes after relapse as patients with T-cell ALL have worse outcomes compared with those with B-cell ALL.54 Thus, T-cell ALL patients are referred to transplant in second complete remission (CR2) regardless of time to relapse. Newer studies suggest that underlying cytogenetic changes also influence outcome after relapse. For example, the presence of ETV6/RUNX1 is associated with excellent prognosis, whereas patients with IKZF1 and P53 mutations have very poor outcomes after relapse.60,61
Several multicenter clinical trials have attempted to compare outcomes for children in CR2 treated with transplantation or chemotherapy (Table 16.1).55,57,62,63,64,65,66,67 Recent studies continue to show benefit for patients with high-risk features. In a recent study from the Berlin-Frankfurt-Münster (BFM) collaborative group, the 10-year EFS was significantly higher for transplant (33%) compared with chemotherapy (20%) for patients with early BM or combined relapse and for patients with T-cell ALL.55 The findings of these and other studies54 support the practice to recommend transplantation for children whose relapse occurs while on therapy or within 6 months of completing therapy and for patients with relapsed T-cell ALL as these patients are unlikely to be cured with chemotherapy alone. For patients with late relapse or when relapse is limited to extramedullary sites, chemotherapy alone can be used as salvage therapy.
More recently, it has been found that end of reinduction MRD identifies an additional subset of patients with late relapses of B-cell ALL who may benefit from transplant. These are patients who otherwise would not have met criteria for transplant.68,69 Based on similar studies,70,71 several cooperative groups are evaluating end-reinduction MRD to risk stratify patients with relapsed pre-B ALL to determine indication for transplant.69,72
Pretransplant MRD appears to adversely influence outcome. In patients with high pretransplant MRD (≥10-4 leukemic cells) versus low MRD (≤10-4 leukemic cells), the probability of EFS was significantly lower (0.27 vs. 0.60, p = 0.04), and cumulative incidence of relapse was significantly higher (0.57 vs. 0.13, p = 0.01).73 For now, it is unclear how treatment should be tailored for patients with persistent detectable MRD before transplant. Additional aggressive therapy may not reduce persistent disease or may lead to complications, which ultimately are prohibitive for transplantation. Newer therapies such as the use of monoclonal antibodies directed against leukemic cells or T cells expressing chimeric antigen receptors offer novel approaches to reduce disease burden while minimizing toxicities.74
Data from a COG/Pediatric Bone Marrow Transplant Consortium (PBMTC) trial showed that in addition to detection of pretransplant MRD, the presence of posttransplant MRD is associated with a 14-fold increase in relapse rate (RR 14.3, 95% confidence interval [CI]: 6.1 to 33.6). Interestingly, while the effect of a GVL for ALL is controversial, patients who did not develop grade II-IV acute GVHD had higher levels of posttransplant MRD and relapsed nearly 5 times more often than those with acute GVHD (30% vs. 6.4% at 2 years).75
Data from the CIBMTR show that the 3-year probability of survival for children in CR2 aged 0 to 11 and 11 to 20 following matched related donor transplant is 61% and 46%, respectively. For similar patients of ages 0 to 11 and 11 to 20 following unrelated donor transplantation, the 3-year probability of survival is 49% and 43%, respectively. While these results indicate that stem cell source determines outcome, recent single-center studies report comparable outcome for patients undergoing transplant from matched related donors, and matched unrelated marrow and cord donors.64,65 In addition, haploidentical donor transplants are also increasingly being used with newer graft engineering techniques successfully employed to maximize a GVL effect while minimizing GVHD.23,76
Similar to patients with BM or combined relapses, the prognosis for patients with isolated extramedullary relapse, most commonly CNS relapse, is correlated with the duration of initial remission. The United Kingdom (UK) ALL group reported 20% EFS for children with isolated CNS relapses when initial remission was <18 months.59 The COG and CIBMTR retrospectively compared the outcome for children with CNS relapse and found EFS was similar, 66% and 58%, after chemotherapy/radiation and transplantation, respectively.66 Regardless of therapy, initial remission lasting <18 months was associated with inferior EFS, as half the patients developed a subsequent occurrence. As the outcomes for chemotherapy and transplant are similar following isolated CNS relapse and data show that for patients with isolated late relapse (>18 months), treatment with chemotherapy alone may be sufficient,77 transplant is generally not recommended for patients with isolated CNS relapse.28 There are, however, insufficient data to make recommendations for patients with isolated CNS relapse of T-cell ALL.
Transplantation for children beyond second remission can be effective in selected cases. Data from the Nordic Society for Pediatric Hematology and Oncology (NOPHO) group reported a 10-year survival of 37% for patients receiving transplant in CR3. Patients who were treated less aggressively or who had longer durations of CR1 and CR2 fared better.78 The CIBMTR reported 5-year estimates of LFS and NRM of 30% and 45%, respectively, for patients undergoing transplant in CR3. The risk of relapse was lowest when both first and second relapses were late. However, survival was impacted by high treatment-related mortality (TRM).79 Therefore, although ALL in CR3 has a poor prognosis, transplant is usually recommended as it can improve survival.
Acute Myeloid Leukemia
Acute Myeloid Leukemia—First Remission
Outcomes for children with AML have also improved with newer therapies, and up to 65% of children are long-term survivors.80 Transplant has been used as consolidative therapy for children with AML in first remission, but in the absence of randomized controlled trials, comparison with chemotherapy suffers from similar limitations as described above. In addition, better understanding of the biology of AML, in particular the relationship between cytogenetic and molecular features and treatment outcome, has prompted a reconsideration of the indications for transplantation.
AML characterized by the presence of t(8;21) or inv(16) or mutations in the NPM and CEBPA genes81 is usually responsive to chemotherapy and is classified as low-risk. On the other hand, the presence of monosomy or deletions of chromosome 5 and 7 or mutations in FLT3/ITD81 indicate high risk with poor outcome with conventional chemotherapy. For the majority of cytogenetic changes, there is no clear association with favorable or unfavorable outcomes. Patients in this group are classified as “standard or intermediate risk.” Similar classification schemes are utilized by most international pediatric cooperative groups, and while differences exist, the major risk features are widely agreed upon.
Previous studies compared the utility of chemotherapy versus transplant for patients in CR1; risk stratification was not used to allocate treatment, and all patients with matched related donors were eligible for allogeneic transplant in CR1 regardless of risk stratification. These studies showed benefit for transplant overall specifically in preventing relapse.82,83 Importantly, some of the early studies showed no benefit for transplant over chemotherapy in patients with low-risk cytogenetic features, thereby allowing for the current practice of less intensive therapies.84 On current treatment protocols, transplant is reserved for these patients in second remission after relapse. For patients with standard risk,
several groups recommend transplant for those patients who have a matched sibling donor. However, unlike for low-risk disease, there is less consensus regarding transplant, with some studies showing benefit and others not. A risk-stratified meta-analysis of 1,373 patients treated on various clinical protocols found that for standard-risk patients, allogeneic transplant from matched related donors in CR1 was superior to chemotherapy, DFS 58% versus 39%.85 In contrast, when data from UK Medical Research Council (MRC) AML 10 and 12 trials were combined together, there was no statistically significant benefit noted for patients for any risk group.82 It is not clear whether all patients with standard-risk disease will benefit from transplant. Ongoing clinical trials that further risk stratify these patients on the basis of response to therapy will help answer this question.
several groups recommend transplant for those patients who have a matched sibling donor. However, unlike for low-risk disease, there is less consensus regarding transplant, with some studies showing benefit and others not. A risk-stratified meta-analysis of 1,373 patients treated on various clinical protocols found that for standard-risk patients, allogeneic transplant from matched related donors in CR1 was superior to chemotherapy, DFS 58% versus 39%.85 In contrast, when data from UK Medical Research Council (MRC) AML 10 and 12 trials were combined together, there was no statistically significant benefit noted for patients for any risk group.82 It is not clear whether all patients with standard-risk disease will benefit from transplant. Ongoing clinical trials that further risk stratify these patients on the basis of response to therapy will help answer this question.
Recently, the COG demonstrated that for patients without known cytogenetic or molecular markers, previously designated “standard-risk” (SR), the presence of MRD at the completion of induction was able to further risk stratify patients. The presence of MRD after induction was associated with significantly worse survival compared with those who were negative for MRD (26% vs. 67%).86 This finding has led to the use of a new classification system in the current frontline COG AML trial to define two risk groups, low risk and high risk. With an expected overall survival (OS) of 71%, the low-risk group comprises 73% of patients with AML and includes those with favorable cytogenetic/molecular features and patients with SR that are MRD negative at end of induction. The high-risk group has an expected OS of 35%, and includes those with unfavorable cytogenetic/molecular features and those with SR that are MRD positive at the end of induction. In ongoing COG clinical trials, only patients with high risk are referred to transplant in CR1 with the best available donor.
Because of the poor outcomes for patients with high-risk disease, transplant is generally recommended in first remission by most centers or collaborative groups such as COG. However, some centers in Europe have reserved stem cell transplantation solely as salvage therapy for relapsed patients or for those with refractory AML based on the results of the BFM-AML 98 trial, in which there was no significant difference in 5-year DFS for children with high-risk AML following treatment with SCT versus conventional chemotherapy. This was an intent-to-treat analysis, with 5-year DFS of 43% versus 47% for patients with and without a donor, respectively.87
In other recent studies, investigators have refined the definition of high-risk and shown a benefit for transplant88 when performed in the context of new risk stratification protocols, several using MRD-based methods to refine stratification.89 Although the St. Jude AML-02 trial did not demonstrate a benefit for transplant in the high-risk group overall, those high-risk patients with MRD >1% at the end of induction had improved outcomes with transplant.89 In the Italian Pediatric Hematology and Oncology Association (AIEOP) AML 2002/01 trial, 242 patients with HR-AML were allocated to transplant. The OS, EFS, and DFS for HR patients were 64%, 53%, and 62%, respectively with low rates of relapse and TRM.90 In the NOPHO-AML 2004 trial, “response guided” time-intensive reinduction therapy followed by transplant for poor responders led to an EFS of 70% at a median follow-up of 2.6 years.91 Similar results of improved survival for patients with high-risk disease have been shown in other large cooperative group and single-center studies, as summarized in Table 16.2.
In addition, while older trials were associated with significant TRM, recent trials have shown improvements in OS and a decrease in relapse rates without an associated increase in TRM.90
A recent study evaluated 190 children with very high-risk leukemia who underwent allogeneic transplant in two sequential treatment eras. For children with high-risk AML, the 5-year OS rates in the recent cohort treated on contemporary protocols compared favorably to children treated in the previous era (5-year OS, 74% vs. 34% respectively, p < 0.0001). Importantly, this improvement was observed regardless of donor type and was attributed to decreased TRM, with a significantly lower hazard of death from infections (HR 0.12, p < 0.005) and regimen-related toxicity (HR 0.25, p < 0.002) while also being associated with a decreased risk of relapse (HR 0.4, p = 0.013).92 It is likely that with improvements in the transplant procedure, outcomes for patients with high-risk disease undergoing transplant will continue to improve.
A recent study evaluated 190 children with very high-risk leukemia who underwent allogeneic transplant in two sequential treatment eras. For children with high-risk AML, the 5-year OS rates in the recent cohort treated on contemporary protocols compared favorably to children treated in the previous era (5-year OS, 74% vs. 34% respectively, p < 0.0001). Importantly, this improvement was observed regardless of donor type and was attributed to decreased TRM, with a significantly lower hazard of death from infections (HR 0.12, p < 0.005) and regimen-related toxicity (HR 0.25, p < 0.002) while also being associated with a decreased risk of relapse (HR 0.4, p = 0.013).92 It is likely that with improvements in the transplant procedure, outcomes for patients with high-risk disease undergoing transplant will continue to improve.
TABLE 16.2 Hematopoietic Stem Cell Transplantation for Acute Myeloid Leukemia | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Related posts:
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
