Chronic myelogenous leukemia (CML) in the pediatric population is rare. It accounts for about 3% of childhood leukemias, which translates into fewer than 100 cases per year in this country. Therefore, reports detailing the results of hematopoietic cell transplantation (HCT) for pediatric patients with CML are few in number. Pediatric outcome data are often dissected from published adult series. While the only known cure for CML is an HCT, the success of imatinib in attaining molecular remission (reviewed in Seminars in Hematology, April 2003) has delayed the timing of HCT for most adults. In fact, many transplant centers offer HCT only for patients with progressive disease. However, for the pediatric patient with a matched sibling, transplantation early in disease is the treatment of choice favored by most, but not all (Thornley et al, Med Ped Oncol. 2003;41:115-117). Is an HCT a rational strategy for children who lack a sibling donor? In this issue of Blood, Cwynarski and colleagues (page 1224) retrospectively review data from the European Bone Marrow Transplant Registry (EBMTR). They report on 314 pediatric CML patients who underwent transplantation with either a matched sibling or a matched volunteer unrelated donor (VUD). These data have some limitations related to retrospective registry studies such as different regimens for conditioning and graft-versus-host disease prophylaxis among institutions, missing data points that excluded patients from the analysis (216 patients), and different HLA-typing methodologies, to name a few. The data clearly point out that HCT for CML in children can be successful with a matched sibling or a VUD.
A number of issues are raised. Is an irradiation-containing preparative regimen necessary? Is there a role for imatinib after transplantation? Are nonmyeloablative transplants associated with lower transplantation-related mortality and improved survival? For a VUD transplant, what is considered “well-matched”? While the number of HLA class I (∼50) and II (∼30) antigens has remained relatively constant since 1988, the number of alleles has grown exponentially. Currently there are 275 HLA-A, 521 HLA-B, and 405 HLA-DRB alleles (Marsh GE. www.anthonynolan.com/HIG/index.html). Additionally, do we consider only HLA-A, HLA-B, and HLA-DR? Molecular typing now allows the evaluation of HLA-C, and there are data to support its inclusion in donor selection (Petesdorf et al, Blood. 1997;89:1818-1823). What about HLA-DQ and HLA-DP? Allele-level matching has altered the landscape of transplantation. Certainly not all allelic differences between donor and recipient are biologically relevant, but how do we determine which ones are? If small, single-institution trials for pediatric CML attempt to resolve these issues, we will be left with contradictory results and without a logical biologic base to develop future trials.
It is therefore imperative that we implement a system to capture outcome data and long-term follow-up for all pediatric patients with CML regardless of treatment. Because there are so few patients, this will necessitate an international collaborative effort. The challenge of constructing a treatment strategy that is acceptable to physicians is more daunting, but if it is a well-designed, evidenced-driven protocol, “build it and they will come.”
