Pediatric aplastic anemia treatment patterns and responses; power in the numbers

In this edition of Haematologica, Rogers et al., representing 25 individual institutions, collectively report on their findings of the diagnostic approaches, applied therapies and responses in a cohort of 314 pediatric patients (aged 1-20 years) with a diagnosis of aplastic anemia (AA) collected through the North American Pediatric Aplastic Anemia Consortium (NAPAAC).1 This study highlights a number of important messages; specifically the power of collated registry data in a rare disease, the need to retest and continually refine diagnostic criteria to be fit for real-world purposes, the patterns of response and relapse to immunosuppressive therapy (IST) in a pediatric setting including the substantial differences in outcomes from IST in children compared with that in adults, and the importance of allogeneic stem cell transplant (HCT) in the therapy of refractory or relapsed disease.

Although AA can affect people at any stage of life, with a well reported bimodal peak of age incidence observed in older children/young adults and those over sixty years, it is a rare condition that often poses a diagnostic dilemma with acquired AA potentially confounded by a differential of inherited bone marrow failure syndromes in younger patients and hypoplastic myelodysplasia in older patients. These variations in clinical presentation and potential differences in pathophysiology between pediatric and adult patients with AA emphasize the importance of having data sets dedicated to pediatric cases on which to analysis patterns of presentation, diagnosis and treatment outcome and thereby recommend consensus-driven therapeutic algorithms, particularly in an environment in which there has been substantial historical variations in practice.2

The insights provided in the Rogers et al. analysis provides substantial clarity around several management issues in pediatric AA, but also pose several other questions either for ongoing analysis as their data set continues to expand and mature or as hypotheses to be tested in prospective studies.

A practical outcome of this consortium analysis is a real-world assessment of the applicability of the modified Camitta criteria for AA diagnosis first described in 1976.3 These diagnostic criteria are still recommended by international guidelines for assessment of AA severity.4 In these criteria, in addition to the depth of marrow hypocellularity and peripheral blood cytopenias, a reticulocyte count is required both as a diagnostic criterion for AA and also to assist with severity classification. However, in a prior analysis by the NAPAAC, it was established that reticulocyte values substantially vary between institutions, making their inclusion in diagnostic criteria uncertain.2 To address this conundrum, Rogers et al. offer the interesting observation from their data set of a lack of correlation between hemoglobin and reticulocyte count in their pediatric cohort at the time of diagnosis, and suggest that hemoglobin may be a more accurate and clinically relevant parameter on which to base management decisions. Whilst the Camitta criteria have stood the test of time, and their use is a strong recommendation, some of their elements are based on relatively low quality C level source data.4 The findings outlined in the Rogers et al. paper re-iterate the importance of ongoing review and modification of diagnostic criteria as new data sets, such as that collated by the NAPAAC, come to hand.

Similarly, collation and description by co-operative groups of current patterns of clinical practice and the degree of its adherence to consensus guideline is an important element of continued improvement in practice, particularly for rare conditions where individual institutional experience may be limited. Currently, one of the most widely accepted management decisions in the treatment of young patients with AA is to offer HCT in patients aged under 40 years with an HLA matched sibling donor (MSD).74 For those without a MSD, IST with anti-thymocyte globulin (ATG), most commonly horse-derived, in combination with cyclosporine is used as initial therapy with HCT from unrelated donors (UD) reserved for those who do not respond or who relapse after IST. Of the cohort outlined in the Rogers et al. analysis, the majority of HCT undertaken as second-line therapy utilized UD, indicating the lack of a MSD for upfront use, and in those who eventually received second-line HSCT from a MSD, it is unclear why this donor was not used in the upfront setting. This particular question may be answerable in future analyses by the NAPAAC.

The Rogers et al. analysis demonstrated a striking difference in outcome following IST in pediatric patients compared to that in a historical cohort of adult patients. While complete response (CR) was only seen in 10% of adults treated with IST,8 the pediatric cohort showed CR rates of nearly 60%. Despite this excellent response rate, a pattern of continual events, including death, relapse or transformation to hematologic malignancy following IST, resulted in a disappointing 5-year event-free survival (EFS) of 62%, similar to the findings showing by Yoshida et al.6 The finding of a continued pattern of events even after apparent successful therapy with IST further reinforces the view that normalization of peripheral blood parameters and marrow cellularity after immunosuppression does not imply normalization of hematopoietic clonality and/or immunological repertoire, and, as a consequence, the once-aplastic marrow remains at ongoing risk of recurrent aplasia, clonal evolution, and/or malignant transformation (Figure 1).

Figure 1.Immune mediated attack on hematopoietic stem cells (HSC) and potential responses to therapy. IST: immunosuppressive therapy; HCT: hematopoietic stem cell transplant.

This description of the patterns of response and subsequent relapse (or other event) in pediatric AA raises at least three important questions. Firstly, what is the most appropriate salvage therapy for those patients relapsing after initial IST? Secondly, given the high rate of relapse/events, should HCT from any matched donor be considered as front-line therapy in children? Thirdly, are there better biomarkers under development that might provide greater guidance in the choice between these treatment options? With regards to the first question, the Rogers et al. paper provides clear guidance. Re-treatment was required in 35%, and second-line therapy with an allogeneic HCT offered superior EFS to pursuing a second course of IST. These combined findings of unstable responses to IST, and the high rate of durable responses to HCT, contribute to the evolving debate as to whether HCT from any matched donor (related or unrelated) is preferable over IST as initial therapy for pediatric patients with proven AA. This question is clearly best answered in a randomized clinical trial, although such an undertaking would require a long-term commitment for feasible accrual and is only likely to succeed through consortia such as the NAPAAC. Lastly, biomarker development is critically required to more accurately determine the degree of clonal restriction (and therefore risk of clonal progression) and/or ongoing potential for immunological attack (and therefore post-IST relapse) to help determine whether IST or HCT should be offered as initial therapy. Clearly, in studies where restricted clonality is evident through genomic or cytogenetic analysis, poorer outcomes to IST are seen, indicating that with more sensitive techniques directed at assessment of the stem cell pool, more informed therapeutic decisions should follow.9 Again, it is likely that only through the co-ordinated efforts of consortia will sufficient biomarker samples be accumulated to begin to address this unmet need.

One clear determinant of clinical outcome from whatever therapy is chosen is the certainty with which the diagnosis is made. Increasingly, there is an appreciation that occult constitutional bone marrow failure syndromes may underlie what is thought to be a presentation of idiopathic AA, with significant implications for patient management. Through the increasingly readily available techniques for telomere length assessment10 and next generation sequencing for assessment of underlying germline lesions11 reclassification of many cases of AA is likely both during the prospective work up of new cases and retrospectively from archival diagnostic samples, which will further inform future treatment algorithms. As a greater clinical appreciation of the importance of diagnostic certainly has been met with greater diagnostic technical capacity, consensus recommendations increasingly incorporate evaluation of constitutional syndromes by chromosomal fragility testing in all AA patients presenting at younger than 50 years of age. Telomere length assessment is likely to be added to the routine work up panel in the near future.4 Reflecting these guidelines, Rogers et al. describe that while chromosomal fragility assessment was performed in most children, telomere length assessment was only performed at diagnosis in one-third of them.

Registries are crucial tools in efforts to improve outcomes for patients with rare diseases and their families. They serve as a means of pooling rare data in a standardized format in order to achieve meaningful sample sizes for subsequent analysis and allow comparison to historical or international cohorts, facilitate collaboration, generate hypotheses for future testing, and provide a framework for annotated sample collection and translational research. Further, participation in registry reporting contributes to achieving consistent and complete work up of new cases and provides a means of formulation and distribution of educational opportunities including multidisciplinary discussions which are so often needed in the management of rare conditions. Registries allow for the identification of patients, informing epidemiology assessments and areas of need, and may assist with allocation of scarce resources. Registries may facilitate feasibility assessments of and planning for clinical trials. The importance of registries focused on AA in particular is reflected in the increasing number of publications describing national outcome data in AA.14122 In this edition of Haematologica, Rogers et al. have made an important contribution to this data pool, informing optimal diagnostic and therapeutic approaches and, equally importantly, highlighting opportunities for further research and discussion in pediatric AA.

Footnotes

• FundingLF is supported by a Higher Degree Fellowship in Bone Marrow Failure from Maddie Riewoldt’s Vision.