Long-term outcomes of patients with refractory cytopenia of childhood under observation only Open Access

Patients with unexplained cytopenia represent a diagnostic challenge in pediatric and adult hematology. A critical step in obtaining the correct diagnosis is to distinguish reactive causes of cytopenia, such as infections, autoimmune disorders, liver disease, and vitamin deficiencies, from clonal hematological disorders. Recently, categories have been introduced to sort out precursor states from myeloid malignancies for better clinical care. Foremost, the term idiopathic cytopenias of undetermined significance was formed to describe adult patients with relevant cytopenias in ≥1 lineages persistent for at least 6 months, which are not explained by any other disease, and the criteria for myeloid neoplasm are not fulfilled.^1,2 The clinical course of idiopathic cytopenias of undetermined significance is variable and unpredictable, but, in a subset, progression to myelodysplastic syndrome (MDS) or acute myeloid leukemia is observed. In pediatrics, refractory cytopenia of childhood (RCC) describes an entity defined by persistent cytopenia, dysplastic changes, and a characteristic histopathological pattern with patchy distribution of hematopoiesis, predominance of left-shifted erythropoiesis with increased mitosis, and markedly decreased granulopoiesis and megakaryopoiesis.^3,4 Approximately 80% of RCC cases have a hypocellular marrow. It is important to appreciate that not all cases of RCC are bona fide MDS carrying acquired somatic mutations or cytogenetic abnormalities.^5,6 The RCC pattern can also be observed in the absence of markers of clonality in some cases of the classical inherited bone marrow failure disorders such as Fanconi anemia, telomere biology disorders, Shwachman-Diamond syndrome, or in the autosomal dominant inherited predisposition syndromes such as GATA2 deficiency, or SAMD9/SAMD9L syndrome.^7-10 Most patients with RCC have a normal karyotype, but monosomy 7 is the most frequent chromosomal abnormality.^11-13 T-cell receptor Vβ skewing and/or small paroxysmal nocturnal hematuria (PNH) clones are observed in 40% and 41% of patients with RCC, respectively, suggesting that an immune-mediated pathomechanism is involved in a subset of patients.^14,15 

Since 1996, the European Working Group on MDS (EWOG-MDS) has prospectively registered patients with RCC, providing a unique opportunity to study the natural history of this rare disorder. In a risk-adapted therapy approach, patients with transfusion dependency, clinically relevant neutropenia (absolute neutrophil count [ANC] of <1 × 10⁹/L), monosomy 7, del(7q), or ≥2 chromosomal aberrations had an indication for early hematopoietic stem cell transplantation (HSCT) or, in selected cases, immunosuppressive therapy (IST), whereas an observational approach was chosen for the remaining patients (supplemental Figure 1).¹⁶ Here, we provide the clinical characteristics and long-term outcome of patients with RCC with normal karyotype and absence of a predisposition syndrome who were followed up with an observational approach for at least 2 years from diagnosis.

Patients with RCC (n = 688) were consecutively enrolled in the EWOG-MDS registries (EWOG-MDS 98 ClinicalTrials.gov identifier: NCT00047268; EWOG-MDS 2006 ClinicalTrials.gov identifier: NCT00662090) in Germany between 1 July 1998 and 30 June 2021. All patients with monosomy 7, del(7q), or ≥2 chromosomal aberrations were excluded from this study, because they had an indication for HSCT. After excluding patients who underwent HSCT or IST, died without therapy, or experienced disease progression within 2 years of diagnosis, 132 patients were identified as having been managed with an observational approach (Figure 1). Patients were generally followed up at a pediatric clinic until last follow-up.

During the study period of 23 years, major advances in molecular diagnostics let to the description of novel predisposition syndromes such as GATA2 deficiency and SAMD9/SAMD9L syndrome.^9,10,17 Biological samples of patients registered with EWOG-MDS before these discoveries had been analyzed retrospectively for presence of these germ line mutations on a research or clinical care bases.^9,10 Blood or bone marrow samples from 127 of 132 patients managed with an observational approach were analyzed for germ line disease (Figure 1, no sample in 5 patients). Predisposition syndrome was diagnosed in 29 patients. Five patients had an abnormal karyotype with a single aberration (3 of whom had a predisposition), and karyotype information was unavailable for 1 patient. After excluding patients with predisposition syndrome and/or an abnormal or unknown karyotype, 100 patients were included in this study. Analysis for presence of PNH clones was performed as previously reported.¹⁵ 

The diagnosis of RCC was established by central review of bone marrow core biopsies and aspirates by EWOG-MDS reference pathologists according to the fourth edition of the World Health Organization 2008 and its 2017 revision,^18,19 more recently according to the 2022 International Consensus Classification of Myeloid Neoplasms and Acute Leukemias.⁴ Hematoxylin and eosin staining, reticulin staining, and naphthol AS-D chloroacetate esterase staining were performed as standard diagnostics. Immunostaining was performed for CD42b, E-cadherin, myeloperoxidase, terminal deoxynucleotidyl transferase, CD20, CD3, CD123, and CD34, as previously reported.²⁰ Results of bone marrow cellularity were analyzed over time in patients with at least 3 bone marrow examinations, with the last examination conducted at least 2 years after diagnosis (n = 57). Specimens collected after HSCT were excluded. We also evaluated the long-term course of blood counts in patients with available data at least 2 years after diagnosis, excluding those who underwent HSCT (n = 74).

In descriptive statistical analyses of demographic and clinical data, location and scale parameters of continuous variables were calculated (median, minimum, and maximum). Absolute and relative frequencies tables of categorical variables were prepared. The χ² test was used to examine the statistical difference of categorized factors. The Mann-Whitney U test was used for comparison of differences in medians of continuous variables. Overall survival was defined as the time from diagnosis to death or last follow-up and transplant-free survival as the time from diagnosis to death, HSCT, or last follow-up. The Kaplan-Meier method was used to estimate survival rates,²¹ and all results are expressed as the estimated probability of survival with the 95% confidence interval. All P values were 2-sided and values of <0.05 were considered statistically significant. Statistical analysis was performed using SPSS for Windows 30 (IBM).

The University of Freiburg institutional ethics committee had approved the research (247/05 and 430/16) and written informed consent had been obtained from patients’ guardians. The study was conducted in accordance with the Declaration of Helsinki.

The study cohort comprised 100 patients with RCC with a median age at diagnosis of 10.9 years (range, 1.4-17.3) and a slight male predominance (male/female = 1.2). Patients presented with varying degree of thrombocytopenia, neutropenia, and/or anemia (Table 1). The age-adjusted mean corpuscular volume (MCV) of red cells was increased in approximately half of patients.²² Seven (7%) and 14 patients (14%) received at least 1 red blood cell or platelet transfusion at the time of diagnosis, respectively. The PNH clone analysis was performed in 15 patients, and a minor PNH clone was detected in 7 patients. The diagnostic bone marrow was hypocellular in 84% (n = 84) of cases, consistent with the typical presentation of RCC. There was no difference in patient characteristics between patients with hypocellular marrow and those with a normocellular or hypercellular marrow (data not shown). In the remaining cases, the marrow was normocellular or hypercellular, including 2 patients with a minor PNH clone.

The median follow-up time was 7.2 years (range, 2.0-18.4) from diagnosis, yielding 788 person-years of longitudinal observations. In median, patients were followed up until the age of 17.9 years (range, 6.1-31.2). Most patients had reached middle adolescence (14-17 years of age, n = 36, 36%), late adolescence (18-21 years of age, n = 29, 29%), or adulthood (>21 years of age, n = 19, 19%) at last follow-up.

All patients were alive at last follow-up (Figure 2). Clonal evolution defined by an abnormal karyotype and/or increase in blasts occurred in 3 patients (3%) (Table 2), 1 of whom (D682) developed progressive cytopenia 9 years from diagnosis with an increased bone marrow blast percentage (5%) and partial trisomy 21q22; MDS with excess blasts was diagnosed. The patient subsequently underwent successful HSCT. The 2 other patients had transient cytogenetic abnormalities without morphological progression. Three patients developed clinical PNH at 1, 2, and 10 years from diagnosis, respectively; 2 of whom had a detectable PNH clone at diagnosis, whereas PNH testing had not been performed in the remainder. Eculizumab therapy was started in 1 patient, and 2 patients received transplantation (Table 2). Except for 1 patient who developed clinical PNH with hemolysis at the age of 8 years, clonal evolution and clinical PNH presented in late adolescence (aged ≥16 years). All patients with clonal evolution or clinical PNH had a hypocellular marrow at diagnosis.

Nine (9%) patients received HSCT for MDS with excess blasts (n = 1), clinical PNH (n = 2), transfusion dependency (n = 4), or progressive neutropenia (n = 2). Of 4 patients who received transplantation for transfusion dependency, 2 had developed transfusion requirements during the clinical course, the other 2 had been transfusion dependent since diagnosis but did not receive transplantation because of lack of donor availability or initial parental refusal. HSCT was performed 2 to 11 years from diagnosis at a median age of 15 years (range, 7-21). Patients with subsequent HSCT had a lower median hemoglobin (Hb; 8.7 vs 10.6 g/dL) and lower platelet count (30 × 10⁹/L vs 64 × 10⁹/L) at diagnosis compared with patients without HSCT during the clinical course. The 5-year and 10-year transplant-free survival were 94% and 88%, respectively (Figure 2). In total, 91 of 100 (91%) patients were continuously observed by a watch-and-wait strategy until last follow-up.

Hematological course was evaluated in 74 patients without HSCT and sufficient data on long-term blood counts (median, 5.2 years; range, 2.0-15.6). Most patients (n = 59) had continuous cytopenia and/or an elevated MCV of red blood cells during the course: 34 patients had significant abnormalities in blood counts (Hb <10 g/dL, ANC <1 × 10⁹/L, and/or platelets <100 × 10⁹/L; last examination at median 4.2 years, maximum 12.2 years after diagnosis), whereas 26 patients had mild abnormalities (Hb, 10-12 g/dL; ANC, 1.0-1.5 × 10⁹/L; platelets, 100-150 × 10⁹/L; and/or persistent high MCV; median, 5.4 years [maximum 14.1 years] after diagnosis). Fifteen patients had normal blood counts and normal MCV at last follow-up (median, 6.5 years [maximum 15.6 years] after diagnosis). In 10 of 15 patients with normalized blood counts, marrow histology showed persistent RCC features, whereas the histological criteria for RCC were no longer met in the remaining 5 patients. Of 12 patients with a mild cytopenia at diagnosis, 4 had a significant cytopenia at the last follow-up. Interestingly, 7 of 15 patients with normalized blood counts had nonhematological abnormalities (supplemental Table 1). There was no significant difference in patient characteristics between patients who normalized their blood counts and those who did not (data not shown).

Bone marrow cellularity changes were assessed in 57 patients over the course of the disease, with the last evaluation conducted at a median of 5.2 years (range, 2.0-13.3) after diagnosis. All 49 patients with hypocellular bone marrow at diagnosis maintained hypocellularity throughout the disease course. Notably, both patients with a transient abnormal karyotype had hypocellular marrow at diagnosis and during follow-up (Table 2). Among the 7 patients with a normocellular marrow at diagnosis, 4 experienced decreased cellularity (1 with a declining neutrophil count, and 3 with stable or improving blood counts), 2 showed no change, and 1 had increased cellularity despite progressive cytopenia. Additionally, 1 patient with initially increased cellularity later showed decreased cellularity without changes in blood counts.

This analysis demonstrates an excellent outcome of 100 consecutive pediatric patients diagnosed with RCC, absence of a known germ line predisposition, normal karyotype, and absence of severe cytopenia who were followed up without therapy such as HSCT or IST within the first 2 years from diagnosis. All patients, including 9 patients who received transplantation during the course, were alive up to 20 years from initial presentation, confirming that the stratification into an observation-only group was safe. Similar results of an expectant strategy have been reported for 27 Japanese patients with RCC followed up from diagnosis without therapy for at least for 2 years; overall survival was 87% ± 9% at 5 years, with 9 patients having received HSCT or IST during the course.²³ 

Regardless of the excellent long-term survival, most patients with an initial histological pattern of RCC continued to have persistent cytopenias and a hypocellular bone marrow as assessed by serial examinations. With the low incidence of evolution of abnormal karyotype and development of clinical PNH, yearly bone marrow examinations previously recommended by EWOG-MDS and performed in many pediatric centers can be omitted. Considering the psychological burden and risk of sedation of repeated marrow investigations in children, it is appropriate for these patients to be followed up by medical history and peripheral blood analysis only. If blood counts obtained every 3 to 6 months alter from the patient’s baseline, further investigations including bone marrow studies may be indicated.

It is of special interest that 15 patients had normal blood counts at last follow-up. The low percentage supports the notion that the RCC pattern is a phenotype of perturbed hematopoiesis and reflects an intrinsic bone marrow disorder. This interpretation is possibly supported by the observation that 7 of 15 patients had nonhematological phenotypic abnormalities such as abnormal skin pigmentation, sensorineural hearing loss, congenital heart disease, clotting disorder, dysmorphic facial features, or a consanguineous family history. It is conceivable that in some of these patients, somatic genetic rescue of an unknown germ line condition improved hematopoiesis and resulted in a normal blood picture.¹⁰ 

This study excluded patients with a known predisposition syndrome. Identifying children with germ line condition at time of diagnosis of RCC is crucial as indicated by early death of 6 of 29 patients with predisposition (Figure 1) because of liver failure, lung fibrosis or pneumonia (4 patients with telomere biology disease), infection (ligase-4 deficiency), or hemophagocytic lymphohistiocytosis (GATA2 deficiency).

Our data emphasize the importance of transition of adolescent patients with RCC to adult hematology care (supplemental Figure 2). Patients and caregivers must understand the significance of further follow-up and surveillance in adult life. Although the risk of MDS and clinical PNH early in life is low, there are currently no data on the incidence of hematological neoplasia in middle and older adulthood. Furthermore, it is conceivable that some patients have a yet unrecognized germ line predisposition. Continuous care by an experienced hematologist for adults may give patients and families the best chance to learn about future discoveries and benefit from clinical research.

The authors thank A.-R. Kaya, M. Teller, and C. Jaeger for excellent laboratory assistance; and A. Breier and W. Truckenmueller for data management (all Medical Center, University of Freiburg). The authors acknowledge the Hilda Biobank at the Department of Pediatrics and Adolescent Medicine, Freiburg, Germany, for specimen processing. Patient care within the European Working Group of Myelodysplastic Syndrome in Childhood (www.ewog-mds-saa.org) would not have been possible without the continuous effort of the National Reference Pathologists, National Reference Cytogeneticists, physicians, nurses, and other staff of pediatric oncology units and transplant centers.

The research presented herein is supported by grants from the German Federal Ministry of Education and Research 01GM1911A “MyPred - Network for young individuals with syndromes predisposing to myeloid malignancies” (G.G., Y.L.B., M.W.W., B.S., C.M.N., M.E., and A.Y.), and 01GM1909B and 01GM2205C “ADDRess” (R.K.); as well as grants from the German Cancer Consortium (M.W.W. and C.M.N.) and the Förderverein für krebskranke Kinder Freiburg e.V., Freiburg, Germany.

The authors acknowledge support by the open access publication fund of the University of Freiburg.

Contribution: B.D., C.M.N., and A.Y. designed the research; B.D. and A.Y. interpreted clinical data and wrote the manuscript; S.S.-F., I.B., and M.R. were involved in the histological diagnosis; D.L., S.R., A.K., S.H., M.W.W., F.B., and R.K. were responsible for molecular diagnosis for predisposition syndromes; Y.L.B. and G.G. were responsible for cytogenetic diagnosis; N.R., F.B., B.S., M.E., C.M.N., and A.Y. were responsible for clinical data collection; P.N. performed the statistical analysis; and all authors contributed to the manuscript and approved its final version.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

The current affiliation for G.G. is Amedes Genetics, Hannover, Germany.

Correspondence: Ayami Yoshimi, Department of Pediatric Hematology, Oncology and Stem Cell Transplantation, Children’s Hospital, Medical Center–University of Freiburg, Faculty of Medicine, University of Freiburg, Breisacherstrasse 62, 79106 Freiburg, Germany; email: ayami.yoshimi@uniklinik-freiburg.de.