Central Morphology Review of Childhood Bone Marrow Failure in Japan

Abstract

Abstract In the diagnosis of childhood bone marrow failure (BMF), differentiating aplastic anemia (AA) from hypoplastic myelodysplastic syndrome (MDS) is challenging. In addition, inherited BMF (IBMF) should be excluded from acquired BMF. The 2008 WHO classification has proposed a provisional entitiy, “refractory cytopenia of childhood (RCC)”. The spectrum of patients with RCC is wide, ranging from patients with severe hypocellular bone marrow (BM) and mild dysplasia to those with normocellular BM and distinct dysplasia meeting the criteria for refractory cytopenia with multilineage dysplasia (RCMD) in adults. Currently, it is recommended that children who meet the criteria for RCMD should be considered as RCC in the WHO classification until the number of lineages involved have been fully evaluated with regard to their relative importance as prognostic factors in childhood MDS. To enable diagnosis based on the WHO classification, the Japanese Society of Pediatric Hematology and Oncology in February 2009 established a central review system of BM morphology, including peripheral blood (PB) and BM smears and specimens from trephine biopsies in childhood BMF. PB and BM smears were reviewed by two pediatric hematologists, and the specimens from BM trephine biopsies were reviewed by a pathologist. In addition, the telomere length of lymphocytes and paroxysmal nocturnal hemoglobinuria (PNH) clones in PB were measured by flowcytometry for patients with BMF. RCC is defined as persistent cytopenia with <2% and <5% blasts in PB and BM, respectively. BM aspirate smears show dysplastic changes in >2 cell lineages or >10% within one cell lineage. On the other hand, the criteria of RCMD in adult MDS is defined as persistent cytopenia with <1% and <5% blasts in PB and BM, respectively. BM smears show >10% dysplastic changes in >2 cell lineages. We introduced the RCMD criteria in this central review. From February 2009 to October 2013, 1,000 cases including 536 males and 464 females were prospectively reviewed. The median age was six years (range, 0–39 years). Of the 1,000 cases, 575 were classified as BMF, and of them, 137 were classified as AA, 236 as RCC, 103 as RCMD, 38 as hepatitis-related BMF, 3 as PNH and 58 as IBMF. Of the 58 cases with IBMF, 21 were diagnosed as Fanconi anemia, 12 as Shwachman–Diamond syndrome, and 8 as dyskeratosis congenita. Seventeen patients suspected of IBMF were undiagnosed. In 97 advanced cases of MDS, 24 were classified as refractory anemia with excess blasts (RAEB), 6 as secondary MDS, and 21 as therapy-related MDS. To determine the clinical differences among AA, RCC, and RCMD, we compared laboratory and clinical data for 476 patients classified as AA, RCC, and RCMD. Median ages in the AA, RCC, and RCMD groups were 9, 8, and 7 years, respectively (p = 0.007). The male/female ratio in AA, RCC, and RCMD groups was 1.1, 1.2, and 3.6, respectively (p = 0.034). When patients were classified according to the disease severity criteria for AA, 78% of the patients with AA had very severe or severe disease, whereas only 38% of the patients with RCC and 28% of the patients with RCMD had very severe or severe disease (p < 0.001). Chromosomal abnormalities were detected in two patients (1%) with AA (trisomy 8), 10 patients (4%) with RCC (monosomy 7, n = 2; trisomy 8, n = 6; other, n = 2), and 12 patients (12%) with RCMD (monosomy 7, n = 5; trisomy 8, n = 2; other, n = 5) (p = 0.001). Out of the 476 patients, 67 (AA, n = 32; RCC, n = 32; RCMD, n = 3) were administered IST with rabbit antithymocyte globulin (ATG) and cyclosporine. After 6 months, the response rate to IST was not significantly different among the three groups; AA, 41%; RCC, 47%; RCMD, 100% (p = 0.142). In conclusion, the entity of RCMD should be applied to childhood MDS because patients with RCMD exhibited a significantly high frequency of chromosomal abnormalities at the time of diagnosis. To definitively determine whether these three diseases are different entities, it would be necessary to prospectively compare the clinical outcomes and biological findings in a larger number of patients with AA, RCC, and RCMD. Disclosures No relevant conflicts of interest to declare.