Hematopoietic stem cell transplantation (HSCT) has improved survival of children with JMML considerably. Since it became apparent that JMML is a heterogeneous disease it has been questioned whether HSCT is the therapy of choice for all subtypes of JMML. Here we report the outcome of children with JMML registered to the retrospective or prospective studies of the European Working Group of MDS and JMML in Childhood (EWOG-MDS) with respect to clinical and mutational subtypes and presence or absence of HSCT. Among the 488 children with JMML studied by EWOG-MDS 43 patients (8.8%) were known to have Noonan syndrome, while 48 (9.7%) carried the clinical diagnosis of neurofibromatosis type 1 (NF1). Mutational analysis was carried out in 254 of the remaining 397 children. Somatic mutations in PTPN11, KRAS and NRAS were seen in 116, 46 and 47 children, respectively; 8 patients were noted to have germline CBL mutations in the absence of Noonan phenotype and 37 children with complete typing (N=203) had none of these abnormalities (all negative group). For the 51 patients with Noonan syndrome and/or germline CBL mutations overall survival at 5 years was 0.72 (0.58-0.86) irrespective whether HSCT was performed or not. A normal karyotype was observed in all but 3 of these 51 children. Normal chromosomal studies were also observed in 74%, 66% and 60% of the NF1, PTPN11 mutated and the all negative group, respectively. Interestingly, monosomy 7, and other aberrations were noted in 52% and 10% of KRAS mutated patients, respectively, but only in 7% and 5% of the NRAS mutated group (P<0.05). Five of the patients with NRAS mutation and normal karyotype are long term survivors without HSCT (5.5 to 27 years after diagnosis) compared to none of the children from the NF1, PTPN11 mutated, KRAS mutated or all negative group. Event-free survival (EFS) following HSCT differed significantly among the mutational groups with 0.38 (0.29-0.49), 0.43 (0.25-0.61), 0.45 (0.24-0.66), 0.69 (0.54-0.84) and 0.72 (0.56-0.88) for children with PTPN11 mutation, NF1, NRAS mutation, KRAS mutation and the all negative group (P<0.01). Likewise, relapse rate varied between 0.41 (PTPN11 mutation), 0.39 (NF1), 0.31 (NRAS mutation), 0.13 (all negative group) and 0.10 (KRAS mutation) (P< 0.01). Multivariate analysis identified mutational type and age at diagnosis as independent prognostic factors for EFS following HSCT. Interestingly, of the 7 children with JMML and NRAS mutation who relapsed post HSCT 5 had a normal karyotype. Considering that other children with JMML sharing the same oncogenic exon 1 mutations in NRAS experienced long-term survival without HSCT these data indicate that a NRAS mutation by itself may not be sufficient to cause lethal disease. Our results also demonstrate that clinical characteristics and outcome following HSCT differ between patients with JMML and oncogenic NRAS and KRAS mutations. Most importantly, we showed that in JMML subtypes with a low relapse incidence following HSCT (KRAS mutated and all negative group) transplant related mortality exceeds the relapse incidence, while in other subtypes (PTPN11 mutations, NF1) leukemic relapse is the most common failure. Current transplant strategies for children with JMML need to be revised to accommodate these differences among mutational subgroups.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.