Genetic Aberrations As Putative MRD Targets in Childhood Acute Myeloid Leukaemia

Background: Pediatric AML is considered as a heterogeneous disease affecting 15% of children with acute leukaemia. Outcome is still unsatisfactory with an overall survival around 70% and a relapse rate of 30-40% after primary treatment. The most important factors determining clinical outcome are genetic aberrations, and early response to treatment. Many studies have demonstrated the prognostic significance of early marrow response by bone marrow morphology, and more recently by flow cytometry for MRD quantification. All studies have demonstrated that rapid early clearance is associated with a better outcome. MRD-PCR is more sensitive than flow MRD but the applicability is lower since recurrent mutational targets, using conventional technology, are present in only 40-50% of patients. Furthermore, AML usually presents with multiple distinct clones at diagnosis varying in number and volume, varying in number between diagnosis and relapse and showing mutational losses or gains at time of relapse. Therefore, only few studies have investigated the significance of response by MRD-PCR. In this study we investigated the applicability and stability of various molecular aberrations as putative MRD targets.

Methods: Dutch patients (n=86) included in the pediatric Dutch Belgium-AML-01 cohort were screened for various aberrations (CBFβ/MYH11; AML1/ETO; MLL/AF4; MLL/AF9; MLL/ENL; FLT3/ITDs; NPM-1exon 12). Patients positive for any of these mutations were followed during treatment using aberration specific quantitative MRD-PCR.

Results: For 53% of the included patients (46/86) molecular aberrations could be identified. For the AML1/ETO gene fusion, 6/8 patients were followed for response to treatment by MRD-PCR until 17-51 weeks after diagnosis. None of these patients developed a hematological relapse in this period while low but stable MRD levels were still detectable in half of the tested patients (1.5E-05 to 8.9E-04). In case of CBFβ/MYH11 gene fusion, for 6/8 patients (4 CBFβ/MYH11 type A patients and 2 CBFβ/MYH11 type D patients) MRD-PCR was performed until 5-130 weeks after diagnosis. Three patients developed a hematological relapse (all type A) which was preceded by clear molecular relapse, in one case even 19 weeks preceding the clinical relapse. All tested patients with MLL fusions (1/1 MLL/AF4; 2/6 MLL/AF9; 0/0 MLL/ENL) showed no relapse until 54 weeks after diagnosis while MRD levels were all negative. For NPM-1, 7/9 patients were tested by MRD-PCR. Four patients developed a relapse also shown by MRD-PCR, indicating the stability of this marker. FLT3/ITD MRD-PCR was tested for 14/14 positive patients. This marker was not as stable as the other aberrations tested: we observed 1 patient with no FLT3/ITD at diagnosis and until week 8, while at hematological relapse (week 41) this marker was gained. In contrast, 2 patients showed presence of FLT3/ITD at diagnosis, with loss at relapse. Another patient showed 3 separate FLT3/ITD clones with different kinetics before and after relapse. Five other relapses (with positive FLT3/ITD at diagnosis) were also shown by MRD-PCR.

Conclusions: In this study we show that, although numbers are low, selected genetic aberrations might be suitable as MRD targets in childhood AML. Thus, it is clear that response evaluation by MRD-PCR is feasible and carries significant prognostic information for a specific group of aberrations such as AML1/ETO, CBFβ/MYH11 and NPM-1. MRD-PCR in combination with MRD-flow will increase the number of patients suitable for MRD and subsequent earlier detection of imminent relapse.