The assessment of minimal residual disease versus that of somatic mutations for predicting the outcome of acute myeloid leukemia patients

Abstract Background In addition to morphological and cytogenetic features, acute myeloid leukemias are characterized by mutations that can be used for target-therapy; also the minimal/measurable residual disease (MRD) could be an important prognostic factor. The purpose of this retrospective study was to investigate if somatic mutations could represent an additional prognostic value in respect of MRD alone. Method At baseline, 98 patients were tested for NPM1, FLT3, and for WT1 expression; 31 for ASXL1, TET2, IDH1, IDH2, N-RAS, WT1, c-KIT, RUNX1, and DNMT3A. The same genes have been also tested after induction and consolidation. Results Overall, 60.2% of our patients resulted mutated: 24.5% carried mutations of FLT3-ITD, 38.7% of NPM1, 48.4% of c-KIT, 25.8% of N-RAS and 19.3% of IDH2. The probability of achieving a complete response (CR) was higher for younger patients, with low ELN risk score, NPM1-mutated, with low WT1 levels, and without FLT3. The presence of additional mutations represented a poor predictive factor: only 19% of these cases achieved CR in comparison to 43% of subjects without any of it. Concerning survival, it was conditioned by a lower ELN risk score, younger age, reduction > 1 log of the NPM1 mutational burden, disappearance of FLT3 mutations and lower WT1 expression. Regarding the role of the additional mutations, they impaired the outcome of 20% of the already MRD-negative patients. Concerning the possibility of predicting relapse, we observed an increase of the NPM1 mutational burden at the time-point immediately preceding the relapse (about 2 months earlier) in 50% of subjects. Similarly concerning WT1, an increase of its expression anticipated disease recurrence in 64% of cases. Conclusions We demonstrated that additional somatic mutations are able to impair outcome of the already MRD-negative subjects. About MRD, we suggest a prognostic role also for the WT1 expression. Finally, we considered as relevant the assessment of NPM1 quantity clearance instead of the presence/absence of mutations alone. Still remains in doubt the utility in terms of long-term prognosis of a baseline more complex mutational screening; we could hypothesize that it would be useful for those patients where other markers are not available or who reached the MRD negativity