Until now, the development of personalized medicine in oncology has relied on the use of somatic biomarkers to help therapists choose the right molecule(s) to administer, based on the genetic and molecular profile of each hematological disease. In this project, investigators propose to extend the strategy of therapeutic individualization to the field of dosage targeting. Today, azacytidine is a standard treatment for patients with acute myeloid leukemia (AML) and/or myelodysplastic syndromes (MDS), usually as monotherapy. According to the treatment regimen, azacytidine is prescribed at a standard dose (DS=75mg/m²/d), administered subcutaneously every day for 7 days. The treatment cycle is repeated every 28 days.

No study has evaluated the relevance of "a priori" dose adjustment on an individual basis, according to each patient's pharmacogenetic data. In current practice, doses are adapted a posteriori, and reduced empirically, following the occurrence of observed toxicity (6 to 71% of patients) (Schuck A et al. 2017). This ex-post adjustment in the face of grade 3-4 toxicity is a loss of chance for the patient. Similarly, under-dosing patients for fear of toxicity is another loss of chance. Investigator's hypothesis is that the optimal dose of azacytidine depends not only on the characteristics of the patient's pathology (risk groups including cytogenetic and molecular biology data), but also on the patient's individual characteristics (genetic status of metabolic enzymes and transporters). A mathematical model of the PK/PD type could, on the basis of early observations of circulating levels, be capable of rapidly predicting the pharmacodynamic repercussions in each patient, thus enabling rapid individualization of dosages. In the future, such a tool could make it possible to propose dosage adjustments rapidly after treatment initiation, before toxicity occurs, by predicting azacytidine exposure levels, themselves correlated with the patient's clinical condition.

Study design: In this open-label, paucicentric, non-randomized study, patients with AML and/or MDS, all of whom are receiving azacytidine-based chemotherapy as part of their standard treatment regimen, will be included. Each patient will be monitored for toxicities (EORTC), treatment response and progression-free survival. In addition to the standard care described above, each patient will undergo a series of constitutional genetic investigations conducted by NGS on markers linked to azacytidine pharmacokinetics (CDA, dCK). Another series of blood samples will be taken to calculate individual azacytidine pharmacokinetic parameters using a Bayesian approach.

Expected results: This study should make it possible to correlate pharmacogenetics with patient plasma exposure, and ultimately improve the molecule's efficacy/toxicity balance by personalizing dosage regimens, which until now have been carried out on an empirical basis.

Prospects: If the data are validated, a pre-therapeutic ADC assay could predict azacytidine pharmacodynamics and enable individual dose and/or dosage adjustment, as is the case with 5-FU and DPD.

Inclusion Criteria:

• Patients over 18 years of age
• Patients with acute myeloid leukemia
• Patient with myelodysplastic syndrome
• Person who has given non-opposition
• Patient who has signed an authorization to perform a constitutional genetic analysis (in the context of care)
• Need for effective contraception in patients of childbearing age

Exclusion Criteria:

• Failure to obtain non-opposition
• Adults under guardianship or safeguard of justice
• Persons deprived of their liberty
• Patient participating in another research project
• Pregnancy in progress