Introduction Acute Myeloid Leukemia (AML) identifies a clinically heterogeneous group of clonal disorders. Advances in Next Generation Sequencing (NGS) have increased the possibility to define the genetic lesions present in each patient, which has led to improved disease classification, risk stratification and identification of new therapeutic targets. The 2017 European LeukemiaNet (ELN) reccommendations for genetic risk stratification of AML have been widely adopted among physicians and investigators, even though patient responses remain heterogeneous. A new edition of reccomendations for diagnosis and management of AML in adults, ELN 2022, has been recently proposed to better define prognostic significance of genomic abnormalities. Aim of the study was to evaluate the usefulness of the new ELN 2022 genetic risk classification in a series of patients diagnosed with AML at two hematological institutions. Methods A retrospective analysis of 158 newly diagnosed AML patients according to WHO 2016 (100 de novo, 58 secondary) and treated from 2013 to 2021 in two hematological centers of Sardinia, Italy, was performed. Male patients were 96 (61%) and 63 out of 158 (40%) were >70 yrs old. One-hundred-five AML diagnosis samples were analyzed by targeted NGS on Ion Torrent platform using the DNA Oncomine Myeloid Research panel which explores 1256 hotspots in 33 key myeloid genes. Demographic and clinical data, disease characteristics at diagnosis, first-line treatment and clinical outcome data were available for all patients. Cytogenetic and molecular characteristics were used to classify patients into ELN 2017 and ELN 2022 risk groups. Results NGS detected 196 clinically revelant mutations in 87% AML patients with a limit of detection of 2%, while fusion genes were detected in 5% of patients (5/105). The most frequent genetic abnormalities were FLT3ITD/TKD and NPM1 mutations which represented the 29% and 28%, respectively. Co-occurrence of NPM1 and FLT3 driver gene mutations was found with a frequency of 16%. IDH1/2 mutations were observed in 20/105 pts (19%), RUNX1 and K/NRAS in 12/105 (11%), ASXL1 mutated in 11 pts (10%), STAG2 and BCOR mutations were present in 4 and 1 pts, respectively. TP53 mutations were observed with a frequency of 13% (14 patients). ELN 2017 classified 24 cases as favorable (23%), 40 as intermediate (38%), and 41 as adverse (39%) risk group. Sixteen out of 105 (15%) patients migrated to a more adverse risk group when ELN 2022 was applied: from favorable to intermediate (7) or adverse (2), 7 pts from intermediate to adverse. ELN 2022 re-classification assigned at favorable 15 pts (14%), 40 at intermediate ( 38%), and 50 at adverse risk groups (48%). In our cohort with the use of the new ELN 2022 classification no patient was reclassified with a better risk profile. Conclusions The recent ELN 2022 classification of AML with updated genetic integration has been proposed as a useful disease classification which impacts on the choice of treatment and outcome of individual patients. Our retrospective analysis confirms that the ELN 2022 is a robust tool of genetic risk classification and redefines a better genomic profile of AML patients prompting to individualized treatments. Re-classification according to ELN 2022 showed shifting from favorable to a worse category in the 10% of AML patients <60 years of age, who could have been eligible to a more intensive treatment. Furthermore, the implementation of the AML screening gene panel through NGS analysis allows to identify molecular abnormalities, such as IDH1/IDH2 mutations, susceptibles of target therapies.