Acute Myeloid Leukemia (AML) is, in fact, a set of diseases with variable clinical and immunophenotypic presentation, determined by genetic alterations. During treatment, new leukemic clones may arise, or even the dominant clone may undergo genetic and immunophenotypic modifications in the process of clonal evolution. This process is responsible for leukemia relapse and failure to detect measurable residual disease (MRD). Pre-leukemic stem cells detected or not at the time of diagnosis may also undergo clonal evolution and determine the relapse of leukemia. This case report exemplifies these situations. A 61-year-old woman was referred to our hospital after 2 months of treatment for a monocytic AML (30% monoblasts, 18% promomocytes) in a fragile clinical status. She had been treated with 2 cycles of Cytarabine and Daunorubicin (5+2). The immunophenotypic diagnosis was made at our institution: the monoblasts expressed CD4, CD36, CD56, CD71, CD97, CD99, dim/partial myeloperoxidase, partial CD7, heterogeneous CD13, bright and homogeneous CD15, CD18, CD33, CD38, CD44, CD45, CD64, LLC1 and HLA-DR. Monoblasts did not express CD2, cyCD3, SmCD3, CD10, CD11b, CD14, CD16, CD19, CD22, CD34, CD35, CD45RA, CD54, cyCD79a, CD96, CD105, CD117, CD123, CD244, CD300e, NuTdT, TIM3, NG2. She did not have access to molecular tests. After the 1st and 2nd cycle of chemotherapy, the MRD results were 0.12% and 0.06% respectively, but the blast cell immunophenotype was different from the diagnosis, not expressing CD7, CD15, CD18, CD36, CD56 and CD64, with asynchronous expression of CD13, CD33, CD117 and HLA-DR in CD34 positive cells, with 40% of CD34 positive blast cells without expression of CD38 and with aberrant expression of CD54, characterizing leukemic stem cells (LSC). She underwent 2 consolidation cycles with Cytarabine (1 g 12/12 hours for 3 days) adapted to her clinical status and had febrile neutropenia after the 1st cycle. Immediately before the 2nd consolidation cycle, the myelogram showed morphological remission and undetectable MRD with a limit of detection and quantification of 10-5. One month after the 2nd consolidation cycle, she was admitted with sepsis and overt disease progression assessed by myelogram. Patient not eligible for intensive chemotherapy due to her clinical and nutritional conditions, she received palliative treatment and died. Unfortunately, immunophenotyping was not performed after relapse, but in the MRD monitoring we could observe immunophenotypic changes from differentiated monocytic cells to myeloid precursor cells at a less mature stage of differentiation including LSC, which support the hypothesis of clonal evolution. In conclusion, to detect MRD in AML patients, who potentially may undergo clonal evolution, it is essential to use a comprehensive panel to detect not only the leukemia-associated immunophenotypes (LAIPs) at diagnosis, but also the “non-normal” immunophenotypes acquired during the treatment in addition to the LSC.