Acute myeloid leukemia (AML) is a group of cancers in which abnormal myeloid precursor cells accumulate in the bone marrow, blood, and sometimes other tissues. Myeloid cells normally develop into granulocytes, monocytes, red blood cells, and platelets. In AML, disrupted differentiation and abnormal survival allow immature cells, often called blasts, to expand while impairing normal blood-cell production. The resulting combination of anemia, infection, and bleeding can become life-threatening over days to weeks.
AML is the most common acute leukemia in adults, although it also occurs in children. It is not a single biological disease: different chromosome rearrangements, gene mutations, and developmental histories produce distinct subtypes with markedly different treatment responses. Diagnosis therefore integrates cell morphology, immunophenotyping, chromosome analysis, and molecular testing rather than relying on the appearance or number of blasts alone.
Treatment ranges from intensive chemotherapy and donor stem-cell transplantation to lower-intensity drug combinations and molecularly targeted medicines. Acute promyelocytic leukemia, a genetically defined subtype, has a particularly distinctive treatment approach based on restoring cell maturation. Across AML subtypes, genetic classification and sensitive measurement of residual disease have become central to estimating relapse risk and selecting subsequent treatment.
Epidemiology, risk factors, and inherited susceptibility
AML occurs at every age, but its frequency rises steeply in later adulthood. United States cancer-registry data place the median age at diagnosis at approximately 69 years, with an annual incidence of roughly four cases per 100,000 people. Rates vary with population age structure, sex, geography, and registry methods. Men are affected more often than women. Comparisons between countries require age standardization, because a population with more older adults can have a higher crude rate without a higher age-specific risk.
Most affected people have no identifiable external cause. Established risk factors include cigarette smoking, substantial benzene exposure, and previous exposure to certain cancer treatments. Alkylating agents, topoisomerase II inhibitors, and radiation therapy can contribute to the development of subsequent myeloid neoplasms. Historically, alkylator-associated disease often appeared after several years and was frequently preceded by abnormal blood counts or myelodysplasia, whereas topoisomerase II inhibitor-associated disease could appear sooner and involve characteristic chromosome rearrangements. These patterns overlap and do not establish the cause of an individual case.
AML may develop after a myelodysplastic syndrome or a myeloproliferative neoplasm. In these settings, a pre-existing abnormal clone acquires additional changes that permit acute leukemic growth. AML arising after cytotoxic treatment and AML following another myeloid disorder often have adverse genetic features, but their outcomes are not determined by clinical history alone. Contemporary classifications record this history alongside the leukemia’s defining genetic abnormalities.
Inherited susceptibility accounts for a minority of cases but has important implications for diagnosis and donor selection. Pathogenic germline variants in genes such as RUNX1, CEBPA, GATA2, and DDX41 can predispose to myeloid malignancy. DDX41-associated disease commonly presents in later adulthood, demonstrating that hereditary predisposition is not restricted to childhood or conspicuous multigenerational families. Some syndromes involve platelet disorders, immune dysfunction, or other clinical features before leukemia develops; others have few preceding signs.
Children with Down syndrome have a distinctive predisposition to myeloid proliferations associated with acquired GATA1 mutations. Transient abnormal myelopoiesis in newborns can resolve spontaneously, but a proportion of affected children subsequently develop myeloid leukemia of Down syndrome. This disease has biological and treatment characteristics different from most adult AML. When inherited predisposition is investigated, testing may require nonhematopoietic tissue, because blood and marrow can contain acquired leukemia-related variants. Genetic counseling also addresses implications for relatives and the suitability of related stem-cell donors.
Cellular origins and molecular mechanisms
Normal blood formation depends on hematopoietic stem and progenitor cells that balance self-renewal with differentiation. AML develops when acquired alterations disrupt this balance, allowing a population of immature cells to persist and expand. The relevant changes affect several interacting processes: transcriptional control, chromatin organization, growth signaling, RNA processing, metabolism, and programmed cell death. No single mechanism explains all AML, and the same clinical label can encompass very different molecular dependencies.
The disease often evolves through several stages rather than beginning with one fully transformed cell. Age-associated clonal hematopoiesis provides an important example. Mutations in genes such as DNMT3A, TET2, or ASXL1 can allow a blood-forming clone to expand without meeting diagnostic criteria for leukemia. Most people with clonal hematopoiesis do not develop AML. In a minority, however, additional alterations produce progressively abnormal differentiation and growth. Some early mutations can remain detectable during remission because they belong to ancestral blood-forming clones rather than to the full leukemic population.
Chromosome rearrangements can create fusion genes that alter gene expression. RUNX1::RUNX1T1 and CBFB::MYH11 disrupt core-binding-factor transcriptional regulation, while PML::RARA produces the differentiation block characteristic of acute promyelocytic leukemia. Rearrangements involving KMT2A alter transcriptional programs associated with immature progenitor cells. These examples demonstrate why chromosome-level changes are not merely descriptive markers: they can identify the mechanism driving the disease.
Other alterations affect signaling or cellular chemistry. Activating FLT3 mutations can sustain proliferative signaling. Mutant IDH1 or IDH2 enzymes produce the abnormal metabolite 2-hydroxyglutarate, which interferes with enzymes involved in epigenetic regulation and contributes to impaired differentiation. NPM1 mutations change the intracellular localization and functions of nucleophosmin and are associated with a characteristic leukemic transcriptional program. TP53 abnormalities impair cellular responses to genomic damage and frequently accompany complex chromosome abnormalities.
AML is also heterogeneous within a single patient. Distinct subclones may carry different combinations of mutations, and their relative abundance can change during treatment. A drug can eliminate a sensitive population while leaving a resistant subclone that later causes relapse. Cells capable of sustaining leukemia are sometimes described functionally as leukemic stem cells, although their surface phenotype and developmental state vary. Interactions with the marrow environment, altered energy metabolism, and resistance to cell death can support their persistence. This biological diversity helps explain why an initial remission does not necessarily mean that every disease-propagating cell has been eliminated.
Classification and the changing definition of AML
AML classification has shifted from a predominantly morphological system to an integrated genetic framework. The French–American–British classification, developed beginning in the 1970s, divided AML into categories M0 through M7 according to maturation and lineage characteristics. These included myeloblastic, promyelocytic, myelomonocytic, monocytic, erythroid, and megakaryoblastic patterns. Although these terms remain useful descriptions, they do not adequately capture the biological and prognostic differences revealed by modern genetic testing.
The fifth edition of the World Health Organization classification and the International Consensus Classification, both introduced in 2022, give greater priority to defining genetic abnormalities. They recognize entities associated with recurrent rearrangements and mutations, including PML::RARA, RUNX1::RUNX1T1, CBFB::MYH11, NPM1 mutation, and selected CEBPA alterations. They also distinguish myelodysplasia-related disease using specified genetic findings and, depending on the system, clinical context. Morphological dysplasia alone is no longer sufficient to assign all cases to a myelodysplasia-related category.
A blast percentage of at least 20% in blood or marrow remains an important diagnostic boundary for many AML categories. However, certain defining genetic abnormalities allow an AML diagnosis below that threshold. The exact requirements differ between the two 2022 classifications and between genetic entities. The International Consensus Classification generally uses a 10% blast threshold for many recurrent genetic categories, while retaining higher thresholds for particular entities. The WHO framework removes a fixed minimum for several genetically defined forms but retains specified thresholds in others. Consequently, a diagnostic report may identify which classification system was applied.
The International Consensus Classification also introduced an MDS/AML category for selected cases with 10–19% blasts. This acknowledges a biological continuum between myelodysplastic neoplasms and AML without implying that all cases in that range are identical. Classification and treatment boundaries are not always the same: age, genetics, clinical status, and trial eligibility may influence management even when two systems assign different labels.
Some related disorders require careful separation from AML. Blast-phase chronic myeloid leukemia arises from an established BCR::ABL1-positive myeloproliferative disease and is not equivalent to newly presenting AML with BCR::ABL1. Mixed-phenotype acute leukemia has convincing features of more than one hematopoietic lineage. Myeloid sarcoma is a tissue mass of myeloid blasts that can accompany marrow AML, precede it, or occur as a relapse. These distinctions matter because superficially similar collections of immature cells can have different disease origins and treatment frameworks.
Clinical manifestations and acute complications
The common manifestations of AML result from failure of normal marrow production. Reduced red-cell production causes anemia, with fatigue, pallor, exertional breathlessness, and sometimes cardiovascular strain. Reduced platelet production causes bruising, petechiae, nosebleeds, gum bleeding, or heavier menstrual bleeding. Reduced numbers or impaired function of mature neutrophils increases susceptibility to infection. A high total white-cell count does not necessarily indicate effective immunity, because much of that count may consist of abnormal blasts.
Symptoms usually develop over weeks, although some patients have a longer history of abnormal blood counts. Fever, loss of appetite, weight loss, and bone discomfort may occur. Gum enlargement and infiltration of the skin are particularly associated with some monocytic forms, but they are not specific enough to establish a subtype. Enlarged liver or spleen can occur; extensive lymph-node enlargement is less characteristic than in many lymphoid malignancies. Myeloid sarcoma may involve skin, soft tissue, bone, lymph nodes, or other sites.
Hyperleukocytosis is often defined as a white-cell count above , but complications depend on blast characteristics as well as the absolute count. Leukostasis describes impaired tissue perfusion associated with circulating leukemic cells and their interactions with vascular endothelium. Pulmonary and neurological manifestations can include hypoxemia, breathlessness, confusion, focal neurological deficits, or intracranial hemorrhage. Clinically important leukostasis can occur below the conventional hyperleukocytosis threshold, particularly in some monocytic leukemias.
Tumor lysis syndrome results from rapid breakdown of malignant cells, spontaneously or after effective therapy. Released intracellular material can cause hyperuricemia, hyperkalemia, hyperphosphatemia, secondary hypocalcemia, and acute kidney injury. Its likelihood reflects disease burden, cell turnover, renal function, and treatment sensitivity. Treatments that rapidly kill or differentiate leukemic cells can change the timing and pattern of these complications, making early laboratory monitoring part of the treatment process.
Disordered blood coagulation is especially important in acute promyelocytic leukemia, where simultaneous activation of coagulation and fibrinolysis can produce severe hemorrhage. Other AML subtypes can also cause disseminated intravascular coagulation, particularly with infection or a large disease burden. Sepsis may be present at diagnosis or arise during treatment-related marrow suppression. Fever can be the only early sign of a serious infection, while inflammation may be muted in profound neutropenia. These acute complications explain why the initial evaluation addresses physiological instability at the same time as definitive leukemia classification.
Diagnostic evaluation and prognostic assessment
Initial testing usually includes a complete blood count with differential and examination of a peripheral-blood smear. The white-cell count may be high, normal, or low. Anemia and thrombocytopenia are common, but their severity varies. Blasts may be visible in blood, although their absence does not exclude AML. Auer rods—needle-like cytoplasmic inclusions composed of granule material—support myeloid differentiation when present, but many AML cases do not contain them.
Bone-marrow aspiration and a core biopsy establish cellular composition, blast proportion, and tissue architecture. Flow cytometry evaluates combinations of cellular antigens to determine lineage and distinguish AML from lymphoblastic and mixed-phenotype leukemias. Myeloperoxidase and markers such as CD13, CD33, CD117, CD34, and HLA-DR contribute to the assessment, but no single surface-marker pattern defines all AML. The interpretation integrates the whole antigen profile with morphology and genetics.
Conventional chromosome analysis identifies large structural and numerical abnormalities. Fluorescence in situ hybridization and molecular assays detect selected rearrangements, while sequencing panels identify clinically relevant gene variants. Rapid testing for PML::RARA is particularly important when acute promyelocytic leukemia is suspected. FLT3, NPM1, CEBPA, IDH1, IDH2, TP53, and genes associated with myelodysplasia-related biology are among the findings that can affect classification, prognosis, or treatment selection. A genetic result must be interpreted in context: a variant may represent inherited predisposition, antecedent clonal hematopoiesis, or the active leukemia itself.
Baseline evaluation also assesses organ function, coagulation, infection, and the ability to tolerate particular treatments. Cardiac assessment may be relevant before anthracycline exposure, while renal and hepatic function influence drug handling and complication risk. Lumbar puncture is not routine for every adult with AML; evaluation of the central nervous system is generally reserved for selected clinical circumstances. Tissue masses may require separate pathological assessment when myeloid sarcoma is suspected.
Prognostic systems combine genetic abnormalities with treatment context. The European LeukemiaNet 2022 framework for intensively treated adults divides AML into favorable-, intermediate-, and adverse-risk groups. Core-binding-factor rearrangements and selected NPM1- or CEBPA-defined cases can fall into favorable categories, whereas TP53 mutation, complex cytogenetics, and specified myelodysplasia-related abnormalities often indicate adverse risk. Coexisting abnormalities can change the assignment. These categories describe group-level outcomes rather than an individual certainty, and their performance differs across treatment regimens. Age, frailty, accompanying illness, and subsequent measurable residual disease provide additional information beyond the baseline genetic category.
Intensive treatment, consolidation, and transplantation
For patients selected for intensive treatment, therapy is commonly divided into remission induction and post-remission treatment. A longstanding induction backbone is “7+3”: seven days of cytarabine with three days of an anthracycline such as daunorubicin or idarubicin. Cytarabine interferes with DNA synthesis, while anthracyclines damage DNA through mechanisms that include topoisomerase II inhibition. Treatment produces a period of profound marrow suppression before normal blood-cell production can recover. Achieving remission is the first objective, but induction alone is generally insufficient for durable disease control.
The induction regimen may be modified according to disease biology. FLT3 inhibitors can be incorporated for appropriate FLT3-mutated AML. Gemtuzumab ozogamicin, a CD33-directed antibody–drug conjugate, is used in selected settings and has particular relevance in favorable-risk disease such as core-binding-factor AML. CPX-351 packages daunorubicin and cytarabine in liposomes at a fixed molar ratio and is used for specified newly diagnosed secondary or therapy-related AML populations. These approaches are not interchangeable: trial eligibility, genetic findings, prior disease, and treatment-related risks determine where the supporting evidence applies.
Post-remission therapy addresses leukemia that remains below the threshold of routine microscopy. Cytarabine-based consolidation is commonly used after intensive induction, particularly in favorable-risk disease. Dose and schedule depend on age, tolerance, protocol, and organ function. Higher-dose cytarabine can cause cerebellar toxicity and ocular inflammation, in addition to marrow suppression. The number and intensity of consolidation courses therefore reflect more than the initial speed of blast clearance.
Allogeneic hematopoietic stem-cell transplantation combines conditioning treatment with blood-forming cells from a donor. Its antileukemic effect comes both from conditioning and from donor immune recognition of residual malignant cells—the graft-versus-leukemia effect. Transplantation can reduce relapse risk, but it also carries risks of infection, organ injury, graft-versus-host disease, and death unrelated to leukemia recurrence. It is therefore not automatically used for every patient who enters remission.
Transplant decisions integrate genetic relapse risk, residual disease, donor availability, organ function, accompanying conditions, and the expected toxicity of the procedure. Reduced-intensity conditioning has extended transplantation to some older adults and people unable to tolerate fully myeloablative regimens, although it does not eliminate transplant-related risks. Some patients receive maintenance therapy after chemotherapy or transplantation. Oral azacitidine, for example, has demonstrated benefit in a defined post-remission population not proceeding to transplantation; it is not pharmacologically interchangeable with injectable azacitidine used on other schedules.
Lower-intensity regimens and molecularly targeted treatment
Treatment intensity is not determined by chronological age alone. Functional status, cognition, organ function, comorbidities, disease genetics, and expected treatment tolerance all contribute. Lower-intensity regimens are important for many older adults and for people whose clinical condition makes conventional induction unsuitable. The term “lower intensity” does not mean low risk: these treatments can still cause prolonged cytopenias, serious infection, hospitalization, and treatment-related death.
Azacitidine and decitabine are hypomethylating agents that influence DNA methylation and cellular differentiation. Their combination with venetoclax changed the treatment landscape for patients ineligible for intensive induction. Venetoclax inhibits BCL-2, a protein that helps cells resist programmed death. In the VIALE-A randomized trial, published in 2020, azacitidine plus venetoclax produced a median overall survival of 14.7 months compared with 9.6 months for azacitidine plus placebo in the studied population. Complete remission with or without full blood-count recovery occurred in approximately 66% and 28%, respectively. These results describe that trial population, not a uniform expectation for every molecular subtype.
Venetoclax-based therapy requires particular attention to tumor lysis risk, marrow suppression, and drug interactions. Strong CYP3A inhibitors, including several antifungal medicines frequently used in AML care, increase venetoclax exposure and require protocol-specific dose modification. After response, persistent low blood counts may reflect treatment effect rather than active leukemia, so marrow findings help distinguish the two. Treatment schedules are often adjusted over successive cycles according to count recovery and disease response.
Targeted cancer therapies exploit specific molecular dependencies. Midostaurin and quizartinib have roles alongside chemotherapy in defined newly diagnosed FLT3-mutated populations, while gilteritinib is established for relapsed or refractory FLT3-mutated AML. Ivosidenib targets mutant IDH1 and enasidenib targets mutant IDH2; olutasidenib is another IDH1 inhibitor. Ivosidenib combined with azacitidine has demonstrated benefit in an appropriate newly diagnosed IDH1-mutated population unsuitable for intensive chemotherapy. The indications and evidence differ by drug, mutation, treatment line, and jurisdiction.
Targeting a mutation does not necessarily eradicate every leukemic clone. Resistance can arise through additional signaling changes, altered drug-binding sites, or expansion of populations that lack the original target. Some targeted agents also cause distinctive toxicities. IDH inhibitors can produce differentiation syndrome as immature cells mature and inflammatory mediators affect tissues. Menin inhibitors target transcriptional dependencies associated with KMT2A rearrangements and some NPM1-mutated leukemias; their development illustrates the growing use of molecularly defined treatment groups. Evidence from one targeted population cannot automatically be extended to all AML.
Acute promyelocytic leukemia as a distinct subtype
Acute promyelocytic leukemia (APL) is usually defined by the PML::RARA fusion, most often produced by the chromosome translocation . The fusion protein interferes with retinoic-acid-dependent transcription and arrests myeloid maturation at the promyelocyte stage. Morphology may show heavily granulated abnormal promyelocytes with bundles of Auer rods, although a microgranular variant can look different and may present with a high white-cell count. Molecular confirmation distinguishes classic APL from other leukemias with similar appearances.
APL has a distinctive early hazard: severe coagulopathy can produce fatal intracranial, pulmonary, or other bleeding before leukemia-directed treatment has had time to work. Abnormal promyelocytes promote coagulation activation and excessive fibrinolysis through several interacting pathways. Clinical protocols therefore treat suspected APL as a hematologic emergency, beginning all-trans retinoic acid while urgent confirmation proceeds and providing intensive blood-product support. Its high curability after successful treatment does not remove this early risk.
All-trans retinoic acid, commonly abbreviated ATRA, reverses the differentiation block imposed by PML::RARA. Arsenic trioxide acts through complementary mechanisms, including effects on the PML component and degradation of the fusion protein. The combination established a major example of treating cancer by reversing its maturation block rather than relying exclusively on conventional cytotoxic chemotherapy. Randomized studies in newly diagnosed non-high-risk APL showed excellent disease control with ATRA plus arsenic trioxide, allowing many such patients to avoid conventional chemotherapy.
Risk-adapted treatment commonly distinguishes patients by presenting white-cell count, with a count above defining high-risk disease in widely used systems. High-risk regimens incorporate additional measures to control leukemic burden, such as anthracycline-based treatment or gemtuzumab ozogamicin within particular protocols. Treatment details differ across cooperative groups, but the central components remain rapid differentiation therapy, control of coagulopathy, and molecular assessment of response.
Differentiation syndrome can occur during ATRA- or arsenic-based therapy. Manifestations include fever, weight gain, edema, respiratory compromise, pulmonary infiltrates, hypotension, and kidney dysfunction. It reflects treatment-associated inflammatory and tissue effects rather than simple progression of the original blast population. Arsenic also requires attention to electrolyte abnormalities and cardiac electrical effects. Molecular monitoring detects PML::RARA transcripts, with results interpreted according to treatment stage; early positivity during differentiation is not equivalent to persistent disease after completion of consolidation. Relapsed APL has its own treatment framework, influenced by prior exposure to arsenic and ATRA and by molecular response.
Supportive care and treatment-related complications
Supportive care is an essential component of AML treatment rather than a secondary addition. Leukemia and its treatment both impair blood-cell production, often for prolonged periods. Red-cell and platelet transfusions support oxygen delivery and reduce bleeding risk, with thresholds adapted to symptoms, bleeding, procedures, and disease subtype. APL requires a particularly intensive approach to coagulation support. Transfusion exposure can also lead to reactions, alloimmunization, and, over time, iron accumulation, making product selection and monitoring relevant.
Infection is a major cause of illness and death. Profound neutropenia, mucosal injury, venous access devices, and altered cellular immunity create different routes of infection. Bacteria can enter through damaged oral or intestinal surfaces, while invasive molds become important during prolonged neutropenia. Antibacterial, antiviral, and antifungal prophylaxis are used in defined treatment settings according to expected immune suppression and local epidemiology. Fever during neutropenia is managed through time-sensitive clinical pathways because deterioration can precede definitive microbiological identification.
Drug interactions are unusually important in AML because antileukemic agents and anti-infective medicines are often administered together. Azole antifungals can alter exposure to venetoclax and several kinase inhibitors. Multiple medicines may also affect cardiac repolarization or liver function. The resulting clinical decisions depend on the complete regimen rather than on the properties of one drug in isolation. Pharmacovigilance and medication reconciliation therefore contribute directly to treatment delivery.
Treatment can injure tissues outside the marrow. Anthracyclines can impair cardiac function, cytarabine can cause neurological and ocular toxicity, and several agents can affect hepatic or renal function. Mucositis, nausea, altered taste, diarrhea, and reduced intake may compromise nutrition and physical function. Tumor lysis prevention and monitoring accompany treatments likely to produce rapid cell clearance. Transplant recipients face additional complications related to conditioning, graft-versus-host disease, immune suppression, and delayed immune reconstitution.
AML also affects fertility, psychological well-being, work, family responsibilities, and physical independence. The relevance of fertility preservation depends on treatment urgency, age, reproductive circumstances, and the therapies anticipated. Rehabilitation can address deconditioning during and after prolonged treatment, while symptom management and psychological support may continue throughout the disease course. Palliative care can be provided alongside active leukemia treatment, addressing symptoms, communication, and quality of life; it is not synonymous with stopping antileukemic therapy. When disease-directed treatment no longer controls leukemia, transfusion requirements, infections, bleeding, and preferred care settings become important components of ongoing clinical care.
Response assessment, relapse, and long-term outcomes
A conventional complete remission requires fewer than 5% marrow blasts, no circulating blasts or extramedullary leukemia, and recovery of specified blood counts. Other response categories recognize marrow remission without full count recovery. These distinctions matter because apparent blast clearance can coexist with substantial treatment-related cytopenias, and response rates calculated with different definitions are not directly interchangeable. Remission is a measured disease state, not necessarily a cure.
Measurable residual disease (MRD) testing detects leukemia below the sensitivity of routine microscopy. Multiparameter flow cytometry identifies abnormal cell populations, while molecular assays track suitable gene mutations or fusion transcripts. Depending on the target, specimen, and method, molecular techniques may detect approximately one leukemic cell among ten thousand to a million cells. Such sensitivity is not universal, and a negative result means disease was not detected within the assay’s capabilities, not that no leukemic cell exists anywhere in the body.
The choice of molecular marker is critical. NPM1 mutations and core-binding-factor fusion transcripts can provide informative targets. In contrast, persistence of DNMT3A, TET2, or ASXL1 mutations may represent antecedent clonal hematopoiesis rather than residual AML. MRD timing also matters: results after induction, after consolidation, and before or after transplantation answer different questions. MRD positivity is generally associated with increased relapse risk, but its clinical interpretation depends on subtype, treatment, assay quality, and the magnitude and trajectory of the finding.
Relapsed AML reappears after a prior remission, whereas refractory AML has not achieved the required response to initial treatment. Repeat genetic assessment is important because the relapse clone may differ from the one detected at diagnosis. Available approaches include salvage chemotherapy, mutation-directed treatment, lower-intensity combinations, and transplantation in suitable circumstances. Clinical trials test new drugs, combinations, immune approaches, and strategies guided by residual disease. Unlike B-cell malignancies, AML presents a particular challenge for immune targeting because many candidate antigens are also expressed on normal myeloid progenitors; eliminating the malignant population can therefore also damage essential blood formation.
Long-term outcomes vary widely. Broad United States registry figures place five-year relative survival at roughly one third, but this average combines children, younger adults, very old patients, multiple genetic subtypes, and different treatment eras. It is not an appropriate estimate for every individual subgroup. Some favorable-risk AML and most successfully treated APL have substantially better outcomes, while adverse-genetic disease and AML in medically frail older adults remain difficult to control. Survivors may require follow-up for relapse, chronic graft-versus-host disease, cardiac effects, endocrine or reproductive consequences, second cancers, and persistent functional limitations. Relapse becomes less common with increasing time in remission, but late recurrence can occur.