Relapse of NPM1-Mutated AML with Extramedullary Manifestation 17 Years after Allogeneic Hematopoietic Stem Cell Transplantation

The majority of patients with acute myeloid leukemia (AML) with the NPM1 mutation achieve remission with intensive chemotherapy. However, many patients subsequently relapse, which occurs frequently within the first 2–3 years after therapy, while late relapse after more than 10 years is rare and can also represent secondary/therapy-associated AML without the NPM1 mutation. Here, we present a case of NPM1-mutated AML that developed medullary and extramedullary relapse 17 years after allogeneic stem cell transplantation, maintaining the NPM1 mutation and all other genetic alterations detected at first diagnosis. This exceptionally long latency between diagnosis and relapse of a genetically highly related leukemic clone implies the existence of therapy-resistant, persisting dormant leukemic stem cells in NPM1 mutant AML.


Introduction
NPM1 mutations are among the most frequent genetic alterations in acute myeloid leukemia (AML) and are associated with distinct morphological, genetic, and clinical features. Intensive chemotherapy with or without allogeneic stem cell transplantation (allo-HSCT) leads to remission in 70-90% of patients with NPM1 mut AML, but about 30-50% of patients relapse due to persistent leukemic stem cells [1]. Most AML relapses occur early, usually within 12-18 months after the frst complete remission (CR1), while late (LR) and very late relapses (VLR), defned as >3 years and >5 years after CR1, respectively, are rare and occur in 1-3% of the patients [2,3], with only a few case studies describing relapses after more than 10 years [2,[4][5][6][7]. Recent studies revealed that many very late relapses are secondary, therapy-related AMLs with a distinct genetic profle as compared to the frst AML at diagnosis [8,9]. Extramedullary (EM) AML is a rare manifestation of AML and occurs in only 2-9% of newly diagnosed patients [10]. Nevertheless, it occurs more frequently in relapsed patients, and EM relapses have been reported more frequently in transplant recipients than in patients treated with chemotherapy alone.
Here, we report the case of a patient with NPM1-mutated AML who developed a bone marrow (BM) and EM relapse 17 years after the frst diagnosis ( Figure 1).

Case Presentation
A 23-year-old male patient was diagnosed with AML (initial white blood cell count: 50/nL; hemoglobin: 3.6 g/dL; platelet count: 67 × 10 3 /μL) in 2003. A BM biopsy showed 80% blasts with monocytic features without evidence of extramedullary disease. Conventional cytogenetics revealed a 46, XY, del(9q) karyotype. Flow cytometry revealed the expression of CD33 and HLA-DR, without CD34 or CD117 expression. Molecular genetics and MRD assessment were initially not performed as they were not available at that time. No relevant comorbidities, no family history of malignancies, and no known environmental exposures to substances that damage DNA were reported. Te patient received intensive induction therapy with cytarabine, daunorubicin, and thioguanine and achieved a complete remission. He then underwent allogeneic HSCT from an HLA-identical sibling, which was well tolerated without the development of GvHD. Bone marrow assessment revealed a CR with complete donor chimerism one month after transplantation until the end of follow-up seven years after transplantation.
In 2020, 17 years after transplantation, the patient was admitted to the hospital sufering from severe renal impairment, urine retention, and a mechanical ileus. MRI of the abdomen revealed a 12 × 9 cm large mass infltrating the lesser pelvis. Furthermore, a computed tomography (CT) scan of the chest revealed lesions in both lungs suspicious of metastases. However, histologic analysis of the tumor that infltrated the colon mucous membrane showed myeloid leukemic cells that were strongly positive for CD33 and CD117 and negative for CD34.
Retrospective analyses of the diagnostic AML sample obtained 17 years before revealed that the DNMT3A R882H and NPM1 Type A mutations were already present at initial diagnosis, while no other mutations were found in FLT3, IDH1/IDH2, RUNX1, TP53, or ASXL1. A drop in donor chimerism revealed that leukemia originated from the host and not the donor. Te persistence of two identical gene mutations and one chromosomal aberration highly implies that relapse evolved directly from persisting cells of the initial leukemic clone rather than an independent clone with newly acquired aberrations.
After induction therapy with cytarabine and daunorubicin, followed by consolidation therapy with intermediatedose cytarabine and radiation of the pelvic chloroma, a second complete hematologic remission was achieved with complete regression of the pelvic tumor. Radiotherapy of the lung and pelvis was performed. A second HSCT from an HLA haploid donor was planned, but unfortunately, the patient died due to neutropenic sepsis during consolidation therapy.

Discussion
Tis case represents one of the latest relapses described in AML with a confrmed NPM1 mutation at both diagnosis and relapse. Te presence of the same gene mutations, NPM1 mutA and DNMT3A R882H , as well as a del(9q), strongly argues that the second AML directly evolved from the initial AML clone [8]. While a complex karyotype that was detected in the relapsed AML in addition to the initial alterations is frequently observed in patients exposed to DNA damaging agents, typical therapy-associated aberrations such as deletions 5 or 7 or mutations in TP53 [11] were absent. Tis is in contrast to our previous fndings that very late reoccurrence of leukemia in NPM1 mut AML often represents a genetically distinct, therapy-associated AML without NPM1 mutation [8,9].

Case Reports in Hematology
Bertoli et al. showed that "true," genetically stable, very late relapses after 5 years are more common in patients with the NPM1 mutation than in AML without the NPM1 mutation [3]. Watts et al. recently described four cases with very late relapse [12] that were similar to our patient: they had a monocytic phenotype, underwent allo-HSCT, and relapsed with the extramedullary disease. While molecular diagnostic data was not available for any of these cases, monocytic diferentiation and extramedullary disease are indicative that these cases harbored an NPM1 mutation. Bolli et al. described a case of a myeloid sarcoma with an NPM1 mutation occurring 20 years after the frst diagnosis of AML but no genetic information was available at the initial diagnosis [5]. EM disease, although a rare manifestation of AML, is more frequently seen at relapse and after allo-HSCT compared to standard chemotherapy, suggesting a difering underlying pathogenic mechanism [10,12]. Characteristics that seem to correlate with EM relapse are VLR, monocytic diferentiation, favorable cytogenetics, younger age, and NPM1 mutations [3,10,12], all of which were present in our case.
Te underlying pathophysiologic mechanisms of such long quiescence of a leukemic stem cell are unclear. Te microenvironment has a strong infuence on normal and leukemic stem cells, and the transition of such a cell outside the bone marrow may have induced an awakening from dormancy. Decreased immunosurveillance may be a possible factor. Graft versus leukemia might not be that efective outside the BM, explaining the higher rates of EM relapse after allo-HSCT [13][14][15].
Very late relapse AML with clear proof of a genetic relationship, as in our case, has several implications. From a clinical perspective, the occurrence of very late hematologic and extramedullary relapse in NPM1 mut AML has to be considered and warrants awareness of the possibility of EM relapse. AML cases with VLR further demonstrate that leukemia stem cells (LSCs) may persist for many years, most likely in a dormant state that is invisible to the immune system, even after allo-HSCT. It is conceivable that many patients in long-term remission also have persisting, dormant LSCs, as also suggested by the detection of low MRD levels in patients in long-term remission [16]. Te reasons for the "awakening" of persisting LSCs are not clear and warrant the further molecular and functional characterization of dormant LSCs. Furthermore, understanding how AML cells persist in therapy over many years may help fnd ways to eradicate them and cure patients.

Data Availability
Te data used to support the fndings of this study are available from the corresponding author upon request.

Conflicts of Interest
Te authors declare that there are no conficts of interest.