MIRAGE Syndrome and SAMD9-SAMD9L-Related Disorders

Gain-of-function syndromes of SAMD9 and SAMD9L: MIRAGE syndrome, monosomy 7 myelodysplasia and leukaemia syndrome, SAMD9L-related syndrome (ataxia-pancytopenia, spinocerebellar ataxia type 49 and SAMD9L-associated autoinflammatory disease) and normophosphataemic familial tumoral calcinosis, within the context of hereditary predisposition to myeloid neoplasms

Literature and regulatory information updated: 31 October 2026.

0. Initial Reference Table

1. Introduction

1.1. From the Li Family to the Discovery of the Genes

The first clinical description of what is now recognised as SAMD9L-related syndrome was published in 1978 by Frederick Li and colleagues, who studied a sibship of five children in which the eldest brother died of acute myeloid leukaemia at 10 years, two siblings died of hypoplastic anaemia at 5 and 9 years, and the proband, aged 6, had mild pancytopenia, platelet dysfunction, immunodeficiency and hypoplastic marrow with about 18% blasts [25]. The marrow of the proband and of one sibling showed monosomy of a group C chromosome, and at least four siblings and the father had cerebellar ataxia with a small cerebellum at autopsy or on computed tomography (CT); the authors excluded ataxia-telangiectasia and Fanconi anaemia and proposed a new syndrome [25]. The Li family was followed for decades: the proband developed hypoplastic anaemia at 3 years and died of acute myelomonocytic leukaemia at 7, with progressive expansion, over 25 months, of a marrow clone with monosomy 7, and the autopsy of one sibling showed cerebellar atrophy with loss of 90% of Purkinje cells [11]. The initial designation was “myelocerebellar syndrome”, later replaced by ataxia-pancytopenia syndrome [4].

In 2016, two groups independently identified the germline variants responsible for these syndromes. Chen and colleagues, in a four-generation North American family (UW-AP), combined linkage analysis and exome sequencing and found the variant SAMD9L c.2640C>A p.(His880Gln), which co-segregated with the disease, and in the Li family identified a second variant, c.3587G>C p.(Cys1196Ser); the patients' lymphoblastoid lines spontaneously lost the mutated allele through copy-neutral loss of heterozygosity of 7q, a first indication of the somatic rescue that defines these diseases [26]. In the same year, Narumi and colleagues described 11 Japanese patients with syndromic adrenal hypoplasia and de novo variants in SAMD9, proposed the acronym MIRAGE and showed that the mutant proteins profoundly inhibited cell growth and that two patients with MDS had lost precisely the chromosome 7 carrying the variant [27]. There followed the series of Buonocore and colleagues, with eight children with MIRAGE in whom progressive loss of the mutated allele in the marrow was associated with longer survival [28], and of Tesi and colleagues, who demonstrated the gain of function of SAMD9L in two Nordic families and described revertant mosaicism as a frequent mechanism of haematological remission [29].

Between 2017 and 2018, SAMD9 and SAMD9L variants were found in families with myelodysplasia and monosomy 7 without malformations, which linked these genes to the former familial monosomy 7 syndrome [30–32]. The French bone marrow failure cohort of Bluteau and colleagues showed that SAMD9 and SAMD9L accounted for 16 of the 86 patients with a molecular diagnosis (18.6%), with frequent transient aplasia and monosomy 7 [24]. In 2020, de Jesus and colleagues, at the National Institutes of Health (NIH), described six patients with de novo frameshift variants in SAMD9L and a severe autoinflammatory picture of perinatal onset, which they named SAMD9L-associated autoinflammatory disease (SAAD) [5]. In 2021, the European childhood MDS group (EWOG-MDS, European Working Group of MDS and SAA in Childhood) determined, in a cohort of 669 paediatric MDS, that germline SAMD9 or SAMD9L variants were present in 8% of consecutive cases [22]. In 2022, a five-generation family from Menorca with adult-onset ataxia and no haematological abnormalities led to the definition of spinocerebellar ataxia type 49 (SCA49) [33].

1.2. Nomenclature and Reclassifications

The accumulated nomenclature reflects the order in which the phenotypes were discovered and not distinct biological entities. The SAMD9L GeneReview, in its August 2026 update, considers that ataxia-pancytopenia, monosomy 7 predisposition syndrome, SAAD and SCA49 represent a continuum of manifestations caused by heterozygous germline gain-of-function variants in SAMD9L, and calls it SAMD9L-related syndrome [4]. The French national protocol for bone marrow aplasias, of March 2023, also adopts the single designation “SAMD9/SAMD9L syndrome”, because none of the clinical signs is specific to one of the genes, and recorded that Orphanet did not yet have its own code for the SAMD9/SAMD9L group, with only the MIRAGE code (ORPHA:494433) existing [23]. The authors of the St. Jude registry proposed, for the same reason, the designation “SAMD9 and SAMD9L syndromes” for the whole spectrum, reserving “syndromic SAMD9” for patients with MIRAGE features, because many severe SAMD9 patients do not have all the elements of the acronym and because SAMD9L variants may cause only haematological abnormalities [2]. OMIM keeps separate entries: MIRAGE (617053) and monosomy 7 myelodysplasia and leukaemia syndrome type 2 (M7MLS2, 619041) for SAMD9; ataxia-pancytopenia (ATXPC, 159550), monosomy 7 myelodysplasia and leukaemia syndrome type 1 (M7MLS1, 252270) and SCA49 (619806, with a provisional gene-phenotype relationship) for SAMD9L [3, 9, 11–13]. SAAD has its own Orphanet code (ORPHA:619367) but no OMIM number [6]. Before this unification, Davidsson and colleagues had already argued that ataxia-pancytopenia and familial monosomy 7 syndrome were unlikely to be distinct entities and reflected different degrees of penetrance and severity of the same spectrum; the classical criterion for familial monosomy 7 syndrome, which requires two siblings with a myeloid neoplasm with monosomy 7, probably led to underdiagnosis, because relatives who do not develop MDS frequently rescue the germline variant [34].

The SAMD9L GeneReview draws attention to the fact that SCA49 was described in a single family with purely neurological manifestations and in one patient with a frameshift variant upstream of the usual region, and that the possibility of protective revertant events was not sufficiently investigated in any of the reports, which leaves open whether SCA49 is a separate entity or an expression of the same spectrum with haematopoietic rescue [4, 35].

1.3. Current Haematological Classifications

The 2016 revision of the World Health Organization (WHO) classification created the category of myeloid neoplasms with germline predisposition, and the 2017 European LeukemiaNet (ELN) added a category of the same name [20]. The 5th edition of the WHO (2022) groups myeloid neoplasms with germline predisposition into three subtypes and places pathogenic or likely pathogenic germline variants of SAMD9 (MIRAGE syndrome) and of SAMD9L (SAMD9L-related ataxia-pancytopenia, with the note that ataxia is not always present) in the subtype “with potential organ dysfunction”, alongside GATA2, the bone marrow failure syndromes, the telomere biology disorders, the RASopathies, Down syndrome and Bloom syndrome [36]. The 2022 International Consensus Classification (ICC) incorporated SAMD9 and SAMD9L as new entities, “myeloid neoplasms with germline SAMD9 mutation” and “… SAMD9L”, in the group of haematological neoplasms with germline predisposition associated with a constitutional disorder affecting multiple organ systems, and requires the predisposition to be written as a qualifier of the neoplasm diagnosis [37]. The 2022 ELN recommendations for adult AML list predisposition due to SAMD9 and SAMD9L as autosomal dominant, with onset in childhood more than in adulthood and with no prior platelet disorder described [38]. The ELN expands the “R” of MIRAGE as renal abnormalities and the “G” as genitourinary anomalies, a reading that differs from Narumi's original definition (restriction of growth and genital phenotypes) [27, 38].

1.4. Definition of the Entities in Scope

MIRAGE syndrome is defined by the association of a heterozygous germline gain-of-function variant in SAMD9 with a multisystem phenotype with six core features: myelodysplasia, infection, restriction of growth, adrenal hypoplasia, genital phenotypes and enteropathy; in the absence of consensus criteria, the GeneReview uses the operational definition of four or more core features in 46,XY individuals and three or more in 46,XX individuals [1]. M7MLS2 corresponds to carriers of gain-of-function variants in SAMD9 who present with refractory cytopenia, MDS or AML with monosomy 7 and few or no MIRAGE features [1, 3]. SAMD9L-related syndrome encompasses any combination of cytopenias, immunodeficiency, predisposition to bone marrow failure and MDS with monosomy 7, cerebellar ataxia and other neurological manifestations, in carriers of a germline gain-of-function variant in SAMD9L [4]. SAAD is the neonatal-onset autoinflammatory subtype caused by truncating SAMD9L variants that remove the autoregulatory domain but preserve enzymatic function [4]. Normophosphataemic familial tumoral calcinosis (NFTC) is an autosomal recessive disease due to biallelic loss of function of SAMD9, with no phenotypic overlap with MIRAGE, and is included in this article as an allelic disorder [1, 7].

1.5. Related Entities

Hereditary predisposition to myeloid neoplasms as a whole includes GATA2, RUNX1, CEBPA, DDX41, ANKRD26 and ETV6 deficiency, alongside the SAMD9 and SAMD9L syndromes; the present article examines the SAMD9/SAMD9L class in depth. Fanconi anaemia, Shwachman-Diamond syndrome, severe congenital neutropenias and Diamond-Blackfan anaemia are part of the differential diagnosis. X-linked congenital adrenal hypoplasia, familial glucocorticoid deficiency and IMAGe syndrome are included here only as differential diagnoses of the adrenal form of MIRAGE.

2. Molecular Genetics and Pathophysiology

2.1. The SAMD9 and SAMD9L genes

SAMD9 and SAMD9L are contiguous paralogues on the long arm of chromosome 7, in band 7q21.2, separated by a few tens of kilobases (GRCh38 coordinates chr7:93,099,512–93,118,022 for SAMD9 and chr7:93,130,053–93,148,400 for SAMD9L) [10, 14]. Each gene has a single coding exon and produces a cytosolic protein of about 1500 amino acids; the two proteins share about 60% sequence identity and overlapping functions in the inhibition of viral replication and the negative control of cell proliferation [2]. The two genes are thought to have arisen from duplication of an ancestral gene in placental mammals; over the course of evolution, species have retained one or both, so that cattle have only SAMD9, mice only Samd9l and primates both, which limits murine models [2]. The MANE Select reference transcripts are NM_017654.4 (SAMD9) and NM_152703.5 (SAMD9L) [8].

Both genes are among the interferon-stimulated genes: type I, II and III interferons increase the expression of SAMD9L, and stimulation of CD34+ progenitors with interferon-α or interferon-γ induces SAMD9L [4, 29]. Expression of SAMD9 is particularly high in the fetal adrenal gland, the colon, the bone marrow and the thymus, the organs most affected in MIRAGE [21]. In the mouse, Samd9l is expressed in most tissues, with a maximum in the kidney, the spleen, the stomach and the adrenal gland [14].

Homology analyses classify SAMD9 and SAMD9L in the STAND family (signal transduction ATPases with numerous domains), with a central nucleoside triphosphatase domain with a P-loop, C-terminal tetratricopeptide repeats that would function as ribonucleic acid sensors, an OB-fold domain binding deoxyribonucleic acid (DNA) and N-terminal domains (SAM and AlbA_2) [2]. In 2023, Zhang and colleagues showed that SAMD9 is an anticodon nuclease that specifically cleaves phenylalanine transfer ribonucleic acid (tRNA-Phe), through the N-terminal effector domain, which causes ribosomal stalling at phenylalanine codons and stress signalling; that activity is latent in normal cells, is activated by poxvirus infection and becomes constitutive with the pathogenic variants [39]. The cryo-electron microscopy structures of SAMD9 published in 2026 show a closed, autoinhibited conformation, stabilised by an ATP-bound nucleotide-binding and oligomerisation domain and by a network of intramolecular interactions; the recurrent gain-of-function variants lie at these interfaces and destabilise them, and compensatory mutations designed from the structure restore autoinhibition [40]. No pathogenic variant of SAMD9L has been described in the SAM domain, which rules out protein-protein interaction through that domain as the main mechanism [4].

2.2. Disease mechanism: toxic gain of function

The variants that cause MIRAGE, M7MLS2 and SAMD9L-related syndrome are gain-of-function: the wild-type protein has a slight antiproliferative capacity, and the mutant protein has an exaggerated antiproliferative capacity [1]. All the germline variants tested functionally in the St. Jude registry (83 of 130) amplified the growth-inhibitory effect relative to the wild-type protein, without altering messenger ribonucleic acid or protein levels [2]. In cells from patients with SAMD9L-related syndrome a global block of translation at the elongation step is observed, and SAMD9L, like SAMD9, cleaves tRNA-Phe and creates ribotoxic stress; the wild-type protein requires viral activation, whereas the gain-of-function variants render it constitutively active [4, 41, 42]. Russell and colleagues showed that repression of translation is cell-autonomous and proportional to the protein level, and that both the truncating variants of SAAD and the missense variant NM_152703.5:c.2956C>T p.(Arg986Cys), recurrent in ataxia-pancytopenia, exaggerate it even at low levels of expression [42]. Narumi and colleagues had previously described, in fibroblasts from MIRAGE patients, restricted growth, less epidermal growth factor receptor at the membrane, enlarged early endosomes and giant vesicles with late endosome markers, and proposed a defect of endosomal fusion [27]; the more recent hypothesis interprets these endosomal changes as a secondary consequence of chronic repression of translation and activation of the integrated stress response, with phosphorylation of the initiation factor eIF2α [21].

Two animal models illustrate the systemic consequences. The mouse carrying the murine equivalent of the human SAMD9 variant NM_017654.4:c.2305G>A p.(Asp769Asn) in constitutional homozygosity has growth restriction, multiorgan failure with testicular atrophy and cystic renal degeneration, a short lifespan, anaemia and B lymphopenia; conditional haematopoietic expression of Samd9l p.(Trp1171Arg) (murine protein numbering), the orthologue of the human SAMD9L variant NM_152703.5:c.3538T>C p.(Trp1180Arg), reduces mature B cells and activates inflammation mediated by transforming growth factor β [2]. Conversely, mice with loss of function of Samd9l, in heterozygosity or homozygosity, develop late myelodysplasia (at around 2 years) in about 40% of the animals, and infection with the MOL4070A retrovirus brings the disease forward, with 84% of the animals dying of myeloid neoplasia within 15 months [2, 14]. These data support the role of SAMD9L as a tumour suppressor gene whose loss, and not gain of function, favours myeloid proliferation. The latency of myelodysplasia did not differ between heterozygotes and homozygotes, which suggests that haploinsufficiency is sufficient, but it was not investigated whether the second allele was lost in heterozygotes during transformation; the same work placed SAMD9L in the complex that promotes homotypic fusion of endosomes and degradation of cytokine receptors, such as KIT, and Samd9l+/− and Samd9l−/− cells had a small advantage in competitive transplantation [34].

2.3. Somatic genetic rescue and monosomy 7

The cellular toxicity of the mutant protein creates intense selective pressure in haematopoietic tissue, which renews itself continuously: cells that eliminate or neutralise the mutated allele gain an advantage and expand. This phenomenon, termed somatic genetic rescue, occurs in at least two thirds of carriers and gives rise to clonal haematopoiesis, which may be adaptive, of indeterminate significance or maladaptive (premalignant) [2]. In the EWOG-MDS cohort, 61% of patients with SAMD9- or SAMD9L-related MDS had somatic rescue; of these, 95% had maladaptive events (monosomy 7, with or without mutations of cancer genes) and 51% adaptive events (revertant uniparental disomy of 7q or somatic SAMD9/SAMD9L variants), often in the same patient; single-cell sequencing showed several competing clones in each patient [22].

There are three routes of rescue. The first is total or partial loss of the chromosome 7 carrying the pathogenic allele (monosomy 7 or del(7q)), an «adaptation by aneuploidy» in which the chromosome lost is always the mutated one [2, 3]. Monosomy 7 eliminates the toxicity of SAMD9 or SAMD9L, but creates haploinsufficiency of EZH2, MLL3 and other 7q genes necessary for normal haematopoiesis, which greatly increases the risk of myelodysplasia; transformation to AML requires additional mutations [4]. Haploinsufficiency of EZH2, the catalytic component of Polycomb repressive complex 2, prepares the ground for leukaemic transformation and is associated with treatment resistance in paediatric AML [20]. The second route is the acquisition of somatic variants on the same allele (in cis) that neutralise the germline variant; about 53% are truncating and 47% missense, distributed throughout the protein, and some inhibit growth when expressed alone but compensate for the effect of the germline variant when co-expressed [2]. The third route is uniparental disomy of 7q (UPD7q), a copy-neutral loss of heterozygosity that replaces the segment carrying the mutated allele with a copy of the normal chromosome and leaves the cell with two wild-type alleles [4]. UPD7q was documented in 21 patients up to 2024, 17 with SAMD9L and 4 with SAMD9, all with missense variants, and is associated with stable remissions of up to 20 years without MDS or AML; it is rare in MIRAGE and in SAAD, probably because the severity and early onset of these forms leave no time for it to emerge [2].

Monosomy 7 associated with SAMD9 or SAMD9L behaves in a way that distinguishes it from monosomy 7 in other predispositions. In young children it may disappear spontaneously (transient monosomy 7) or progress to leukaemia, and for this reason the St. Jude group classifies it as of indeterminate risk; transient monosomy 7 is highly specific to these two syndromes [2, 20]. Pastor and colleagues described two patients in whom monosomy 7 disappeared and was replaced by UPD7q with the wild-type allele, with spontaneous recovery [31]. When monosomy 7 progresses, it acquires mutations in leukaemia driver genes such as SETBP1, ASXL1, RUNX1, PTPN11, CBL, EZH2, ETV6 and genes of the RAS pathway, which are not specific to these syndromes and also arise in paediatric MDS with GATA2 deficiency or without known predisposition [2, 43]. Kotmayer and colleagues emphasise that the spectrum of somatic mutations is determined by monosomy 7 and not by the germline predisposition, and that, in an analysis of 50 primary paediatric MDS, children with monosomy 7 had more somatic mutations than those with a normal karyotype (56% versus 18%) [43].

Somatic rescue has limits. Adaptive clones do not always produce clinical remission, perhaps because they arise in committed progenitors and not in stem cells, and there are patients in whom only the lymphoid compartment is rescued [2]. Rescue restores haematopoietic and immune function but does not correct the other organs: neurones, which are post-mitotic cells, do not renew themselves, which explains the progression of the neurological deficits [2]. Rescue is clonal and mosaic, and may also occur in people without apparent disease: Wimalachandra and colleagues described a 3-year-old boy with severe aplastic anaemia and a novel SAMD9L variant, shared by his father and two healthy siblings, who received a transplant from his matched sister and, about two years later, developed donor-type aplasia [4, 44]. Conversely, Tan and colleagues reported donor-derived somatic rescue with prolonged haematopoietic stability after a sibling transplant in a SAMD9 syndrome [45].

The timing and the cellular level of rescue have been studied in the Nordic families with SAMD9L. The allelic fraction of the variant in Guthrie cards collected at birth was similar for the wild-type and mutated alleles, which excludes constitutional mosaicism and indicates that the UPD7q found later is an acquired clonal event, probably between 1 and 4 years of age, the age at which many carriers had had transient pancytopenia followed by recovery [34]. In lymphoblastoid lines from carriers, UPD7q with loss of the mutated allele arose between 3 weeks and 6 months of culture [34]. Droplet digital polymerase chain reaction in subpopulations separated by cytometry suggested that rescue had occurred in multipotent progenitors and not in haematopoietic stem cells, a conclusion based on few patients; among mature cells, NK cells had the highest proportion of rescued alleles, in keeping with the maximal expression of SAMD9L in NK cells and monocytes and with the deficiency of these lineages in symptomatic carriers [34]. Some revertant variants in cis affected the very codon of the germline variant and produced a substitution with lesser gain of function, and a single patient could have clones with UPD7q and an MDS clone with an abnormality of chromosome 7, including an unbalanced translocation der(1;7) with loss of the segment carrying the mutated allele [34].

2.4. Spectrum of variants

The St. Jude registry had brought together by 2024 the data of 243 published patients and 62 referred patients, with 95 distinct variants in SAMD9 and 67 in SAMD9L in 305 individuals [2]. Most germline variants are missense (149 of 162, 92%) and cluster from the middle region to the C-terminus (from amino acid 600 onwards, in the shared P-loop NTPase domain), with an additional cluster in SAMD9 between amino acids 390 and 486; most are private to each family [2]. The variants recurrent in more than four patients, regarded as hotspots, are, in SAMD9, R459Q, I773T (NM_017654.4:c.2318T>C p.(Ile773Thr)), R982C, R982H (NM_017654.4:c.2945G>A p.(Arg982His)), E1136Q and R1293Q, and, in SAMD9L, S626L, H880Q, F886LfsX11, R986C, R986H, R1281K (NM_152703.5:c.3842G>A p.(Arg1281Lys)) and T1474P (published protein designation NP_689916.2:p.(Thr1474Pro)) (one-letter notation as in the source; the remainder are standardised in the following table and in the paragraph that follows it) [2, 8]. The frameshift variants of SAMD9L between amino acids 876 and 889 are specific to the autoinflammatory phenotype [2]. The following table presents the variants cited in this article as standardised in VariantValidator (GRCh38, MANE Select transcript).

Other variants cited in the sources only at protein level include, in SAAD, p.(Asn885Thrfs*6) and p.(Lys878Serfs*13) (published protein designations, without a uniquely mappable nucleotide) [42]; in SAMD9L familial MDS, p.(His880Gln) (NM_152703.5:c.2640C>A), p.(Arg986His) (NM_152703.5:c.2957G>A), p.(Arg986Cys) (NM_152703.5:c.2956C>T) and p.(Val1512Met) (NM_152703.5:c.4534G>A) [8, 31]; in M7MLS2, SAMD9 NM_017654.4:c.3406G>C p.(Glu1136Gln) and NM_017654.4:c.2026A>G p.(Lys676Glu) [3]; and, in ataxia-pancytopenia, SAMD9L NM_152703.5:c.2672T>C p.(Ile891Thr) [11]. The SAMD9L variant NM_152703.5:c.1216C>T p.(Arg406Ter) is found in about 1 in 350 healthy people and is expressed as a shorter protein, an example of a truncating variant without pathogenic effect [2].

2.5. Truncating variants: loss and gain of function

As the two genes have a single coding exon, truncating variants escape nonsense-mediated decay and produce a shortened protein, whose stability and function have to be assessed case by case [4]. Many N-terminal stop variants of SAMD9L are loss-of-function and arise as somatic rescue events, but some frameshifts in a narrow region of the NTPase-like domain act as gain of function and cause SAAD; the variant c.2849_2852dup, also truncating, causes severe aplastic anaemia without cutaneous or neurological abnormalities [4]. Allenspach and colleagues identified, by mutational scanning, the region of SAMD9L in which stop variants cause arrest of translation [41]. Outside SAAD, there is no evidence that germline truncating variants of SAMD9 or SAMD9L cause disease, and truncating variants of SAMD9 do not correspond to the expected mechanism [2].

2.6. Penetrance, expressivity and genotype-phenotype correlations

The de novo variants of MIRAGE and of SAAD have complete penetrance [2]. The penetrance of bone marrow failure and of MDS is incomplete: it was estimated at 65% (22 of 34 carriers) for SAMD9L in eight independent pedigrees, and in the series of Pastor and colleagues 3 of 10 carriers had no manifestations [2, 31]. Trottier and colleagues report, attributing this to earlier series, almost complete penetrance for early MDS with monosomy 7 in SAMD9 and about 70% in SAMD9L [48]. The SAMD9L GeneReview considers it impossible to estimate penetrance accurately, because of the variable expressivity, the episodic and asymptomatic cytopenias, the scarcity of neurological and imaging assessments and the effect of somatic rescue, but indicates that most carriers manifest some feature and that there are no differences between the sexes [4]. Completely unaffected carriers are rare, and non-penetrance or prolonged spontaneous remission is explained, in most documented cases, by haematopoietic clones with rescue [4, 48].

The MIRAGE GeneReview identified no genotype-phenotype correlations [1]. Rentas and colleagues, in an analysis of their patient and of 28 published MIRAGE patients, observed that the presence of MDS did not alter overall survival, but that the in silico pathogenicity scores of the variants distinguished the patients with the worse prognosis [49]. In SAMD9L-related syndrome, only the inflammatory subtype has a restricted genotypic spectrum (frameshift between amino acids 876 and 889), and not all frameshifts in that region cause SAAD [2, 4]. The overlap between the genes is partial: multisystem involvement is more frequent with SAMD9, progressive neurological disease characterises SAMD9L, but there are SAMD9 patients with marked neurological deficits and isolated SAMD9L patients with a MIRAGE pattern [2].

Rescue may occur very early, and even in the germline of a parent. Roucher-Boulez and colleagues described an asymptomatic mother carrying the SAMD9 variant NM_017654.4:c.2333C>T p.(Thr778Ile), which she transmitted to a son with MIRAGE; the mother also had a nonsense variant NM_017654.4:c.661C>T p.(Arg221*) in cis, probably acquired in utero and not transmitted, and the child acquired another nonsense variant, NM_017654.4:c.853C>T p.(Arg285*), at around the 20th day of life, with rapid improvement of the phenotype [46]. Shima and colleagues described two MIRAGE patients without haematological abnormalities, each with two variants on the same allele, a gain-of-function missense and a somatic nonsense restricted to blood, which acts as a second-site reversion [50]. In the French PFMG2025 cohort, a patient with a known pathogenic SAMD9 variant also had a de novo truncating variant with a similar allelic fraction (54%), which suggests a rescue already present in the germline and may explain the absence of severe systemic manifestations [47].

2.7. Allelic Disorders of SAMD9 and SAMD9L

Biallelic loss of function of SAMD9 causes NFTC, a disease clinically unrelated to MIRAGE [1, 7]. In five Yemenite Jewish families, Topaz and colleagues mapped the disease to a 7.6 Mb homozygous interval at 7q21–q21.3 and identified the variant NM_017654.4:c.4483A>G p.(Lys1495Glu) in homozygosity; dizygotic twins of the same origin were compound heterozygotes for p.(Lys1495Glu) and NM_017654.4:c.1030C>T p.(Arg344Ter) [7]. The disease is characterised by reddish or hyperpigmented skin lesions in the first year of life, followed by periarticular calcified nodules, mainly in the extremities, with normal calcium, phosphate, vitamin D metabolites and parathyroid hormone, and by severe conjunctivitis and gingivitis; it differs from the hyperphosphataemic form (GALNT3, FGF23) and resembles the dystrophic calcinosis that follows tissue injury or inflammation [7]. Germline loss-of-function variants of SAMD9 and SAMD9L, which increase proliferative capacity, have been associated with adult MDS [1, 51]. The SAMD9L GeneReview does not recognise other phenotypes for germline SAMD9L variants beyond the gain-of-function spectrum [4].

2.8. Alternative Mechanisms Proposed for SCA49

In the Menorca family, Corral-Juan and colleagues localised SAMD9L to mitochondria, observed reduced protein levels and mitochondrial changes in patients' fibroblasts, and impaired mobility and vestibular function in zebrafish with the p.(Ser626Leu) variant [33]. OMIM records that the authors proposed haploinsufficiency or a dominant-negative effect [13]. This interpretation diverges from the gain of function demonstrated in the other forms, and the same residue (S626L) is among the hotspots of the St. Jude registry, in patients with the usual haematological spectrum [2]. The sources consulted do not resolve the divergence; the GeneReview considers SCA49 part of the gain-of-function continuum [4].

3. Clinical Spectrum and Manifestations by System

3.1. Overview and Limits of the Data

The SAMD9 and SAMD9L syndromes share cytopenias of variable degree (from isolated thrombocytopenia to pancytopenia), recurrent infections, growth defects, typically hypocellular marrow, a risk of bone marrow failure and MDS, and monosomy 7; non-haematopoietic manifestations can affect almost every organ, with growth restriction and adrenal insufficiency typical of SAMD9 and progressive neurological deficits typical of SAMD9L [2]. Onset ranges from the fetal period, in MIRAGE, to adulthood, in SAMD9L neurology, but most patients present between infancy and adolescence [2]. Symptoms disappear and reappear without a predictable pattern, which confounds diagnosis and causes stress in families [2]. Published frequencies suffer from ascertainment bias: the initial series recruited patients through haematological or neurological presentations, early neurological manifestations are subtle, cytopenias fluctuate and follow-up of children is short [4]. Survival data mostly describe patients treated with supportive care and, in part, with HSCT, and not untreated natural history.

3.2. MIRAGE Syndrome

3.2.1. Frequency of the Core Features

The GeneReview, with the definition of four core features in 46,XY or three in 46,XX, brought together 44 patients with a SAMD9 variant published up to 2020 [1]. Yamada and colleagues updated the review in 2026 to 60 published patients, excluding fetal deaths [21]. The table compares the two series.

3.2.2. Pregnancy, Birth and Growth

Growth restriction begins in utero: typically, patients are born preterm by emergency caesarean section for fetal growth failure and suspected fetal distress [1]. The growth deficit persists after birth, with weight, length or height and head circumference usually below −2.0 standard deviations despite adequate caloric intake; two patients grew normally after birth, both with somatic reversion [1]. In the 22-year-old survivor described by Yamada and colleagues, height at 20 years was 138.3 cm (−5.6 standard deviations) and body mass index 12.0 kg/m², with normal growth hormone secretion and low insulin-like growth factor 1 attributed to malnutrition [21]. Trapani and colleagues documented growth hormone deficiency in an 11-year-old boy with MIRAGE and a history of HSCT, which broadens the endocrine phenotype [52]. The sources also describe congenital renal hypoplasia in a fetus with intrauterine growth restriction [53].

3.2.3. Haematopoietic System

The age at onset of haematological abnormalities varies, and the GeneReview distinguishes, on the basis of Rentas and colleagues, a group with severe cytopenias between birth and 2 years (about 60%) and another with later diagnosis in childhood (about 40%), with different prognoses [1]. Before clonal evolution, the marrow may be hypoplastic with a picture of congenital amegakaryocytic thrombocytopenia [1]. Thrombocytopenia is the most frequent manifestation, followed by anaemia and pancytopenia, and may be present at birth, transiently or chronically [1, 2]. Severe cytopenias occur in acute episodes associated with myelosuppressive infections or are chronic and require frequent transfusions, and may improve with age [1]. In Narumi's original series, all patients had thrombocytopenia and/or anaemia requiring transfusions in early childhood, which resolved spontaneously [9]. Progression to MDS is defined by acquisition of monosomy 7 or del(7q) with multilineage dysplasia; monosomy 7 may be transient if the clone is small or persist for years before transformation [1]. Rentas and colleagues followed a patient with untreated MDS for nine years, with progressive hypocellularity, erythroid and megakaryocytic dysplasia and monosomy 7; at 7 years several pre-leukaemic clones emerged and, through sequential mutations in ETV6 and RUNX1, AML developed at 9 years [49]. Nakata and colleagues described a SAMD9-associated MDS with monosomy 7 that, over 16 years and with somatic mutations in RUNX1 and PPM1D, progressed to AML with BCR::ABL1 and, after venetoclax and azacitidine, to B-cell acute lymphoblastic leukaemia with the same rearrangement [54].

3.2.4. Immunity and Infections

Most patients have recurrent bacterial infections from early childhood: pneumonia, urinary tract infection, gastroenteritis, meningitis, otitis media, dermatitis, subcutaneous abscess and sepsis, mainly due to enteric agents such as Klebsiella and Enterococcus, with an increased risk of complications from cytomegalovirus, respiratory agents and, in severe native immunodeficiency or after HSCT, fungi such as Candida [1]. In the St. Jude registry, 47 of 56 patients with syndromic SAMD9 (84%) had severe invasive infections, associated with high mortality [2]. The cause of the susceptibility is not established; contributing factors include hypogammaglobulinaemia, B, T and natural killer (NK) cell lymphopenia, thymic hypoplasia (described in four patients), incomplete and short-lived vaccine response (two patients), deficient T-cell responses and reduced neutrophil phagocytosis [1, 2]. In the 22-year-old survivor, immunophenotyping showed an inverted CD4/CD8 ratio (0.46), few recent thymic emigrants, reduced transitional B cells, class-switched memory B cells and plasmablasts, reduced NK cells, markedly reduced dendritic cells and immunoglobulin G of 682 mg/dL, despite three revertant variants acquired in the marrow [21]. In a pair of identical twins with MIRAGE, blood transcriptomic analysis showed a reduction of naïve and memory B cells [55]. Gao and colleagues reported a newborn girl diagnosed at 23 days by rapid trio exome, without immunological abnormalities or haematological malignancy at that age, and recommended close surveillance because of the possibility of their later emergence [56].

3.2.5. Adrenal and Other Endocrine Glands

About 80% of patients have primary adrenal insufficiency, which presents with diffuse skin hyperpigmentation at birth and may progress to severe dehydration and life-threatening hypotension; diagnosis rests on low cortisol and markedly raised corticotropin (ACTH), with aplastic or hypoplastic adrenals on ultrasound, and rarely the insufficiency appears only in late childhood [1]. In Narumi's series, hyperpigmentation led to suspicion before salt-wasting symptoms, and the autopsy of two patients showed very small, disorganised adrenals, with medulla only partly surrounded by cortex and foamy-looking cortical cells [9]. Endocrine complications go beyond the adrenal and gonads: thyroid dysfunction in nine published patients, abnormalities of glucose metabolism in seven, including two with autoantibody-negative insulin-treated diabetes, and primary hypogonadism with osteoporosis in the adult survivor [21].

3.2.6. Gonads and Genitalia

In 46,XY individuals, the external genitalia range from hypospadias, micropenis and shawl bifid scrotum to ambiguous genitalia and complete female genitalia; the testes are usually hypoplastic or dysgenetic, with undetectable testosterone in a 3-day-old newborn and testicular failure in a 16-year-old adolescent [1]. In Narumi's series, 6 of 7 46,XY patients had genital underdevelopment and one had female external genitalia and had been registered as female [9]; Li and colleagues described another 46,XY newborn with female external genitalia [57]. 46,XX individuals have no external genital anomalies, but four had hypoplastic or dysgenetic ovaries, with few primordial follicles at autopsy [1, 9].

3.2.7. Digestive Tract

Chronic diarrhoea occurs in about 80% of patients, usually appears after the start of enteral feeding and often causes severe nappy dermatitis; in Narumi's series, nine patients had chronic diarrhoea with colonic dilatation [1, 9]. About one third have oesophageal dysfunction with recurrent aspiration pneumonia: dysphagia, gastro-oesophageal reflux, vomiting, hypoperistalsis, oesophageal stricture and achalasia [1]. The 22-year-old survivor needed a nasogastric tube and, at 2 years, fundoplication and gastrostomy [21]. The enteropathy does not improve with HSCT [2].

3.2.8. Nervous System, Autonomic System and Sense Organs

Most patients have moderate to severe developmental delay; many do not acquire walking and have absent or limited language, and in Narumi's series 4 of the 8 who survived the first year had no head control and said no words [1, 9]. Neuropathology shows microcephaly, hydrocephalus, white matter changes and perivascular calcifications [1]. Garavatti and colleagues described, in three children, hypotonia, seizures, cerebral and cerebellar hypoplasia, lenticulostriate vasculopathy, ventriculomegaly, early deaths from non-infectious causes and one episode of malignant hyperthermia [58]. The 22-year-old survivor had a developmental quotient of 38 at 1 year, autism spectrum disorder, white matter and basal ganglia lesions with extensive calcifications and an asymptomatic occipital infarct at 5 years, seizures and lower limb weakness from 7 years, refractory migraines and transient extension of the lesions to the cortex after a seizure at 20 years; he mainly used a wheelchair [21]. Symptoms of autonomic dysfunction (hypolacrimation with corneal ulcer, hypohidrosis or anhidrosis with temperature instability, or hyperhidrosis) were described in seven patients [1].

3.2.9. Kidney, Lung, Heart and Other Organs

Renal dysfunction, with proteinuria and renal tubular acidosis, was described in 5 of the 17 patients who lived more than 3 years; of the four biopsies, two showed focal segmental glomerulosclerosis, one interstitial nephritis and one C1q nephropathy [1]. The 2026 review counted 11 patients with renal complications, six with biopsy (focal segmental glomerulosclerosis in four, interstitial nephritis and C1q nephropathy in the others), congenital renal hypoplasia in two and severe renal failure after HSCT in one [21]. The 22-year-old survivor progressed from proteinuria at 6 years, treated with an angiotensin receptor blocker, to end-stage kidney disease at 20 and received a kidney from his father at 21, the first kidney transplant described in the syndrome [21]. Ishimori and colleagues described nephrotic syndrome with a pattern of membranous nephropathy and mesangial proliferation in an infant who died at 21 months of multi-organ failure [59]. The St. Jude registry includes in the SAMD9 genotype valvulopathies, cardiomyopathy, arrhythmias, ventricular septal defect, interstitial pulmonary inflammation and bronchopulmonary dysplasia [2], and a lung transplant for MIRAGE-associated lung disease has been published [60]. Orphanet adds dysmorphic features such as downslanting palpebral fissures, low-set and rotated ears, anteverted nares, camptodactyly and arachnodactyly [15]; the adult survivor had hypertelorism, a flat nasal bridge, esotropia, scoliosis, tapering fingers, hallux valgus and metaphyseal dysplasia [21].

3.2.10. Survival

The median age at death in the 44 GeneReview patients was 3 years, almost 60% of deaths were due to infections and the oldest patient was alive at 20 years [1]. In the first two series of syndromic SAMD9, only 5 of 19 patients were alive at last follow-up, most died in infancy, and the two transplanted patients survived [2]. In Narumi's series, six patients died before 2 years, mainly of invasive infection, and the two with monosomy 7 died of MDS [9]. The Japanese survivor is now 22 years old [21]. These figures come from published series biased towards severe cases and are not population estimates.

3.3. SAMD9 Patients without the MIRAGE Phenotype

Carriers of gain-of-function variants in SAMD9 may present with myelodysplasia alone and few or no MIRAGE features [1]. Schwartz and colleagues described three siblings, aged 14 months to 4 years, with refractory cytopenia and monosomy 7, the two eldest with neutropenia, macrocytosis, thrombocytopenia, hypocellularity and trilineage dysplasia; one had hypospadias and bifid scrotum and none had MIRAGE; the variant NM_017654.4:c.3406G>C p.(Glu1136Gln) had been inherited from a clinically unaffected mother [3, 30]. Wong and colleagues described two siblings with AML and MDS at 4 and 2.5 years, with loss of the paternal chromosome 7, of whom one is alive at 33 years after HSCT and the other died during induction [3, 32]. In the EWOG-MDS cohort, SAMD9 and SAMD9L patients had similar 5-year overall survival (84% and 93%) [2].

3.4. SAMD9L-Related Syndrome

3.4.1. Haematopoietic System

More than 150 individuals have been described, some in multigeneration families, with marked inter- and intrafamilial variability in age at onset, severity and rate of progression; the penetrance of haematological and neurological manifestations is not concordant, even within the same family, partly because of somatic rescue [4]. Most people with manifestations present with haematological abnormalities in infancy and childhood: single-lineage cytopenias, hypocellular bone marrow failure with pancytopenia mimicking aplastic anaemia, and myelodysplastic neoplasm [4]. In the series of Pastor and colleagues, seven patients from four lineages had MDS with loss of 7/7q, median age 2.1 years (1 to 42), thrombocytopenia in all and hypocellular marrow without excess blasts; two progressed to leukaemia or accumulated driver mutations, four retained monosomy 7 and two recovered spontaneously with wild-type UPD7q, and none had dysmorphic features or neurological symptoms [31]. The marrow frequently reveals rescue events, even with a normal blood count, and several clones may arise, expand, contract or disappear over time [4]. Red cells may be macrocytic and mimic vitamin B12 or folate deficiency [4]. MDS and leukaemia are diagnosed mainly in early childhood and are rare in adults; the latest onset of MDS described was at 42 years [4]. In the 2018 compilation of the four families then published, 18 of the 25 carriers had had transient or permanent cytopenias, eight of them with recurrent or severe infections or bleeding, and five developed haematological neoplasms at 1.5, 4, 7, 10 and 56 years (the type of neoplasm at 56 years is not specified); all variants were missense and none recurred between families up to that date [34].

3.4.2. Immunity

Immunodeficiency was observed in about a quarter of patients in a retrospective review, probably an underestimate because many were not assessed; the most consistent abnormalities are B-cell and NK-cell deficiency [4]. Tesi and colleagues observed that three carriers developed MDS after loss of the mutated allele following intracellular infections, with myeloid, B and NK deficiency [29]. In young children, infections often precede cytopenia and monosomy 7 [2]. The French case of Dellal and colleagues presented at 5 months as an inborn error of immunity, with Pneumocystis jirovecii pneumonia, systemic cytomegalovirus infection, B and NK lymphopenia and T-cell dysfunction with normal immunoglobulins, before pancytopenia appeared [47].

3.4.3. Nervous System

Neurological involvement was found in most adult carriers examined carefully; children ascertained through haematological disease were not routinely examined neurologically [4]. Neurological onset ranges from infancy to 62 years [4]. Horizontal and vertical nystagmus is present in most adults and in some children, and brisk tendon reflexes are, with nystagmus, the earliest neurological signs; ankle clonus is easily elicited in some and there are extensor plantar responses [4]. Cerebellar gait ataxia appears later and progresses slowly; some patients end up in a wheelchair [4]. The French protocol adds to the spectrum a pseudo-parkinsonian syndrome which, like ataxia, may appear only in adulthood or old age, and also mentions behavioural disorders, mild dysmorphism and almost constant cerebellar atrophy [23]. There is also sensorimotor polyneuropathy (axonal or demyelinating), dysautonomia and paracentral retinal dysfunction, with complaints of difficulty reading and focusing [4]. In the UW-AP family, symptoms began between 6 and 62 years; one sibling with pancytopenia at 3 months had balance problems only at 30 years, and another, with thrombocytopenia from the first year, died at 16 of retroperitoneal haemorrhage without clinical ataxia but with cerebellar atrophy and severe loss of Purkinje cells at autopsy [11]. Among the 29 children with SAMD9L myelodysplasia in the EWOG series, one had ataxia and cerebellar hypoplasia, two arachnoid cysts, one demyelinating polyneuropathy, three macrocephaly and one microcephaly [4]. Tesi and colleagues described a child who, after successful HSCT for MDS with monosomy 7, developed nystagmus, unsteady gait, rigidity and lower limb weakness, with white matter changes and progressive cerebellar degeneration [11]. Thunström and Axelsson reported a patient with leukoencephalopathy, demyelinating peripheral polyneuropathy, dural ectasia and multiple cerebral cysts, with a de novo variant NP_689916.2:p.(Phe896Val) (published protein designation) and a picture dominated by neurology [61]. In the same 2018 compilation, all 25 carriers, except one aged 4 and another aged 16, had gait abnormality, nystagmus or both; the other cerebellar signs (dysdiadochokinesia, dysmetria, tremor) were less marked than the imbalance or absent, which makes the term “ataxia” debatable, and the authors accept that part of the dysmetria described may result from visual control limited by paracentral retinal dysfunction, which they identified in two carriers and which was suggested by the complaints of others [34]. Some carriers of specific variants had behavioural changes, mainly in childhood and adolescence [34].

Brain imaging may show cerebellar atrophy and periventricular and deep white matter changes and cysts before clinical signs [4]. Magnetic resonance imaging (MRI) is more sensitive than CT for early white matter changes, and CT is more sensitive for small basal ganglia calcifications [4]. At autopsy of four individuals aged 7 to 16 years from two families there was marked cerebellar atrophy with loss of Purkinje cells, and in two of them moderate neuronal loss in the inferior olives and central nuclei [4]. In a family studied by Chen and colleagues, a 32-year-old man with severe ataxia had marked midline cerebellar degeneration, and his 38-year-old sister, with mild signs, moderate atrophy; the surviving brother in the Li family needed a walker at 54 years and MRI at 52 showed severe diffuse cerebellar atrophy and diffuse white matter hyperintensities [4]. Dural ectasia and arachnoid cysts have also been described [4]. In 2026, Alm and colleagues proposed dentate nucleus hyperintensities as a diagnostic clue in SAMD9L ataxia [62]. Imaging changes do not reflect clinical severity or predict progression [4]. Cerebellar atrophy begins in early childhood, progresses slowly and is disproportionate to the imbalance, which is generally mild; the neuropathological correlate of the white matter hyperintensities is unknown, because autopsies were performed at ages at which these hyperintensities are not usually yet present, and the authors hypothesise a delay in the physiological myelination of childhood and adolescence [34].

3.4.4. Spinocerebellar Ataxia Type 49

In the Menorca family (11 affected in five generations, nine described in detail, all adults), the first signs, between 12 and 60 years, were unsteady gait, horizontal and vertical gaze-evoked nystagmus and hyperreflexia; gait ataxia developed between 30 and 60 years, followed by dysarthria, dysmetria, diplopia, oscillopsia and strabismus, and in some dysdiadochokinesia, pyramidal signs, clonus and pes cavus [13, 33]. Five patients had moderate axonal sensory polyneuropathy, first in the lower and then in the upper limbs, with abnormal sympathetic skin response; two became wheelchair users in the eighth decade; all five with imaging had cerebellar atrophy and diffuse cerebral demyelination [13, 33]. No patient had seizures, cognitive impairment, extrapyramidal signs or haematological abnormalities [13]. Garg and colleagues described an Indian patient with SCA49, ataxia, early-onset dystonia and bradykinesia [35].

3.4.5. Other Organs

The initial evaluation recommended by the GeneReview includes various congenital findings (micro- and macrocephaly, genitourinary anomalies), and obstetricians should consider that macrocephaly, hydrocephalus and intrauterine growth restriction may influence delivery [4]. The MIRAGE pattern has been described in isolated SAMD9L patients [2]. In the French PFMG2025 cohort, a 60-year-old adult with hereditary spastic paraparesis and the SAMD9L variant NP_689916.2:p.(Thr936Ile) (published protein designation) had no haematological abnormalities, which may reflect reduced penetrance, an atypical neurological presentation or an incidental finding [47].

3.4.6. Survival

Increased mortality is due to haematopoietic and immunological disease [4]. In one of the first SAMD9L series, 9 of 10 patients were alive with a median follow-up of 31 years [2]. Of the 41 published patients who received HSCT for myelodysplasia, cytopenias, mosaic monosomy 7 or autoinflammation, 36 survived the procedure, with follow-up of up to 16 years; haematological reconstitution does not improve the neurological or inflammatory manifestations [4].

3.5. SAMD9L-Associated Autoinflammatory Disease

De Jesus and colleagues identified de novo frameshift variants in SAMD9L in six patients, among 66 with undifferentiated systemic autoinflammatory disease studied at the NIH; all had been diagnosed with CANDLE (chronic atypical neutrophilic dermatosis with lipodystrophy and elevated temperature) because of clinically and histologically indistinguishable neutrophilic panniculitis, but four also had severe early-onset interstitial lung disease, which does not occur in CANDLE, and all developed progressive B and NK cytopenias; one patient died of respiratory failure at 2 months [5]. Patients with a high interferon-response gene score had, compared with those with a negative score, more neutrophilic panniculitis (40% versus 0%), basal ganglia calcifications (46% versus 0%), interstitial lung disease (47% versus 5%) and myositis (60% versus 10%); in SAAD the score was lower than in CANDLE, with raised C-reactive protein, and the pattern suggested NF-κB activation [5]. Allenspach and colleagues described an infant with B-cell aplasia and autoinflammatory features who died of respiratory failure with chronic rhinovirus infection; autopsy showed absence of B cells in marrow and periphery and selective loss of Langerhans and Purkinje cells [41]. Russell and colleagues described two children with leukocytoclastic panniculitis, basal ganglia calcifications, raised inflammatory markers, neutrophilia, anaemia, thrombocytopenia and almost no B cells [42].

The GeneReview summarises SAAD as B-cell aplasia with hypogammaglobulinaemia, severe skin inflammation (neutrophilic panniculitis, with selective loss of Langerhans cells on biopsy), bone marrow failure and basal ganglia calcifications, sometimes with hepatosplenomegaly, interstitial lung disease, neutrophilic tissue infiltration and non-infectious epiglottitis [4]. Skin inflammation, with erythematous papules, pustules or plaques that may ulcerate and scar (Sweet syndrome-like) and CANDLE-like leukocytoclastic or nodular panniculitis, is also frequent in other SAMD9L patients [4]. Orphanet defines SAAD by nodular panniculitis, lipoatrophy, severe early interstitial lung disease and basal ganglia calcifications, with respiratory failure described [6]. SAAD confers a risk of severe viral infections [2]. Of the 12 patients published up to 2024, four of five transplanted early had a good outcome, two responded partially to anti-inflammatory drugs without HSCT, one recovered spontaneously and three died of progressive disease with recurrent infections [2]. In 2026, thrombotic microangiopathy was described in an adult with SAAD and a novel SAMD9L variant [63].

4. Clinical Recognition and Warning Signs

4.1. When to Suspect MIRAGE

MIRAGE should be suspected in the newborn or infant with intrauterine growth restriction and prematurity, diffuse skin hyperpigmentation, dehydration and hypotension due to primary adrenal insufficiency, atypical genitalia in a 46,XY individual, thrombocytopenia or other cytopenias, recurrent bacterial infections, intractable chronic diarrhoea, dysphagia and aspiration pneumonia [1]. In the laboratory, suggestive signs are single- or multilineage cytopenia, transient or persistent; hypocellular or aplastic marrow without megakaryocytes, or MDS/AML with monosomy 7; and primary adrenal insufficiency with hyponatraemia, hyperkalaemia, hypoglycaemia, low cortisol and markedly raised ACTH [1]. On imaging, the adrenals appear aplastic or hypoplastic on ultrasound and brain MRI may show microcephaly, hydrocephalus and white matter changes [1]. Li and colleagues propose considering MIRAGE in every newborn with unexplained prematurity and growth restriction, thrombocytopenia, recurrent infections and a 46,XY karyotype with female or ambiguous genitalia [57]. The initial picture overlaps with that of IMAGe syndrome; recurrent infection, chronic diarrhoea and haematological abnormalities appear with age and point towards MIRAGE [64].

4.2. When to Suspect SAMD9L-Related Syndrome

The GeneReview recommends suspecting it in the presence of one or more of the following: autoinflammatory disease of neonatal or infantile onset, especially with B-cell aplasia or hypogammaglobulinaemia; hypocellular bone marrow failure in childhood; MDS in childhood or young adulthood, particularly with monosomy 7, del(7q) or copy-neutral UPD 7q; nystagmus and brisk tendon reflexes, the earliest neurological signs, with cerebellar ataxia later; cerebellar atrophy, white matter changes or basal ganglia calcification on brain imaging; and a family history compatible with autosomal dominant transmission, although the absence of a family history does not exclude the diagnosis [4]. The combination of mild cytopenias and ataxia in the same patient or in different members of a family is the classical pattern [4]. Neurology may be the only manifestation, and becomes evident mainly when multigeneration families are examined [2].

4.3. When to Suspect It in the Haematological Setting

In the child or young adult with MDS, especially with monosomy 7, the SAMD9 and SAMD9L syndromes are among the first hypotheses: they account for 8% of consecutive paediatric MDS and, among paediatric MDS with monosomy 7, 21% of the total, 33% of refractory cytopenias and 6% of MDS with excess blasts [2, 22]. About half of paediatric MDS with monosomy 7 have a germline predisposition, mainly GATA2 or SAMD9/SAMD9L [43]. The SAMD9/SAMD9L syndromes are mutually exclusive with GATA2 deficiency in paediatric cohorts and are associated more with hypocellular refractory cytopenia (90%) than with high-risk MDS, and the typical presentation is that of a preschool child with cytopenias after an infectious illness [22, 43]. Monosomy 7 that disappears spontaneously is a highly specific sign [2]. Davidsson and colleagues proposed, in 2018, testing in non-haematopoietic DNA all children with MDS with monosomy 7, del(7q) or der(1;7), and probably also adults with these abnormalities, with panels including GATA2, SAMD9 and SAMD9L; monocytopenia and NK-cell deficiency reinforce the suspicion, but should not be a condition for testing, because rescue often normalises the counts, and the inclusion of a neurological examination focused on balance and eye movements in the evaluation of cytopenias or MDS may bring forward the suspicion of SAMD9L [34]. The subsequent analysis of 799 adult MDS without gain-of-function variants limits the yield of that test in adults [2]. In paediatric severe aplastic anaemia without clinical or laboratory features of inherited failure, the probability of a hidden inherited syndrome is low: among 113 children without these features none had a diagnostic variant, and among 36 with at least one suggestive feature one had a pathogenic SAMD9L variant [65]. In adults, haematological manifestations are rare (nine published patients, median 46 years, six with bone marrow failure, two with MDS and one with AML) and, in an analysis of 799 adult MDS, no gain-of-function variants were found [2]. In a child with bone marrow aplasia, the French protocol for bone marrow aplasias lists the elements that point towards a SAMD9/SAMD9L syndrome: suggestive family history; age under 5 years, although presentation may be later; absence of neonatal thrombocytopenia (which would point to congenital amegakaryocytosis), of a severe Hoyeraal-Hreidarsson-type picture (telomere biology disorder) and of MECOM syndrome, the three other constitutional bone marrow failures of the young child; relatively acute onset, sometimes after an infection or a vaccination, whereas gradual onset is what usually suggests a constitutional form; monosomy 7 present from diagnosis; and rapid, spontaneous improvement of the blood count [23]. Acute onset may mimic acquired aplasia, and the distinction has immediate consequences: constitutional bone marrow failures have no indication for immunosuppressive treatment, and the absence of a systematic indication for transplantation in the SAMD9/SAMD9L syndromes justifies, according to the same protocol, early genetic diagnosis [23]. Monosomy 7 in myeloid neoplasms considered therapy-related should also prompt investigation of predisposition, because up to 20% of these neoplasms probably have a germline basis [20].

4.4. Clues in the History, Examination and Blood Count

The monosomy 7 GeneReview proposes a targeted history that includes results of previous blood counts and bone marrow aspirates, macrocytosis, atypical infections, intrauterine growth restriction, malformations, adrenal insufficiency, short stature, neurological and developmental problems, environmental exposures and a three-generation family history looking for deaths in early childhood, miscarriages, early cancers or excessive treatment toxicity, consanguinity, transient neonatal cytopenias, bone marrow failure, immunodeficiency, lymphoedema, atypical bleeding and severe developmental delay [20]. On examination, short stature points to MIRAGE, Fanconi or Shwachman-Diamond; cerebellar ataxia to SAMD9L; and lymphoedema or pulmonary changes to GATA2 [20]. Carrier relatives may have a normal blood count or subtle changes, such as macrocytosis (mean corpuscular volume above 94 fL), raised fetal haemoglobin and mild cytopenias [20].

4.5. Warning and Emergency Signs

In MIRAGE, dehydration, hypotension, hyponatraemia, hyperkalaemia and hypoglycaemia in a hyperpigmented infant indicate adrenal crisis, a life-threatening emergency [1]. Fever in a patient with severe neutropenia, hypogammaglobulinaemia or cellular immunodeficiency requires immediate assessment, because invasive infections are the main cause of death in MIRAGE and also drive mortality in SAAD [1, 2]. In all forms, a new or progressive cytopenia, blasts in peripheral blood, the emergence of monosomy 7 or of leukaemia driver mutations, transfusion dependence and severe bleeding require haematological reassessment with bone marrow aspirate, because progression to advanced MDS or AML may occur within months of monosomy 7 [20, 66]. Families should be taught to seek medical attention in the presence of fatigue, pallor, unexpected bleeding or recurrent infections [4]. In SAAD, respiratory difficulty is a sign of severity, because early interstitial lung disease is associated with high mortality [5].

4.6. Newborn Screening

There is no newborn screening targeted at these syndromes in the sources consulted. Neonatal diagnosis is made by rapid exome or genome sequencing in newborns with the picture described [56].

5. Diagnosis: Criteria, Tests and Algorithms

5.1. Diagnostic Criteria

There are no formal clinical criteria for any of the syndromes [1, 4]. The diagnosis of MIRAGE is established in a proband with suggestive signs and a heterozygous germline gain-of-function variant in SAMD9 [1]. The diagnosis of SAMD9L-related syndrome is established in a proband with a pathogenic or likely pathogenic heterozygous germline gain-of-function variant in SAMD9L [4]. In both cases, a variant of uncertain significance (VUS) neither establishes nor excludes the diagnosis [1, 4]. The operational definition of MIRAGE used in the reviews (at least four of the six core features in 46,XY or unknown karyotype; at least three in 46,XX) serves to compare series and does not replace molecular confirmation [1, 21]. Classification of the associated neoplasm follows the 2022 WHO or the 2022 ICC, with the germline predisposition as a qualifier [36, 37].

5.2. Molecular Testing Strategies

When the phenotype is suggestive, sequencing of SAMD9 or SAMD9L or a multigene panel for bone marrow failure, hereditary leukaemia, adrenal hypoplasia or cerebellar ataxia that includes the gene may be performed; some adrenal hypoplasia or myelodysplasia panels do not include SAMD9 [1, 4]. When the phenotype does not distinguish between the many diseases with growth restriction, myelodysplasia, adrenal hypoplasia, ataxia or cytopenias, exome sequencing is the most used option, and the genome is an alternative [1, 4]. All pathogenic SAMD9L variants described are coding and detectable on the exome, provided that somatic reversion or loss of part of chromosome 7 has not occurred in the tissue analysed [4]. Sequencing detects 44 of 44 MIRAGE probands and all SAMD9L variants described; analysis of intragenic deletions and duplications is not indicated, because a deletion would cause loss of function and not the gain-of-function diseases [1, 4]. These proportions refer to variants detectable by the method in patients already diagnosed and are not equivalent to the clinical sensitivity of the test in a patient with suspected disease. The French panel for suspected constitutional bone marrow failure, presented as an open list, includes SAMD9 and SAMD9L alongside ANKRD26, RUNX1, ATG2B, ATM, ATR, ATRX, BRCA1, BRCA2 and the other Fanconi anaemia genes, DDX41, ERCC6L2, GATA2, LIG4, DNAJC21, MECOM, MPL, THPO, SBDS, SRP54, SRP72 and the telomere biology disorder genes (DKC1, TERC, TERT, TINF2, RTEL1, CTC1, PARN, USB1, NOP10, NHP2 and others) [23].

5.3. Choice of Tissue: the Problem of Somatic Rescue

The central technical limitation of these syndromes is that haematopoietic tissue tends to lose the germline variant. Monosomy 7, del(7q), UPD7q and revertant variants reduce the variant allele fraction in blood, marrow, buccal mucosa or saliva, which may prevent detection or lead to its misclassification as somatic [1, 4]. In a comparison of 65 patients, the mean allele fraction of the germline variant was 33% in the 37 with monosomy 7 and 50% in the 28 with a normal karyotype (P < 0.00001) [2]. Therefore, skin (considered the reference standard) or buccal fibroblasts, or hair follicles, are the preferred tissues for germline testing; when a therapeutic decision is urgent, testing may begin in blood or marrow and be confirmed later in cultured fibroblasts, which take longer [2, 20]. Davidsson and colleagues observed that buccal swabs sometimes gave allele fractions skewed towards the wild-type allele, probably through contamination with blood cells rescued by UPD7q, and prefer skin fibroblasts [34]. St Martin and colleagues described in 2026 rapid germline testing with DNA from hair root bulbs in hereditary predisposition and bone marrow failure syndromes [67]. If suspicion is high and the test negative, a search for genomic changes by single nucleotide polymorphism microarray (SNP array) and for low-fraction variants by deep sequencing (more than 1000 reads) and testing in a non-haematopoietic tissue are recommended [1, 4]. UPD7q is detected only by SNP array or by genome-wide polymorphism analysis, and not by array comparative genomic hybridisation; a low allele fraction without monosomy 7 suggests UPD7q [2]. The case of Dellal and colleagues shows the usefulness of trio genome sequencing with a mosaic disomy detection tool, which revealed mosaic segmental UPD7q in the asymptomatic mother and in the patient [47].

5.4. Cytogenetics and Bone Marrow Study

Monosomy 7 is detected by G-banded karyotype or by deletion/duplication methods (microarray, fluorescence in situ hybridisation, or FISH, and quantitative polymerase chain reaction), in unstimulated samples, because phytohaemagglutinin and other mitogens may mask monosomic cells; at least 20 metaphases are analysed, and at least 3 in 20 cells without chromosome 7 confirm monosomy [20]. Complete loss of chromosome 7 is the most frequent, but del(7q), der(1;7)(q10;p10) and other unbalanced translocations with loss of 7q belong to the same spectrum [20]. A normal karyotype at onset does not exclude future monosomy 7, and corticosteroids may mask it on the karyotype; FISH or microarray are preferable for longitudinal monitoring of the clone percentage [20]. The initial bone marrow study includes aspirate and biopsy, cellularity and dysplasia, blast count by flow cytometry or immunohistochemistry, reticulin and iron staining, karyotype, FISH for chromosome 7 and other chromosomes (1, 3, 5, 8, 17, 20) and testing for somatic variants in SETBP1, ASXL1, RUNX1 and the RAS pathway (CBL, KRAS, NRAS, PTPN11) [20]. The SAMD9L GeneReview recommends morphology, flow cytometry, karyotype, interphase FISH for 7q and a next-generation sequencing panel of myeloid genes including SAMD9L, and SNP array in blood or marrow for copy-neutral loss of heterozygosity [4]. In children, low-grade MDS with hypocellular marrow is frequent in these syndromes: in the EWOG-MDS cohort, 38 of 42 patients (90%) had refractory cytopenia of childhood, 4 of 42 (10%) MDS with excess blasts, all had dysplasia and 32 of 42 (76%) hypocellular marrow [2]. The ICC retains refractory cytopenia of childhood (more than 80% hypocellular, requires biopsy) and recognises that it may represent progression of a predisposition such as SAMD9/SAMD9L [37]. In the French pathway, the bone marrow aspirate shows poor or hypoplastic marrow, sometimes with the dysmegakaryopoiesis typical of monosomy 7; trephine biopsy confirms the poverty of the marrow, but is not systematic in constitutional forms, especially in young children, in whom it requires general anaesthesia; monosomy 7 is sought on the karyotype and also by interphase FISH (chromosome 7 and 8 probes, useful when the karyotype fails in a poor marrow); and the detection of other clonal abnormalities should be followed by a myeloid sequencing panel to estimate the risk of transformation [23]. As in other constitutional bone marrow failures, fetal haemoglobin may be raised at diagnosis and should be measured, if possible, before the first transfusion [23]. Apparent dysplasia may exist at baseline evaluation of carriers without MDS, which requires review in an experienced centre [66].

5.5. Variant Interpretation

Interpretation of SAMD9 and SAMD9L variants is difficult: they are mainly gain-of-function missense variants with no effect on expression, the two genes are poorly conserved and conservation-based CADD or REVEL scores have little value; the ACMG/AMP (American College of Medical Genetics and Genomics/Association for Molecular Pathology) criteria may be misleading, because acquired truncating variants, which would be considered deleterious, have a rescue effect; and there are no standardised functional assays for clinical use [2]. With the usual population threshold of 0.1% in gnomAD, about 70% of the patients' variants in the registry were VUS; the authors provisionally propose not to consider as causal variants with a frequency above 0.005% in gnomAD [2]. The table summarises the criteria proposed by the St. Jude group.

Source: [2].

The detection of UPD7q or of a somatic SAMD9/SAMD9L variant in a patient with a VUS works as a “natural functional assay”, because these events are not found in healthy people or in other syndromes (with the anecdotal exception of UPD7q in Shwachman-Diamond syndrome), and allows the variant to be reclassified as likely pathogenic [2, 43]. This is how Dellal and colleagues reclassified the SAMD9L variant NP_689916.2:p.(Ser1550Pro) (published protein designation), initially a VUS, in a patient with transient pancytopenia and progressive neurology, after detecting mosaic UPD7q and a somatic truncating variant [47]. Functional assays in cell models, such as heterozygous CRISPR editing in myeloid lines with measurement of protein translation, are beginning to support the classification of novel variants [68]. Rare SAMD9 and SAMD9L variants are frequently found in large-scale sequencing programmes: in the French PFMG2025 plan (11,913 genomes from 60 rare disease categories), after excluding frequencies above 0.005%, there were 33 rare variants in SAMD9 and 29 in SAMD9L, almost all missense VUS, and only a minority had a convincing genotype-phenotype correlation [47]. An ultra-rare SAMD9 or SAMD9L VUS in a patient with isolated cytopenia should not be automatically discarded, given the incomplete penetrance, but its interpretation requires integrating the clinical picture, segregation, tissue and somatic genetics [2].

5.6. Report Content

The report should state the gene, the transcript with version (NM_017654.4 or NM_152703.5), the variant in HGVS nomenclature at the DNA and protein level, zygosity, the tissue analysed, the allele fraction, the ACMG/AMP class and, when the tissue is haematopoietic, the caveat that a reduced allele fraction may reflect somatic rescue and not constitutional mosaicism [2, 4]. When somatic data are available (karyotype, SNP array, myeloid panel), the integrated report should distinguish maladaptive events (monosomy 7, driver mutations) from adaptive ones (UPD7q, compensatory variants) [2, 66].

5.7. Complementary Evaluation after Diagnosis

In MIRAGE, the initial evaluation includes a blood count with differential and bone marrow examination, lymphocyte subsets, immunoglobulin G and vaccine responses, growth charts, sodium, potassium, glucose, cortisol and ACTH with adrenal ultrasound, genital examination and karyotype (mandatory even with female genitalia), gastroenterological and nutritional assessment of aspiration risk, developmental assessment, autonomic and ophthalmological assessment (hypolacrimation, keratoconjunctivitis sicca, corneal ulcer, dyshidrosis, temperature instability) and renal function with creatinine, urea and urinalysis [1]. In SAMD9L-related syndrome, it includes a full physical examination, blood count, T, B and NK lymphocyte subsets, immunoglobulins, vaccine responses (tetanus, diphtheria, pneumococcus), bone marrow examination with the studies mentioned (which may be unnecessary in older children or adults identified by cascade testing or through neurology), SNP array, early human leukocyte antigen (HLA) typing and germline genotyping of at-risk relatives to define donors, neurological examination with an ataxia scale (SARA, Scale for the Assessment and Rating of Ataxia; ICARS, International Cooperative Ataxia Rating Scale; or BARS, Brief Ataxia Rating Scale), MRI with and without contrast, non-contrast CT for calcifications, electromyography and conduction studies if there are sensory or motor changes, and ophthalmological assessment of retinal function if there are visual complaints [4]. The 2018 Nordic proposal, common to SAMD9 and SAMD9L, adds sequencing in blood to measure the variant allele fraction and detect revertant variants, bone marrow biopsy with karyotype, 7q FISH and myeloid gene panel, SNP array in blood or marrow to estimate the extent of UPD7, HLA typing, chest radiograph looking for alveolar proteinosis (because of monocyte and macrophage deficiency), baseline brain MRI with T2 and FLAIR (fluid-attenuated inversion recovery) sequences in SAMD9L carriers and attention to human papillomavirus (HPV) vaccination, because of NK deficiency; the authors acknowledge that the radiograph and the attention to HPV are based on the experience of a very small cohort [34]. In SAAD, the interferon-response gene score and the cytokine profile support the differential diagnosis from other interferonopathies [5].

The French protocol sets, for the SAMD9/SAMD9L group, a minimum organ evaluation at diagnosis, guided by the clinical picture: abdominopelvic ultrasound (looking for adrenal hypoplasia, with low diagnostic value), echocardiogram (rare cardiac malformations have been described, without proof that they belong to the phenotype), endocrine assessment with testing for adrenal insufficiency as soon as the diagnosis is suspected, because of the risk of acute adrenal insufficiency, neurological and neurosensory assessment with brain MRI, and psychological assessment [23]. The initial evaluation form for any bone marrow aplasia in the same protocol adds urgent HLA typing of the patient and siblings, T, B and NK lymphocyte phenotyping, measurement of IgG, IgA and IgM, testing for a paroxysmal nocturnal haemoglobinuria clone, haemoglobin electrophoresis or fetal haemoglobin measurement before transfusion, alpha-fetoprotein, abdominal ultrasound (especially renal), echocardiogram and chest radiograph, and, depending on the context, chromosome breakage testing with mitomycin C, telomere length and sequencing in blood and fibroblasts [23]. Surgery needed for malformations should, when possible, be performed before thrombocytopenia worsens [23].

5.8. Practical Algorithm

In the face of clinical suspicion, the sequence proposed by the sources is: (1) targeted clinical evaluation and three-generation family history [20]; (2) blood count with reticulocytes and smear, and bone marrow aspirate with biopsy, cytogenetics, chromosome 7 FISH, SNP array and somatic panel [4, 20, 66]; (3) germline testing of SAMD9 and SAMD9L, alone, in a panel or by exome/genome, ideally in fibroblasts or hair roots, or in blood with later confirmation if urgent [2, 20]; (4) if negative and suspicion persists, deep sequencing, SNP array and testing in non-haematopoietic tissue [1, 4]; (5) if a VUS is found, search for UPD7q and somatic SAMD9/SAMD9L variants, and study of segregation and de novo occurrence [2]; (6) exclusion of alternative diagnoses, including chromosome breakage with mitomycin C or diepoxybutane (Fanconi anaemia) and telomere length [20]; and (7) organ evaluation according to the phenotype [1, 4].

5.9. Prenatal and Preimplantation Diagnosis

Once the germline variant has been identified in an affected relative, prenatal diagnosis and preimplantation genetic testing are possible [1, 4]. Monosomy 7 is not expected in most sampled fetal tissues [20]. Prenatally, severe intrauterine growth restriction with renal hypoplasia may be the first manifestation of MIRAGE [53].

6. Differential Diagnosis

6.1. Comparative Table

6.2. MIRAGE and the Other Causes of Neonatal Adrenal Insufficiency

The neonatal presentation of MIRAGE (growth restriction, adrenal hypoplasia and genital anomalies) is initially similar to that of IMAGe syndrome, and recurrent infections, chronic diarrhoea and haematological abnormalities only appear with age; in IMAGe, metaphyseal dysplasia is absent in the neonatal period and appears in early childhood, and hypercalcaemia or hypercalciuria may cause nephrocalcinosis [64]. MIRAGE is distinguished from POLE-related IMAGe-I by the haematological abnormalities and diarrhoea, and from triple A syndrome by all the extra-adrenal features except achalasia and hypolacrimation, which occur in both [1]. In the Japanese MIRAGE survivor, a targeted adrenal insufficiency panel (STAR, HSD3B2, NR0B1, NR5A1, AAAS and CDKN1C) was negative, and the diagnosis was made only at 12 years [21]; some adrenal hypoplasia panels still do not include SAMD9 [1]. Conversely, the MIRAGE pattern has been described in patients with SAMD9L variants [2].

6.3. SAMD9/SAMD9L and GATA2 Deficiency

The two are the most frequent germline causes of paediatric MDS with monosomy 7, together about 15% of primary paediatric MDS, and are mutually exclusive [22, 43]. GATA2 deficiency is associated more with advanced MDS, with a risk that increases with age and is rare before 6 years, whereas the risk of MDS in the SAMD9/SAMD9L syndromes is highest in young children and decreases with age owing to rescue [43]. In GATA2 deficiency, monosomy 7 requires timely HSCT and watchful waiting is contraindicated; in the SAMD9/SAMD9L syndromes at preschool age monosomy 7 may regress [2]. UPD7q and somatic SAMD9/SAMD9L variants are not seen in GATA2 deficiency, where somatic STAG2 variants appear to play an indirect rescue role [43].

6.4. SAMD9L-Related Syndrome and Other Ataxias with Cytopenias

Ataxia-telangiectasia, dyskeratosis congenita in the Hoyeraal-Hreidarsson form, Fanconi anaemia and X-linked sideroblastic anaemia with ataxia combine neurological and haematological abnormalities, and the Li family was initially compared with ataxia-telangiectasia and Fanconi anaemia [4, 25]. Chen and colleagues showed that ataxia-pancytopenia is distinct from the telomere biology disorders, with which it shares some features [26]. Gonzalez-del Angel and colleagues described in 2000 a Mexican girl with ataxia at 3 years and pancytopenia at 13, severe cerebellar atrophy, no monosomy 7 and increased bleomycin-induced chromosomal aberrations, a picture clinically compatible with ataxia-pancytopenia but suggestive of chromosomal instability, without molecular confirmation in the sources consulted [11]. Adult-onset ataxia without cytopenias, as in SCA49, enters the differential of the dominant spinocerebellar ataxias, and ataxia at any age requires the exclusion of treatable acquired causes [4, 13].

6.5. SAAD and Interferonopathies

The six patients of the founding SAAD series had been diagnosed with CANDLE, and the distinction rested on early interstitial lung disease, progressive B and NK cytopenias, the lower interferon score with raised C-reactive protein and the genetics [5]. In the same NIH cohort, LRBA and IKBKG variants caused pictures also resembling CANDLE with panniculitis and cytopenias, which makes the genetic panel indispensable [5].

7. Treatments and Risk Management Strategies

7.1. General Principles and Quality of Evidence

There is no treatment directed at the cause of any SAMD9 or SAMD9L syndrome, nor any drug approved for these indications; allogeneic HSCT is the only treatment capable of replacing the predisposed haematopoiesis, but it does not correct the non-haematological manifestations [1, 4]. There are no clinical practice guidelines with graded evidence: the available recommendations are those of the GeneReviews, based on the literature and the authors' experience, those of the 2024 American Association for Cancer Research (AACR) working group on paediatric predisposition, which result from expert consensus, and the proposals of Nordic experts and of groups with large series (St. Jude, EWOG-MDS) [1, 2, 4, 34, 66]. The authors of the St. Jude registry consider it still too early to recommend an optimal treatment and surveillance approach and advocate prospective multicentre studies, especially of patients with monosomy 7 [2]. The French national protocol for bone marrow aplasias (Haute Autorité de Santé, March 2023 version), prepared by the reference centre for acquired and constitutional bone marrow aplasias, devotes an appendix to these syndromes, acknowledges that there are no established recommendations and refers the haematological management of each patient to a multidisciplinary team meeting [23]. Treatment is multidisciplinary and should take place in centres with experience in hereditary haematological predisposition, with early consultation of a transplant specialist [2, 66].

7.2. Haematological and Anti-infective Supportive Care

Severe anaemia and thrombocytopenia are treated with red cell and platelet transfusions according to severity, after excluding and correcting secondary causes of anaemia (iron, vitamins), bearing in mind that the macrocytosis of the disease may mimic vitamin B12 or folate deficiency [1, 4]. In severe neutropenia, antibacterial prophylaxis and a fever action plan are recommended [1]. There are no data supporting erythropoietin or granulocyte colony-stimulating factor in SAMD9L-related syndrome, and these growth factors may increase the risk of myelodysplasia and myeloid leukaemia [4]. Infections are treated with antibiotics, antivirals and antifungals according to the agent, and prophylactic intravenous immunoglobulin should be considered when endogenous immunoglobulins are low [1, 4]. In the 22-year-old MIRAGE survivor, monthly intravenous immunoglobulin, started at 6 years, markedly reduced hospital admissions [21]. Yamada and colleagues stress that the evidence is limited and that immunoglobulin should be decided case by case, especially when there is documented B-cell or antibody dysfunction [21]. The same authors recommend inactivated vaccines whenever possible, even with an attenuated response, and consider live attenuated vaccines generally contraindicated when cellular immunodeficiency is suspected or confirmed, to be considered only after careful immunological assessment [21]. The French protocol applies to these syndromes the supportive care common to bone marrow aplasias: transfusions adjusted to the blood count, with transfusion sparing to avoid alloimmunisation; red cells for clinical intolerance of anaemia or, systematically, with haemoglobin below 7 g/dL (higher threshold with comorbidities); platelets for bleeding or, prophylactically, below 20 × 10⁹/L in adults and 10–20 × 10⁹/L in children, a threshold that may fall to 10 × 10⁹/L in mildly symptomatic patients transfused regularly; and hospital assessment of any fever with neutrophils below 0.5 × 10⁹/L [23]. Haematopoietic growth factors have no place in correcting the cytopenias, but granulocyte colony-stimulating factor may be used occasionally in a severe infection, and the place of thrombopoietin receptor agonists is not established, so the protocol does not recommend them [23]; the SAMD9L GeneReview is more restrictive and advises against granulocyte colony-stimulating factor because of a lack of data and the theoretical risk of myelodysplasia [4]. Immunosuppressive treatment with horse antilymphocyte globulin and ciclosporin, the reference treatment in acquired aplasia without a family donor, is not indicated in constitutional bone marrow failures [23]. In multiply transfused patients, the protocol recommends ferritin and transferrin saturation every 3 months, drawn before a transfusion, chelation from ferritin above 500 µg/L and MRI for hepatic and cardiac iron concentration above 1000 µg/L; desferrioxamine is the first choice, despite the subcutaneous route and the risk of Mucor and Yersinia enterocolitica infections, deferasirox exposes to renal, hepatic and, rarely, marrow toxicity, and deferiprone should be avoided because of the risk of agranulocytosis and aplasia in predisposed patients (doses in section 7.10) [23].

7.3. Monosomy 7: Monitor or Transplant

The central decision is the timing of HSCT. The monosomy 7 GeneReview summarises the general rule for predisposition syndromes: success depends on early diagnosis and on performing HSCT before a leukaemic clone emerges, because after transformation to AML the probability of transplant failure increases greatly; any child or young adult with monosomy 7, cytopenias and multilineage dysplasia should undergo HSCT as early as possible, unless rising blasts justify prior cytoreductive chemotherapy [20]. EWOG-MDS recommends HSCT as the preferred initial treatment in MDS with monosomy 7 [43]. In the SAMD9 and SAMD9L syndromes, monosomy 7 behaves distinctively, because it may disappear spontaneously in young children, and the sources propose an approach that depends on age and context. The table compares the proposals.

The indicators of progression that lead to HSCT are worsening cytopenias, growth of the monosomy 7 clone, new chromosomal abnormalities, somatic variants in MDS genes, and progression to multilineage dysplasia or excess blasts [20]. Almost half of MDS with SAMD9/SAMD9L variants and monosomy 7 have additional mutations in SETBP1, ASXL1, RUNX1 or RAS pathway genes, and the detection of somatic mutations with prognostic value should lead to more frequent surveillance and to considering HSCT before transformation [66]. The St. Jude authors acknowledge that watchful waiting requires rigorous blood and marrow assessments and that, with limited data, it is reasonable to follow the general recommendation of timely HSCT in MDS with monosomy 7 [2]. In a longitudinal study of seven patients, three had spontaneous haematological remission at a median of 0.6 years (0.4 to 2.9), with disappearance of monosomy 7 and expansion of somatic SAMD9/SAMD9L variants [2]. The French protocol describes the surveillance that underpins the decision: blood count from once a week, or more often, in the platelet transfusion-dependent patient, to once a month in moderate pancytopenia; bone marrow cytology every 3 months while monosomy 7 is present, with the possibility of review by a cytologist at the reference centre; serial bone marrow karyotype; and molecular surveillance with a two-sided reading, unfavourable (somatic mutations in leukaemic progression genes) and favourable (somatic revertant SAMD9 or SAMD9L variants, whose clones are quantified and followed) [23]. Improvement may be abrupt, shortly after diagnosis, or very gradual, complete or partial, and monosomy 7 may persist for more than a year; reassuring signs are continuous improvement of the blood count, absence of new chromosomal abnormalities with a progressive fall in the percentage of monosomy 7 on FISH, and absence of myeloid mutations with revertant clones of increasing allele fraction [23]. The protocol stresses that data are lacking on the proportion of patients who correct their haematopoiesis, on criteria predicting complete correction or transformation and on a reasonable waiting time; unlike the St. Jude and EWOG-MDS proposals, it does not set an age limit for watchful waiting and decides case by case [2, 23, 43]. The presence of UPD7q is considered a favourable prognostic factor; in the case of Dellal and colleagues, the planned HSCT was abandoned in favour of surveillance with serial blood counts and repeated deep sequencing of SAMD9L in the marrow, with progressive improvement [47].

7.4. Haematopoietic Stem Cell Transplantation: Outcomes

In the EWOG-MDS cohort of 42 patients (26 SAMD9 and 16 SAMD9L), 5-year overall survival was 84% and 93%, respectively, and the prognostic factor was the karyotype: 95% with a normal karyotype versus 76% with monosomy 7; in the 29 transplanted patients, 93% versus 77% [2]. Trottier and colleagues cite a 5-year overall survival of 85% after HSCT in SAMD9/SAMD9L paediatric MDS [48]. In 40 children with refractory cytopenia and monosomy 7, 5-year overall survival after HSCT was 69% with GATA2 deficiency, 77% with SAMD9/SAMD9L and 82% with no known aetiology, without significant differences; in a 2016 EWOG study of 100 children with MDS and monosomy 7, 5-year event-free and overall survival after HSCT were 66% and 69% [43]. Ahmed and colleagues described 12 allogeneic HSCTs in patients with SAMD9 (six, four with MIRAGE) or SAMD9L (six) variants, with a median age of 2.8 years (1.2 to 12.8), myeloablative conditioning in nine and reduced-intensity conditioning in three: 11 engrafted, one SAMD9L patient with MDS and morbid obesity had graft failure and died of refractory AML, another SAMD9L patient died of diffuse alveolar haemorrhage, and 10 of 12 were alive with a median follow-up of 3.1 years (0.1 to 14.7), with resolution of the haematological disease and sustained donor chimerism [71]. The French protocol considers the prognosis of these transplants similar to the usual, except in children with severe extra-haematological involvement (MIRAGE), and describes no particular sensitivity to conditioning or increased risk of graft-versus-host disease [23]. The SAMD9L GeneReview counts 41 published patients transplanted for myelodysplasia, cytopenias, mosaic monosomy 7 or autoinflammation, of whom 36 survived the procedure, with follow-up of up to 16 years [4].

In MIRAGE the outcomes are worse. In the series of Sarthy and colleagues and of Ahmed and colleagues, the eight patients with MIRAGE all had severe complications (adrenal crisis with blood pressure instability and electrolyte disturbances, severe systemic infections, hepatic veno-occlusive disease and respiratory failure), and three died of these complications [1, 71, 72]. The GeneReview considers that HSCT should be considered in severe haematological disease, but that families should be informed that it may worsen pre-existing organ dysfunction and lead to organ failure and death, and that, in some cases, non-curative supportive care may be preferable to preserve quality of life; adrenal hypoplasia requires extreme care during transplantation [1]. The St. Jude registry adds that HSCT in severe SAMD9 cases improves the haematological and immunological problems, but does not alter comorbidities such as enteropathy [2], and the 2026 review includes a patient with severe renal failure after HSCT [21].

HSCT does not halt the progression of the non-haematological manifestations: haematological reconstitution does not improve the neurological and inflammatory manifestations of SAMD9L-related syndrome, and some patients developed ataxia only after a successful transplant [4, 11]. In SAAD, two patients in the NIH series were transplanted (at 20 months, from a matched sibling, and at 9 months) with a positive outcome, and mortality was high in those with severe lung disease, which justifies early diagnosis [5]; across the 12 published SAAD patients, four of five transplanted early had good outcomes [2].

7.5. Transplantation: Conditioning, Prior Therapy and Donor

The optimal pre-transplant therapy in monosomy 7 predisposition syndromes is not defined; several of these syndromes have increased morbidity from prolonged cytopenias after intensive chemotherapy or from prior immunodeficiency, and therefore direct HSCT, without cycles of cytoreductive chemotherapy, is a standard approach in many of them [20]. In children with AML and monosomy 7, high-risk protocols and HSCT with the best available donor in first remission are used [20]. In MDS or AML with a germline SAMD9L variant, Trottier and colleagues recommend limiting high-dose cytarabine, because of the risk of severe cerebellar dysfunction or worsening of ataxia [48]. Data are insufficient to know whether myeloablative conditioning should be modified in these syndromes, unlike Fanconi anaemia and Shwachman-Diamond syndrome, in which toxicity requires reduced intensity [20]. The phase 2 trial NCT04965597, of the Fred Hutchinson Cancer Center, evaluated conditioning with treosulfan, fludarabine and rabbit antithymocyte globulin in bone marrow failure disorders, including neoplasms with germline SAMD9, SAMD9L or GATA2 mutation; it ended in February 2026 with 40 participants and had no results posted in the registry in October 2026 [73]. The Nordic experience illustrates the neurological dilemma: a 4-year-old Swedish child transplanted from a matched unrelated donor, with conditioning according to the EWOG-MDS guidelines (antithymocyte globulin, fludarabine and thiotepa) and without prior MRI or neurological examination, because the diagnosis was not yet known, later developed marked cerebellar atrophy with ataxia and nystagmus, with no established causal relationship (he had had cytomegalovirus infection in infancy), whereas the transplanted Finnish boy had no neurological worsening [34]. The authors advise avoiding high doses of irradiation, especially cranial, in SAMD9L-related syndrome, and suggest, without direct evidence, that HSCT before transformation to MDS could use non-myeloablative conditioning similar to that of Fanconi anaemia or GATA2 deficiency, given the lower competitiveness of the patient's stem cells; after transformation to MDS with monosomy 7, most transplanters opt for myeloablative conditioning, and the decision requires an individual risk-benefit analysis [34].

The French protocol, with the caveat that the data do not allow solid recommendations, proposes for the child standard or reduced-toxicity myeloablative conditioning, with two intravenous regimens: busulfan adjusted to weight according to the summary of product characteristics (SmPC), with pharmacokinetics and an area-under-the-curve target of 900–1350 µmol·min, for 4 days (D-5 to D-2), combined with fludarabine 40 mg/m²/day for 4 days (160 mg/m² per course) and serotherapy; or the EWOG-MDS regimen, with thiotepa on D-7, treosulfan 14 g/m²/day from D-6 to D-4 (42 g/m² per course), fludarabine 40 mg/m²/day from D-6 to D-3 (160 mg/m² per course) and serotherapy [23]. There is no reason for total body irradiation; reduced-intensity conditioning is discussed at a national meeting in the presence of severe comorbidities, especially in MIRAGE, or of prolonged severe aplasia without myelodysplasia [23]. Graft-versus-host disease prophylaxis is standard, with a calcineurin inhibitor combined with short-course methotrexate, or with mycophenolate mofetil when the graft is cord blood; bone marrow is the preferred cell source, to reduce extensive chronic graft-versus-host disease; and alternative donors (unrelated with matching below 10/10 or haploidentical) are discussed at a multidisciplinary meeting in the presence of MDS, acute transformation or prolonged severe aplasia with no tendency to correct [23]. When choosing an HLA-identical sibling, a normal blood count is not enough to exclude carrier status and genetic testing is systematic; in case of doubt in an urgent situation, a 10/10 unrelated donor may be preferred [23]. Before transplantation, a reproductive medicine consultation should be offered (section 8.6) [23].

These proposals diverge from the other sources and from the labels on several points. Davidsson and colleagues suggest, without direct evidence, non-myeloablative conditioning before transformation, whereas the French protocol reserves reduced intensity for severe comorbidities and prolonged aplasia without dysplasia [23, 34]. The thiotepa dose is written in the protocol as 8 mg/m², a value incompatible with conditioning doses, which are expressed in mg/kg: the European label of treosulfan combines it, in the intensified regimen, with thiotepa 5 mg/kg twice on D-2 (10 mg/kg in total), and paediatric series with thiotepa, treosulfan and fludarabine or with thiotepa, busulfan and fludarabine use 8–10 mg/kg on a single day [74, 75]. The protocol, from 2023, stated that the EWOG regimen does not adjust treosulfan to body surface area, contrary to the SmPC for children (10 g/m²/day below 0.4 m², 12 g/m²/day between 0.4 and 1 m² and 14 g/m²/day above 1 m², over 3 days, i.e. a maximum of 42 g/m² per course), and that the drug was not authorised in paediatric non-malignant diseases [23]; the current European SmPC of Trecondi®, renewed in January 2024, now includes malignant and non-malignant diseases in children older than 1 month, in combination with fludarabine with or without thiotepa, uses limits of 0.4 and 0.9 m² and provides for fludarabine 30 mg/m²/day for 5 days (D-7 to D-3; 150 mg/m² per course) instead of the protocol's 40 mg/m²/day for 4 days (160 mg/m²) [74]. The European label of intravenous busulfan refers to a therapeutic window of 900–1500 µmol/L·min, wider than the protocol's target, and does not establish the combination with fludarabine in children [76].

Selection of the related donor is critical, because transplantation from a carrier relative may fail or transmit the disease. The monosomy 7 GeneReview recommends testing potential related donors for the familial variant, preferably in non-haematopoietic DNA; when the patient's variant is not known, the matched related donor should have a blood count and consider bone marrow aspirate, and a fully matched unrelated donor may be more appropriate [20]. The SAMD9L GeneReview recommends early HLA typing of the patient and germline genotyping of at-risk relatives to define the pool of possible donors [4]. The case of Wimalachandra and colleagues, with donor-type aplasia two years after transplantation from a healthy carrier sister, shows that a carrier donor with a normal blood count may be protected only by mosaic rescue [4, 44]. The 2023 British recommendations on patients with germline predisposition proposed for allogeneic transplantation include SAMD9L among the genes with a high risk of non-haematological manifestations and advocate testing related donors for the familial variant [77]. In a St. Jude series of 286 children transplanted for haematological neoplasia between 2017 and 2023, 36 (19%) had a germline pathogenic variant in a predisposition gene and, of five transplanted from a carrier relative, one developed donor-derived leukaemia [78].

7.6. Treatment of AML and Advanced MDS

The associated neoplasms are treated with the protocols for the corresponding neoplasm, by haematology, oncology and transplant teams experienced in MDS and AML [4, 20]. The prognosis of AML with monosomy 7 is poor; even when treated, almost half of patients with monosomy 7 as the main abnormality die of MDS/AML or of treatment complications, and mortality is higher with a complex karyotype [20]. In bone marrow failure syndromes with AML, conditioning may have to start in the aplastic phase after one cycle of chemotherapy, because haematological recovery may not occur [20]. Published cases of rare courses, such as AML with BCR::ABL1 controlled with ponatinib in a SAMD9 patient, do not define general practice [54].

7.7. MIRAGE: Management by Organ

Adrenal insufficiency is treated with glucocorticoid and mineralocorticoid replacement (hydrocortisone and fludrocortisone), guided by endocrinology, with dose increases in stress situations (surgery, infection, trauma) [1, 79]. Nutrition requires follow-up by a dietitian to ensure caloric intake; in chronic diarrhoea, an elemental diet is recommended after excluding other treatable causes; in recurrent aspiration pneumonia, duodenal tube feeding may be considered [1]. In the Japanese survivor, fundoplication and gastrostomy were needed at 2 years [21]. For genital anomalies, gonadectomy of dysgenetic gonads and genital surgery may be considered, and sex of rearing and the management plan are decided by an interdisciplinary team with genetics, endocrinology, surgery and mental health [1]. Developmental delay requires early intervention with occupational therapy, physiotherapy, speech and feeding therapy, and autonomic dysfunction requires artificial tears, ophthalmological follow-up and control of ambient temperature [1]. Renal dysfunction is followed by nephrology [1]; the Japanese survivor received an angiotensin receptor blocker for proteinuria, was dialysed at 20 years and received a living-donor transplant at 21, with an estimated glomerular filtration rate of 80 mL/min/1.73 m² or more after transplantation [21]. Adult primary hypogonadism and osteoporosis were treated with gonadal hormone replacement, active vitamin D and trace elements (zinc and selenium) [21]. Recombinant growth hormone is controversial because of the theoretical risk of MDS and AML: it was given to a patient born small for gestational age, refused in another for fear of tumorigenesis, and Trapani and colleagues treated a boy with documented growth hormone deficiency, without signs of clonal evolution before starting, with improved growth velocity and stable haematological parameters after six years; up to 2026 no case of neoplasia attributed to growth hormone in MIRAGE has been published [21, 52]. A lung transplant for MIRAGE-associated lung disease has also been published [60]. The French protocol, which covers these manifestations for the SAMD9/SAMD9L group as a whole, attributes adult short stature to multiple causes (syndromic short stature associated with intrauterine growth restriction, rarely chronic anaemia and sequelae of corticosteroids and conditioning), recommends systematic follow-up by a paediatric endocrinologist, with screening for delayed puberty and hormone replacement when needed, and accepts that growth hormone may be indicated, without discussing the theoretical risk of clonal evolution raised by the Japanese series; the enteropathy may require parenteral nutrition [21, 23]. In infants, severe Riga-Fede disease (traumatic ulceration of the tongue by the incisors) required excision of the mucosa, frenuloplasty and extraction of the lower incisors in a patient with MIRAGE [80].

7.8. SAMD9L-Related Syndrome: Ataxia and Other Neurological Manifestations

Treatment of ataxia is supportive and aims to maintain independence in walking and activities of daily living and to prevent falls: physiotherapy with balance exercises, gait training and strengthening, occupational therapy, assistive devices (canes, walkers, ramps for powered wheelchairs, weighted cutlery, dressing hooks), home adaptations (grab bars, raised toilet seats), weight control and, in severe ataxia, inpatient rehabilitation [4]. The GeneReview suggests considering a therapeutic trial of riluzole, at a dose of 50 mg twice daily, as it is the only drug with demonstrated improvement of ataxia in some forms (not all), with monitoring of liver enzymes [4]. The systematic review supporting this indication included three trials in hereditary ataxias, two with statistically significant improvement on the ICARS and SARA scales and one without clinical or radiological improvement, with good safety [81]. There are no studies of riluzole in SAMD9L ataxia, and its use is off-label. Dysarthria is treated with speech therapy and alternative means of communication, dysphagia with modification of food consistency (videofluoroscopic swallow study helps choose the consistency), and poor weight gain with nutritional assessment and supplements [4]. Davidsson and colleagues describe the neurology as generally mild and very slowly progressive, with less disability than other genetic ataxias and with a course that usually remains mild to moderate into old age, which allows the patient to be reassured, whereas the GeneReview describes patients who end up in a wheelchair; retinal dysfunction justifies specific visual rehabilitation measures, and access to neurology, ophthalmology and rehabilitation should be ensured [34].

7.9. SAAD and Autoinflammation

The GeneReview recommends considering immunomodulators, corticosteroids or Janus kinase (JAK) inhibitors in autoinflammatory disease, because SAMD9L is interferon-inducible and increased expression of the gain-of-function alleles worsens the manifestations; hypogammaglobulinaemia is treated with immunoglobulin replacement, which also reduces the infections that may trigger flares [4]. The evidence is scarce: in the NIH series, the only SAAD patient treated with a JAK inhibitor had a partial response, with persistent skin disease and a continued need for corticosteroids [5]; two published patients responded partially to anti-inflammatory drugs without HSCT [2]. Experience with baricitinib in the interferonopathies CANDLE and SAVI comes from an expanded access programme in 18 patients, in which the dose started at 100 μg once daily and was increased according to pharmacokinetics and response up to exposures 1.83 times higher than those of adults with rheumatoid arthritis treated with 4 mg/day; the daily symptom score fell from 1.3 to 0.25, prednisone from 0.44 to 0.11 mg/kg/day, and the most frequent adverse effects were upper respiratory infections, gastroenteritis and BK virus viruria and viraemia [82]. This experience did not include SAAD patients and its application to SAAD is an extrapolation. The European label of baricitinib (Olumiant®) has paediatric doses for other indications and prohibits initiation with lymphocytes below 0.5 × 10⁹/L, neutrophils below 1 × 10⁹/L or haemoglobin below 8 g/dL, thresholds that many SAAD patients with cytopenia do not meet [83]. Early HSCT is the option with the best published outcomes in SAAD [2].

7.10. Dosage of the Drugs Mentioned in the Sources

The label doses of hydrocortisone and fludrocortisone are the standard replacement doses for primary adrenal insufficiency and apply to MIRAGE without extrapolation of indication. The riluzole dose in the GeneReview coincides with that of the adult labels, but the indication (ataxia) and use in children are outside the label conditions. None of the sources states a corticosteroid dose for SAAD. The conditioning doses of the French protocol, for children and intravenously, are intended for transplant centres and differ from the paediatric SmPC of treosulfan in fludarabine (40 mg/m²/day × 4, 160 mg/m² per course, versus 30 mg/m²/day × 5, 150 mg/m² per course) and in the absence of body surface area adjustment; the thiotepa dose is discussed in section 7.5. For chelation, the protocol starts treatment with ferritin above 500 µg/L, whereas the deferasirox label starts treatment with ferritin above 1000 µg/L or after about 20 units transfused, and the guideline prevails as the reference [23, 90].

7.11. Regulatory Status

No agency has approved drugs for the SAMD9 or SAMD9L syndromes, and the sources consulted do not describe orphan medicinal product designations for these diseases.

7.12. Trials and Studies with Results

These studies are retrospective or single-arm, with heterogeneous populations and selection of patients referred to reference centres; they do not allow comparisons between them or estimation of the effect of HSCT versus surveillance. There are no randomised trials or trials of targeted therapies in these syndromes; a ClinicalTrials.gov search for SAMD9, SAMD9L, MIRAGE syndrome, ataxia-pancytopenia and monosomy 7 found no other specific interventional studies, which does not prove that none exist [73].

7.13. Developments in the Last 12 Months

Between November 2025 and October 2026, the full update of the SAMD9L GeneReview (August 2026) consolidated the continuum concept, revised the initial evaluation and surveillance and added SAAD and SCA49 to the scope [4]. At the structural (preclinical) level, cryo-electron microscopy structures of SAMD9 explained how gain-of-function variants destabilise autoinhibition and how compensatory mutations restore it [40]; gene-edited cell models allowed novel SAMD9L variants to be classified by measuring translation [68]. At the clinical level (series and cases), publications included the first adult MIRAGE survivor with a kidney transplant [21], the first prolonged growth hormone treatment in MIRAGE [52], a lung transplant [60], new neurological manifestations [58], membranous-type nephropathy [59], familial somatic rescue with mosaic UPD7q detected by trio genome sequencing in the French cohort [47], donor-derived rescue after transplantation [45], thrombotic microangiopathy in an adult with SAAD [63] and dentate nucleus hyperintensities as an imaging clue in SAMD9L ataxia [62]. In diagnostic practice, rapid germline testing in hair bulbs [67] and the low frequency of hidden inherited bone marrow failure in children with severe aplastic anaemia without suggestive features [65] were published. There was no regulatory application, opinion or marketing authorisation for these syndromes in the sources consulted.

7.14. Agents and Circumstances to Avoid

Non-steroidal anti-inflammatory drugs, anticoagulants and thrombolytics are contraindicated in severe thrombocytopenia and should be used with caution outside it, given the fluctuation of the cytopenias [4]. Alcohol and centrally acting sedative drugs worsen gait and coordination and should be avoided in patients with ataxia [4]. Erythropoietin and granulocyte colony-stimulating factor have no data on benefit and may increase the risk of MDS and leukaemia [4]. High-dose cytarabine should be limited in SAMD9L carriers [48], and high doses of irradiation, especially cranial, should be avoided [34]. Live attenuated vaccines require prior immunological assessment in MIRAGE [21]. Hepatotoxic drugs increase the risk of liver injury with riluzole [85]. An episode of malignant hyperthermia was described in a patient with MIRAGE [58]. The French protocol adds deferiprone, because of the risk of agranulocytosis and aplasia, advises against thrombopoietin receptor agonists for lack of data, excludes total body irradiation from conditioning and draws attention to the reduced marrow reserve of these patients (pauciclonal haematopoiesis or transplant), with poor haematological tolerance of chemotherapy [23].

7.15. Pregnancy, Perioperative Period and Transition

There is no information on the effect of pregnancy in SAMD9L-related syndrome; anaemia, thrombocytopenia and neutropenia may increase the risk of complications [4]. Obstetricians should know that fetal macrocephaly, hydrocephalus and intrauterine growth restriction may influence delivery and benefit from perinatal follow-up [4]. In MIRAGE, patients with the typical phenotype have not reproduced, probably because of the severity of the disease and primary gonadal dysfunction [1]. In the perioperative period, patients with adrenal insufficiency need stress doses of glucocorticoid [79, 87], and thrombocytopenic patients need transfusion support adjusted to the procedure [4]. Transition to adulthood is a new problem as more MIRAGE patients survive: the Japanese survivor accumulated renal, endocrine, metabolic and neurological complications with age, and the authors recommend long-term annual follow-up of all organs, even without apparent cardiac, pulmonary or hepatic involvement [21]. In SAMD9L-related syndrome, the haematological risk decreases in adulthood but the neurology progresses, and the St. Jude authors recommend lifelong neurological follow-up [2]. The French protocol recommends a structured transition programme: a preparatory phase in paediatrics (adolescent autonomy, part of the consultation without the parents, early naming of the adult physician, summary of the record), flexible transfer between 16 and 20 years, preferably in a calm phase of the disease, with a joint consultation, alternating consultations in the first year and a liaison nurse, and reception in the adult service with an extended consultation; follow-up is lifelong, even in non-transplanted patients [23]. Therapeutic education relies on the reference centre's booklets (bone marrow aplasia, updated in October 2022), on the bone marrow aplasia emergency card, which should be given to all patients, and on the national follow-up booklet of the Francophone Society of Bone Marrow Transplantation and Cellular Therapy (SFGM-TC) for transplanted patients [23].

8. Reproductive Options and Family Counselling

8.1. Mode of Inheritance

MIRAGE, M7MLS2 and SAMD9L-related syndrome are transmitted in an autosomal dominant manner [1, 3, 4]. MIRAGE results, in the great majority of cases, from a de novo variant, and all SAAD patients described up to 2026 had de novo variants [1, 4]. In SAMD9L-related syndrome, many variants are de novo, but others are inherited from a parent with or without manifestations, and in all multigeneration families described there was marked intrafamilial variation in age at onset, neurology and haematology [4]. NFTC, due to loss of function of SAMD9, is transmitted in an autosomal recessive manner, with the classical 25% risk for each sibling of a patient when both parents are carriers [7, 20]. The same SAMD9 gene therefore has two inheritance patterns depending on the mechanism (dominant gain of function and recessive loss of function).

8.2. Parents of a Proband

Molecular testing of both parents is recommended, to define their status, the need for haematological surveillance and the recurrence risk, preferably in DNA from non-haematopoietic tissue [4, 20]. A proband may appear to be the only affected member of the family because relatives with subtle neurological manifestations or episodic cytopenias have not been recognised, because penetrance is reduced (often due to a second protective event), or because the parent died early or has late-onset disease; an apparently negative family history is confirmed only by testing both parents in non-haematopoietic DNA [4]. If the variant is not found in the parents, there are three hypotheses: de novo variant (confirmed only if biological parentage has been verified; otherwise, “presumed de novo”), germline mosaicism in a parent, or a false-negative result in a carrier parent whose haematopoietic tissue has lost the variant through monosomy 7 or UPD7q [1, 4]. Parental germline mosaicism was presumed in a family with two siblings affected by MIRAGE [1, 9]. Transmission from a healthy parent with early reversion was demonstrated in MIRAGE, by Roucher-Boulez and colleagues, and in SAMD9L-related syndrome, by Dellal and colleagues, whose asymptomatic mother had an allele fraction of 23.4% in blood, a somatic truncating variant and mosaic segmental UPD7q [46, 47]. Testing the parent may reveal pauciclonal haematopoiesis, a sign that the parent has had an aplasia, recognised or not, and corrected it [23].

8.3. Siblings and Other Relatives

If a parent is a carrier, the risk for each sibling of inheriting the variant is 50%; in MIRAGE, carrier siblings will probably manifest the phenotype, and in SAMD9L-related syndrome expression varies greatly within the family [1, 4]. If the proband's variant is not detected in the leukocyte DNA of either parent, the risk to siblings is slightly greater than that of the population, because of germline mosaicism or a parental false negative; if the parents have not been tested and are clinically healthy, siblings are at increased risk [1, 4]. The risk to other relatives depends on the status of the parents, and carrier relatives are at risk of myelodysplasia [1, 4].

8.4. Offspring

Each child of a person with a pathogenic SAMD9 or SAMD9L variant has a 50% chance of inheriting it [1, 4]. People with typical MIRAGE have not had children, because of the severity of the disease and/or primary gonadal dysfunction [1]. In carriers with somatic rescue, the variant may be present in the germline even with a low allele fraction in blood, and the risk of transmission remains [46, 47]. The offspring of an NFTC patient are obligate heterozygotes [20].

8.5. Cascade Testing and Testing of Minors

Testing targeted at the familial variant should be offered to asymptomatic at-risk relatives, to identify those who benefit from haematological surveillance and early treatment of severe cytopenias or myeloid neoplasia; if the blood test is negative, it may need to be repeated in non-haematopoietic DNA, because somatic rescue may give false negatives [1, 4]. When germline testing has been performed in non-haematological DNA and is negative, the relative is no longer at increased risk and can be discharged from haematological surveillance [20]. In families in which the proband has a predisposition to monosomy 7 without an identified molecular cause, all asymptomatic first-degree relatives, adults and children, should have a blood count; if normal, it is repeated annually; if it shows cytopenias or macrocytosis, bone marrow aspirate with cytogenetics and FISH is performed [20]. Testing of at-risk minors is appropriate because it guides surveillance, with prior education of the child and parents, and the consequences of predictive testing (socioeconomic, long-term follow-up) are discussed in formal counselling before sampling [4]. Evaluation of siblings as potential bone marrow donors includes testing for the variant [20]. Counselling explains that a relative may be a carrier and at risk of MDS or severe cytopenias even without previous symptoms, and in families with ataxia-pancytopenia subtle neurological signs may indicate carrier status [34].

8.6. Reproductive Planning

The ideal time to define the risk and discuss prenatal diagnosis and preimplantation genetic testing is before pregnancy, and counselling should be offered to affected or at-risk young adults [1, 4]. Once the familial variant has been identified, prenatal diagnosis and preimplantation testing are possible; opinions on their use vary between professionals and families, and the decision is personal [1, 4]. When the molecular diagnosis is not confirmed, storage of the proband's DNA for future testing should be considered [20]. The existence of healthy carriers with somatic rescue, of incomplete penetrance and of variable expressivity within the same family makes it difficult to predict the phenotype of a fetus carrying an inherited SAMD9L variant, and this uncertainty should be discussed in counselling [2, 4]. According to the French protocol, preimplantation diagnosis requires an identified variant affirmed as pathogenic, analysable in the DNA of a single embryonic cell, which excludes families with a VUS, and the double selection of unaffected embryos that are HLA-identical to an affected sibling is no longer practised in France [23]. Hypogonadism, urogenital malformations and transplant sequelae reduce fertility, and patients should have access to assisted reproduction consultations [23]. Before a transplant, fertility preservation should be offered: before menarche in girls, or before 12–13 years and Tanner stage P3-T3 in boys, only cryopreservation of gonadal tissue is possible (ovarian cortex, with pregnancies described after reimplantation; testicular tissue, still experimental), and in adolescents and adults sperm cryopreservation or ovarian stimulation for oocyte vitrification; profound thrombocytopenia or neutropenia may contraindicate these techniques [23].

9. Special Considerations: Population Diversity and Epidemiology

9.1. Prevalence and Relative Frequency

The population prevalence of the SAMD9 and SAMD9L syndromes is not known. Orphanet classifies MIRAGE, ataxia-pancytopenia, SAAD, SCA49 and familial monosomy 7 syndrome in the prevalence class below 1 per 1,000,000, based on counts of published cases (19, 25, 6, 9 cases and 14 families, respectively) [6, 15–18]. The MIRAGE GeneReview counted 44 patients in 2020, the review by Yamada and colleagues 60 in 2026, the SAMD9L GeneReview more than 150 individuals and the St. Jude registry 305 individuals (243 published and 62 unpublished) in 2024 [1, 2, 4, 21]. The GeneReview considers SAMD9L-related syndrome almost certainly underdiagnosed, because somatic rescue may mask the germline variant [4].

The available frequencies refer to hospital cohorts and not to the general population. In the European EWOG-MDS cohort of 669 paediatric MDS, SAMD9/SAMD9L variants were present in 8% of consecutive cases [22]. Among paediatric bone marrow failure and MDS cohorts, the proportions range from 8% to 18.6% [2, 43]; the AACR working group cites 8%–18% of childhood MDS [66]; the GeneReview attributes to SAMD9L alone 5%–12% of childhood MDS [4]. In the French cohort of suspected inherited bone marrow failure, 16 of the 86 patients with a molecular diagnosis (18.6%) had SAMD9 or SAMD9L [24]. In the same French reference centre cohort, 179 patients, mostly children (median 11 years; 21% under 2 years), with aplasia of constitutional appearance after exclusion of Fanconi anaemia and the other classical forms, SAMD9/SAMD9L variants accounted for 9% of the total and 18% of the 48% with a confirmed genetic cause; the French protocol accepts a relevant proportion of undiagnosed cases, in the form of rapidly resolving pancytopenias [23, 24]. Among bone marrow aplasias, with an incidence of 2–3 cases per million inhabitants per year in Europe and the USA, 15%–20% are constitutional, and the SAMD9/SAMD9L syndromes are among the main ones, with Fanconi anaemia, the telomere biology disorders and congenital amegakaryocytosis [23]. In multicentre cohorts of children and young adults, monosomy 7 is present in about 20% of MDS and 5% of AML, and in a single-centre series it reached 41% of primary paediatric MDS [20]. Considering all chromosome 7 abnormalities, the proportion is 30%–40% in paediatric MDS and about 10% in adult MDS [34]. Among all monosomy 7 predisposition syndromes, the GeneReview attributes 10%–20% to SAMD9 and SAMD9L together, 35%–40% to GATA2 and 5%–15% to Fanconi anaemia [20]. In adults, haematological manifestations are rare and gain-of-function variants were not found in 799 adult MDS [2]. In the French genomic medicine plan (PFMG2025), 11,913 genomes of patients with rare diseases had 33 rare variants in SAMD9 and 29 in SAMD9L, almost all of uncertain significance, a number that measures the burden of rare variants and not the prevalence of the disease [47].

9.2. Geographical Distribution and Founder Populations

The founding series come from Japan (MIRAGE), the USA (Li family, UW-AP family of German, Irish and Native American ancestry), Sweden and Finland (Tesi), Germany and other European countries (EWOG-MDS), France (Bluteau, PFMG2025) and Spain (Menorca family with SCA49) [11, 22, 24, 27, 29, 33, 47]. The sources consulted do not describe founder gain-of-function variants; most variants are private to each family, with a few recurrent hotspots [2]. NFTC is the exception: the SAMD9 variants NM_017654.4:c.4483A>G p.(Lys1495Glu) and NM_017654.4:c.1030C>T p.(Arg344Ter) have been found only in people of Yemenite Jewish origin, with a carrier frequency of about 1% (1 in 92 controls) and origin in two independent events, which suggests positive selection, genetic drift or a population modifier [7]. The sources consulted do not include data from Portugal or Brazil on these syndromes.

9.3. Special Populations

Preschool children are the group at greatest haematological risk and the only one in which watchful waiting with monosomy 7 is acceptable [2, 43]. Adults rarely develop myeloid neoplasia, but accumulate neurological disability [2, 4]. MIRAGE survivors who reach adulthood develop progressive renal, endocrine, metabolic and neurological complications [21]. Healthy carrier family donors constitute a risk group of their own in the transplant setting [44, 78].

9.4. European Networks, Consortia and Patient Organisations

The relevant European Reference Networks are ERN EuroBloodNet, for rare haematological diseases and bone marrow failures (https://eurobloodnet.eu/); ERN PaedCan, for paediatric cancer and predisposition (https://paedcan.ern-net.eu/); ERN RITA, for immunodeficiencies and autoinflammatory diseases, including SAAD (https://ern-rita.org/); ERN-RND, for ataxias and other rare neurological diseases (https://www.ern-rnd.eu/); Endo-ERN, for adrenal insufficiency and differences of sex development (https://endo-ern.eu/); and ERKNet, for kidney disease (https://www.erknet.org/). The reference clinical consortia are EWOG-MDS (https://ewog-mds-saa.org/) and the North American Pediatric Aplastic Anemia Consortium (https://napaac.org/) [2, 20]. St. Jude Children's Research Hospital maintains a variant database and a registry of patients with SAMD9/SAMD9L syndromes (www.stjude.org/samd9) and organised in January 2024 a virtual conference for patients and families, with 230 participants from 27 countries [2]. The patient organisations listed in the GeneReviews are the Aplastic Anemia & MDS International Foundation (https://www.aamds.org/) and the National Ataxia Foundation (ataxia.org), and the National Center for Child Health and Development of Japan maintains a page on MIRAGE (https://www.ncchd.go.jp/) [1, 4]. In France, patients are followed by the reference centre for acquired and constitutional bone marrow aplasias, with sites at Saint-Louis Hospital and Robert-Debré Hospital, in Paris, and a network of competence centres, part of the MaRIH rare immuno-haematological diseases health filière (https://www.marih.fr; https://www.aplasiemedullaire.com); cases are discussed at the reference centre's national multidisciplinary team meetings, the national RIME observatory, created in 2017, anonymously registers patients with acquired and constitutional bone marrow failures and maintains a biobank with blood, marrow and fibroblasts from constitutional forms, and bone marrow aplasia may be recognised as a long-term condition [23]. The French patient organisation indicated for bone marrow aplasia is HPN France – Aplasie médullaire (https://hpnfrance.com) [23].

10. Illustrative Clinical Case

10.1. Infant with Opportunistic Infections, Pancytopenia and Somatic Rescue Inherited from the Mother

Dellal and colleagues described in 2026 a French infant girl who, at 5 months, had severe opportunistic infections (Pneumocystis jirovecii pneumonia and systemic cytomegalovirus infection), with B and NK lymphopenia and T-cell dysfunction despite normal immunoglobulins, which led to suspicion of an inborn error of immunity and inclusion in a national trio genome sequencing programme [47]. She later developed pancytopenia; the marrow showed mild dysplasia without cytogenetic abnormalities, and a sequencing panel on the marrow identified the SAMD9L variant NM_152703.5:c.4529G>C p.(Gly1510Ala) with an allele fraction of 24%, initially classified as of uncertain significance [47]. The trio genome confirmed the variant in blood with an allele fraction of 40.7% and showed that it had been inherited from the asymptomatic mother, in whom the allele fraction was only 23.4%; germline origin was confirmed in DNA from skin fibroblasts, and the variant was absent from gnomAD and ClinVar, at a conserved C-terminal residue [47]. A targeted search for rescue events revealed in the mother a second SAMD9L variant, somatic frameshift c.3650dup p.(His1218Profs*5) with a fraction of 15.4%, probably in cis, and, with a mosaic disomy detection tool applied to the genome data, mosaic segmental UPD7q with a breakpoint between 7q21.2 and 7q21.3; in the patient, the same analysis showed low-level mosaic UPD7q, which explains the fraction of 40% instead of the expected 50% [47]. As UPD7q is considered a favourable factor, the initially planned HSCT was not performed, and the team opted for surveillance with serial blood counts and repeated deep sequencing of SAMD9L in the marrow; eight months later, the allele fraction in the marrow had fallen to 14%, with expansion of the population rescued by UPD7q and the appearance of a new somatic truncating variant, c.985_986insC p.(Lys329Thrfs*3), with a fraction of 6%, in parallel with progressive improvement of haemoglobin, neutrophils and platelets between 5 and 21 months [47]. The same study, in an analysis of the 11,913 genomes of the French plan, reclassified as likely pathogenic a de novo SAMD9L variant in a patient with transient pancytopenia and progressive neurology, upon detecting mosaic UPD7q and a somatic truncating variant despite an apparently normal allele fraction [47].

The figures of the article, whose legends were read in the full text, show the pedigree with the segregation of the variants and the conservation of residue Gly1510 (https://pmc.ncbi.nlm.nih.gov/articles/PMC13340485/figure/bjh70563-fig-0001/), the SNP array-type plot of chromosome 7 with mosaic segmental UPD7q in the mother and in the patient (https://pmc.ncbi.nlm.nih.gov/articles/PMC13340485/figure/bjh70563-fig-0002/) and the longitudinal course of the allele fractions and blood counts (https://pmc.ncbi.nlm.nih.gov/articles/PMC13340485/figure/bjh70563-fig-0003/) [47].

The case brings together the elements that make these syndromes difficult: the initial presentation as immunodeficiency, the variant classified as a VUS on first analysis, the different allele fraction in marrow, blood and skin, transmission from a healthy mother protected by two independent rescue events, and haematological remission without transplantation. The decision not to transplant was taken in a patient with documented adaptive rescue and without monosomy 7, and should not be generalised: in children with monosomy 7, the sources propose specific criteria that depend on age and somatic genetics [2].

10.2. Adult MIRAGE Survivor

Yamada and colleagues described a 22-year-old Japanese man with MIRAGE, born at 33 weeks by emergency caesarean section, with intrauterine growth restriction, generalised hyperpigmentation, atypical genitalia (micropenis, hypospadias, bilateral cryptorchidism), 46,XY karyotype, thrombocytopenia and adrenals not visible on ultrasound, treated from birth with hydrocortisone and fludrocortisone [21]. He had repeated urinary tract infections and aspiration pneumonias, needed tube feeding, fundoplication and gastrostomy, and started monthly intravenous immunoglobulin at 6 years, with a reduction in hospital admissions [21]. An adrenal insufficiency panel was negative, and only at 12 years did SAMD9 sequencing identify the p.(Gln1286Lys) variant; at 21 years, the marrow had three acquired SAMD9 variants, interpreted as reversions, which attenuated the haematological complications without restoring immunity [21]. With age he accumulated developmental delay and autism, white matter lesions and basal ganglia calcifications, seizures, migraines and wheelchair dependence, proteinuria from 6 years progressing to end-stage kidney disease and a kidney transplant from his father at 21 years, severe short stature, primary hypogonadism, osteoporosis and impaired β-cell function [21]. The figures of the article, whose legends were read, show the growth chart, the hyperpigmentation, the genitalia, the tube feeding and the wheelchair (https://pmc.ncbi.nlm.nih.gov/articles/PMC13478688/figure/F1/) and brain MRI and CT at 5 and 20 years (https://pmc.ncbi.nlm.nih.gov/articles/PMC13478688/figure/F2/) [21]. The case documents the treated natural history of a MIRAGE patient without HSCT, with spontaneous haematological rescue and progressive disease in the non-haematopoietic organs.

11. Prognosis and Long-Term Follow-Up

11.1. Prognosis

Prognosis differs greatly between the forms. In MIRAGE, the median age at death in the 44 patients published up to 2020 was 3 years, with almost 60% of deaths due to infection, and in the first two series only 5 of 19 patients were alive at last follow-up [1, 2]; long-term survivors, such as the 22-year-old Japanese patient, accumulate renal, endocrine, metabolic and neurological disease [21]. In paediatric MDS cohorts, 5-year overall survival was 84% in SAMD9 patients (including forms without complete MIRAGE) and 93% in SAMD9L patients, and the karyotype is the main prognostic factor: 95% with a normal karyotype versus 76% with monosomy 7 [2]. In SAMD9L-related syndrome, 9 of 10 patients in one of the first series were alive with a median follow-up of 31 years, the risk of myeloid neoplasia decreases with age owing to rescue, and adult morbidity is dominated by progressive ataxia [2, 43]. UPD7q is associated with stable remissions of up to 20 years, without MDS or AML in the documented cases [2]. Monosomy 7 is associated with rapid progression to advanced MDS or AML in many patients with other predispositions, but in the SAMD9/SAMD9L syndromes at preschool age it may regress [2, 20]. In SAAD, three of the 12 published patients died of progressive disease with infections, and severe interstitial lung disease is associated with high mortality [2, 5]. The sources do not allow estimation of the lifetime risk of myeloid neoplasia or of life expectancy; the absence of events in a patient under follow-up does not equal zero risk.

11.2. Surveillance Recommendations and Divergences

Surveillance recommendations are expert-based and diverge in their intervals. The AACR working group (2024) recommends, for SAMD9 and SAMD9L carriers, a blood count with differential and reticulocytes every 3–4 months, bone marrow aspirate with biopsy, morphology and cytogenetics annually and an annual somatic gene panel, starting at diagnosis, with physical examination at least annually, education about signs and symptoms, follow-up in an experienced centre and early consultation with a transplant specialist [66]. The same group allows the frequency to be reduced, never below annual, in patients with stable cytopenias without transformation on morphology, karyotype or panel, and increased (blood count every 2–4 weeks until resolution, or bone marrow aspirate if the abnormality persists) in the presence of worrying signs or worsening cytopenias [66]. The MIRAGE GeneReview proposes a blood count every 4–6 months and an annual bone marrow aspirate with testing for chromosome 7 abnormalities in patients with cytopenias [1], and Trottier and colleagues propose a blood count every 4–6 months and an annual bone marrow aspirate with molecular and cytogenetic study for all carriers [48]. The SAMD9L GeneReview, updated in 2026, proposes a haematology consultation and blood count every 6–12 months depending on age and the presence of cytopenias or monosomy 7, and considering an annual bone marrow aspirate by shared decision with the patient and family or in the presence of a significant or persistent change in the blood count; surveillance should be tighter with monosomy 7 or other high-risk somatic changes, and immunological surveillance may be unnecessary in patients with extensive rescue [4]. The 2018 Nordic proposal, the oldest, recommends for all carriers without MDS a haematology consultation and blood count with differential every 6–12 months, bone marrow karyotype every 2 years or in the presence of a persistent change in the blood count, a myeloid gene panel every 3 years, a chest radiograph every 3 years (every 5 if the initial one is normal), an annual gynaecological examination in women, adjusted to HPV vaccination and typing, and neurological and ophthalmological rehabilitation support, with involvement of neurology and ophthalmology (SAMD9L), endocrinology (SAMD9), gynaecology, genetic counselling and reproductive medicine [34]. In children aged 5 years or under with monosomy 7 under watchful waiting, the St. Jude group repeats blood count and bone marrow aspirate with somatic testing every 3–6 months [2], and the monosomy 7 GeneReview proposes bone marrow aspirate every 1–2 months until the clone stabilises and then twice a year while it persists [20]. The AACR intervals, the most recent from a paediatric consensus group, serve as the reference in the following table, with the GeneReview alternatives shown alongside. The 2023 French protocol proposes, in a child with aplasia in a phase of possible correction, a weekly to monthly blood count depending on severity and quarterly bone marrow cytology while monosomy 7 persists, with karyotype and molecular surveillance; after normalisation of the blood count and disappearance of monosomy 7, a blood count every 3 months and bone marrow evaluation every 12–18 months; annual measurement of IgG, IgA and IgM, because of the risk of a predominantly B-cell immune deficiency in adulthood, which is treated according to the recommendations for common variable immunodeficiency; systematic neurological follow-up in children and adults and systematic paediatric endocrinological follow-up; and lifelong follow-up of all patients, even without transplantation [23]. The protocol diverges from the other sources on two points: it considers that, in the current state of knowledge, there does not appear to be a risk of MDS or AML in adulthood or an increased risk of spontaneous solid tumours (which may exist after transplantation), whereas the AACR and the GeneReviews maintain haematological surveillance without an age limit and the St. Jude registry describes nine adults with haematological manifestations, two of them with MDS and one with AML; and the bone marrow interval of 12–18 months in remission is longer than the annual interval of the AACR [2, 4, 23, 66].

11.3. Surveillance Table

Carriers identified by cascade testing who are haematologically asymptomatic should have a baseline blood count and bone marrow aspirate with somatic testing and immunological assessment, with surveillance adapted to age and manifestations; SAMD9L carriers need lifelong neurological follow-up, and SAMD9 carriers with endocrine problems need endocrinological follow-up [2]. All surveillance strategies should be adjusted to the clinical context and to the preferences of the patient and family [66].

12. Summary and Final Considerations

The SAMD9 and SAMD9L syndromes are autosomal dominant diseases caused by germline gain-of-function variants in two contiguous genes at 7q21.2 that encode interferon-inducible antiviral proteins; the mutant proteins, active without viral stimulus, cleave phenylalanine tRNA, block translation and suppress the proliferation of renewing tissues. The spectrum ranges from MIRAGE, a neonatal multisystem disease with cytopenias, infections, growth restriction, adrenal insufficiency, atypical genitalia and enteropathy, through paediatric MDS with monosomy 7 without malformations and ataxia-pancytopenia with late and highly variable expression, to SAAD, autoinflammatory and of neonatal onset; NFTC, due to biallelic loss of function of SAMD9, is the allelic disorder with the opposite mechanism. The feature that unifies and distinguishes these syndromes is somatic genetic rescue: in at least two thirds of carriers, haematopoietic tissue eliminates or neutralises the mutated allele through monosomy 7, UPD7q or revertant variants in cis, which may cure haematopoiesis or, in the case of monosomy 7, predispose to MDS and leukaemia.

This rescue has three practical consequences. First, germline testing should be performed on fibroblasts, hair roots or other non-haematopoietic tissue, with SNP array and deep sequencing when suspicion persists, because the variant may disappear from blood and be classified as somatic. Second, UPD7q and somatic SAMD9/SAMD9L variants act as natural proof of the pathogenicity of a variant of uncertain significance. Third, monosomy 7 in these syndromes in young children may be transient, which allows intensive surveillance in selected cases, unlike GATA2 deficiency, while the indication for HSCT remains in older children and in the presence of clinical or molecular progression. HSCT corrects the haematological and immunological disease with high survival outside MIRAGE, but does not alter the neurology or the lesions of other organs, and in MIRAGE it has high morbidity and mortality. Haematological surveillance with blood count, bone marrow aspirate and somatic genetics follows expert recommendations with divergent intervals, and organ management (adrenal, nutrition, kidney, immunity, neurology) is multidisciplinary. Genetic counselling must take into account de novo variants, healthy carriers protected by rescue, germline mosaicism and false negatives in parents' blood, and the choice of related donors requires testing for the familial variant in non-haematopoietic tissue.

13. References

1. Tanase-Nakao K, Olson TS, Narumi S. MIRAGE syndrome. In: Adam MP, Feldman J, Mirzaa GM, Pagon RA, Wallace SE, Amemiya A, editors. GeneReviews®. Seattle (WA): University of Washington, Seattle; 1993–2026. Published 25 November 2020. NBK564655. PMID: 33237688. https://www.ncbi.nlm.nih.gov/books/NBK564655/
2. Sahoo SS, Erlacher M, Wlodarski MW. Genetic and clinical spectrum of SAMD9 and SAMD9L syndromes: from variant interpretation to patient management. Blood. 2025;145(5):475-485. doi:10.1182/blood.2022017717. PMID: 39475954.
3. Online Mendelian Inheritance in Man, OMIM®. #619041 Monosomy 7 myelodysplasia and leukemia syndrome 2; M7MLS2. Baltimore (MD): Johns Hopkins University; created 05-10-2020; last edited 14-12-2020. https://www.omim.org/entry/619041 (accessed October 2026).
4. Raskind WH, Keel S, Allenspach EJ. SAMD9L-related syndrome. In: Adam MP, Feldman J, Mirzaa GM, Pagon RA, Wallace SE, Amemiya A, editors. GeneReviews®. Seattle (WA): University of Washington, Seattle; 1993–2026. Published 1 June 2017; updated 20 August 2026. NBK435692. PMID: 28570036. https://www.ncbi.nlm.nih.gov/books/NBK435692/
5. de Jesus AA, Hou Y, Brooks S, Malle L, Biancotto A, Huang Y, et al. Distinct interferon signatures and cytokine patterns define additional systemic autoinflammatory diseases. J Clin Invest. 2020;130(4):1669-1682. doi:10.1172/jci129301. PMID: 31874111.
6. Orphanet. SAMD9L-associated autoinflammatory syndrome (ORPHA:619367). https://www.orpha.net/en/disease/detail/619367
7. Online Mendelian Inheritance in Man, OMIM®. #610455 Tumoral calcinosis, normophosphatemic, familial; NFTC. Baltimore (MD): Johns Hopkins University; created 28-09-2006; last edited 01-12-2020. https://www.omim.org/entry/610455 (accessed October 2026).
8. Freeman PJ, Hart RK, Gretton LJ, Brookes AJ, Dalgleish R. VariantValidator: accurate validation, mapping, and formatting of sequence variation descriptions. Hum Mutat. 2018;39(1):61-68. doi:10.1002/humu.23348. https://variantvalidator.org
9. Online Mendelian Inheritance in Man, OMIM®. #617053 MIRAGE syndrome; MIRAGE. Baltimore (MD): Johns Hopkins University; created 25-07-2016; last edited 14-09-2021. https://www.omim.org/entry/617053 (accessed October 2026).
10. Online Mendelian Inheritance in Man, OMIM®. *610456 Sterile alpha motif domain-containing protein 9; SAMD9. Baltimore (MD): Johns Hopkins University; created 28-09-2006; last edited 19-02-2021. https://www.omim.org/entry/610456 (accessed October 2026).
11. Online Mendelian Inheritance in Man, OMIM®. #159550 Ataxia-pancytopenia syndrome; ATXPC. Baltimore (MD): Johns Hopkins University; created 02-06-1986; last edited 10-12-2020. https://www.omim.org/entry/159550 (accessed October 2026).
12. Online Mendelian Inheritance in Man, OMIM®. #252270 Monosomy 7 myelodysplasia and leukemia syndrome 1; M7MLS1. Baltimore (MD): Johns Hopkins University; created 04-10-1989; last edited 19-02-2021. https://www.omim.org/entry/252270 (accessed October 2026).
13. Online Mendelian Inheritance in Man, OMIM®. #619806 Spinocerebellar ataxia 49; SCA49. Baltimore (MD): Johns Hopkins University; created 31-03-2022; last edited 24-03-2026. https://www.omim.org/entry/619806 (accessed October 2026).
14. Online Mendelian Inheritance in Man, OMIM®. *611170 Sterile alpha motif domain-containing protein 9-like; SAMD9L. Baltimore (MD): Johns Hopkins University; created 05-07-2007; last edited 27-07-2022. https://www.omim.org/entry/611170 (accessed October 2026).
15. Orphanet. MIRAGE syndrome (ORPHA:494433). https://www.orpha.net/en/disease/detail/494433
16. Orphanet. Familial monosomy 7 syndrome (ORPHA:495930). https://www.orpha.net/en/disease/detail/495930
17. Orphanet. Ataxia-pancytopenia syndrome (ORPHA:2585). https://www.orpha.net/en/disease/detail/2585
18. Orphanet. Spinocerebellar ataxia type 49 (ORPHA:631106). https://www.orpha.net/en/disease/detail/631106
19. Orphanet. Familial normophosphatemic tumoral calcinosis (ORPHA:306658). https://www.orpha.net/en/disease/detail/306658
20. Olson TS, Dickerson KE, Nakano TA, Wlodarski M. Monosomy 7 predisposition syndromes overview. In: Adam MP, Feldman J, Mirzaa GM, Pagon RA, Wallace SE, Amemiya A, editors. GeneReviews®. Seattle (WA): University of Washington, Seattle; 1993–2026. Published 10 June 2021. NBK571103. PMID: 34110727. https://www.ncbi.nlm.nih.gov/books/NBK571103/
21. Yamada Y, Kondo H, Nishida M, Kizaki Z, Tanase-Nakao K, Narumi S, et al. Long-term survival in MIRAGE syndrome: Insights into systemic manifestations and management with review of literature. Endocr J. 2026;73(8):975-982. doi:10.1507/endocrj.ej25-0656. PMID: 42021048.
22. Sahoo SS, Pastor VB, Goodings C, Voss RK, Kozyra EJ, Szvetnik A, et al. Clinical evolution, genetic landscape and trajectories of clonal hematopoiesis in SAMD9/SAMD9L syndromes. Nat Med. 2021;27(10):1806-1817. doi:10.1038/s41591-021-01511-6. PMID: 34621053.
23. Haute Autorité de Santé. Protocole National de Diagnostic et de Soins (PNDS): Aplasies médullaires acquises et constitutionnelles. Saint-Denis La Plaine: HAS; 2023 (version of 1 March 2023, published 2023-05-04; reference centre for acquired and constitutional bone marrow aplasias, MaRIH filière). Available at: https://www.has-sante.fr/jcms/p_3431517
24. Bluteau O, Sebert M, Leblanc T, Peffault de Latour R, Quentin S, Lainey E, et al. A landscape of germ line mutations in a cohort of inherited bone marrow failure patients. Blood. 2018;131(7):717-732. doi:10.1182/blood-2017-09-806489. PMID: 29146883.
25. Li FP, Potter NU, Buchanan GR, Vawter G, Whang-Peng J, Rosen RB. A family with acute leukemia, hypoplastic anemia and cerebellar ataxia: association with bone marrow C-monosomy. Am J Med. 1978;65(6):933-940. doi:10.1016/0002-9343(78)90744-1. PMID: 283689.
26. Chen DH, Below JE, Shimamura A, Keel SB, Matsushita M, Wolff J, et al. Ataxia-Pancytopenia Syndrome Is Caused by Missense Mutations in SAMD9L. Am J Hum Genet. 2016;98(6):1146-1158. doi:10.1016/j.ajhg.2016.04.009. PMID: 27259050.
27. Narumi S, Amano N, Ishii T, Katsumata N, Muroya K, Adachi M, et al. SAMD9 mutations cause a novel multisystem disorder, MIRAGE syndrome, and are associated with loss of chromosome 7. Nat Genet. 2016;48(7):792-797. doi:10.1038/ng.3569. PMID: 27182967.
28. Buonocore F, Kühnen P, Suntharalingham JP, Del Valle I, Digweed M, Stachelscheid H, et al. Somatic mutations and progressive monosomy modify SAMD9-related phenotypes in humans. J Clin Invest. 2017;127(5):1700-1713. doi:10.1172/jci91913. PMID: 28346228.
29. Tesi B, Davidsson J, Voss M, Rahikkala E, Holmes TD, Chiang SCC, et al. Gain-of-function SAMD9L mutations cause a syndrome of cytopenia, immunodeficiency, MDS, and neurological symptoms. Blood. 2017;129(16):2266-2279. doi:10.1182/blood-2016-10-743302. PMID: 28202457.
30. Schwartz JR, Wang S, Ma J, Lamprecht T, Walsh M, Song G, et al. Germline SAMD9 mutation in siblings with monosomy 7 and myelodysplastic syndrome. Leukemia. 2017;31(8):1827-1830. doi:10.1038/leu.2017.142. PMID: 28487541.
31. Pastor VB, Sahoo SS, Boklan J, Schwabe GC, Saribeyoglu E, Strahm B, et al. Constitutional SAMD9L mutations cause familial myelodysplastic syndrome and transient monosomy 7. Haematologica. 2018;103(3):427-437. doi:10.3324/haematol.2017.180778. PMID: 29217778.
32. Wong JC, Bryant V, Lamprecht T, Ma J, Walsh M, Schwartz J, et al. Germline SAMD9 and SAMD9L mutations are associated with extensive genetic evolution and diverse hematologic outcomes. JCI Insight. 2018;3(14):121086. doi:10.1172/jci.insight.121086. PMID: 30046003.
33. Corral-Juan M, Casquero P, Giraldo-Restrepo N, Laurie S, Martinez-Piñeiro A, Mateo-Montero RC, et al. New spinocerebellar ataxia subtype caused by SAMD9L mutation triggering mitochondrial dysregulation (SCA49). Brain Commun. 2022;4(2):fcac030. doi:10.1093/braincomms/fcac030. PMID: 35310830.
34. Davidsson J, Puschmann A, Tedgård U, Bryder D, Nilsson L, Cammenga J. SAMD9 and SAMD9L in inherited predisposition to ataxia, pancytopenia, and myeloid malignancies. Leukemia. 2018;32(5):1106-1115. doi:10.1038/s41375-018-0074-4. PMID: 29535429.
35. Garg D, Sharma P, Faruq M, Agarwal A, Garg A, Punith SB, et al. Spinocerebellar ataxia type 49 presenting with ataxia, early onset dystonia, and bradykinesia in an Indian female patient. Parkinsonism Relat Disord. 2025;130:107188. doi:10.1016/j.parkreldis.2024.107188. PMID: 39514912.
36. Khoury JD, Solary E, Abla O, Akkari Y, Alaggio R, Apperley JF, et al. The 5th edition of the World Health Organization Classification of Haematolymphoid Tumours: Myeloid and Histiocytic/Dendritic Neoplasms. Leukemia. 2022;36(7):1703-1719. doi:10.1038/s41375-022-01613-1. PMID: 35732831.
37. Arber DA, Orazi A, Hasserjian RP, Borowitz MJ, Calvo KR, Kvasnicka HM, et al. International Consensus Classification of Myeloid Neoplasms and Acute Leukemias: integrating morphologic, clinical, and genomic data. Blood. 2022;140(11):1200-1228. doi:10.1182/blood.2022015850. PMID: 35767897.
38. Döhner H, Wei AH, Appelbaum FR, Craddock C, DiNardo CD, Dombret H, et al. Diagnosis and management of AML in adults: 2022 recommendations from an international expert panel on behalf of the ELN. Blood. 2022;140(12):1345-1377. doi:10.1182/blood.2022016867. PMID: 35797463.
39. Zhang F, Ji Q, Chaturvedi J, Morales M, Mao Y, Meng X, et al. Human SAMD9 is a poxvirus-activatable anticodon nuclease inhibiting codon-specific protein synthesis. Sci Adv. 2023;9(23):eadh8502. doi:10.1126/sciadv.adh8502. PMID: 37285440.
40. Mou Z, Zhang F, Morales M, Sahoo B, Dai X, Xiang Y. Structural mechanisms of SAMD9 autoinhibition and pathogenic dysregulation. Sci Adv. 2026;12(39):eaeg3967. doi:10.1126/sciadv.aeg3967. PMID: 42777028.
41. Allenspach EJ, Soveg F, Finn LS, So L, Gorman JA, Rosen ABI, et al. Germline SAMD9L truncation variants trigger global translational repression. J Exp Med. 2021;218(5):e20201195. doi:10.1084/jem.20201195. PMID: 33724365.
42. Russell AJ, Gray PE, Ziegler JB, Kim YJ, Smith S, Sewell WA, et al. SAMD9L autoinflammatory or ataxia pancytopenia disease mutations activate cell-autonomous translational repression. Proc Natl Acad Sci U S A. 2021;118(34):e2110190118. doi:10.1073/pnas.2110190118. PMID: 34417303.
43. Kotmayer L, Kennedy AL, Wlodarski MW. Germline and somatic genetic landscape of pediatric myelodysplastic syndromes. Haematologica. 2025;110(9):1974-1986. doi:10.3324/haematol.2024.285700. PMID: 40568716.
44. Wimalachandra M, Dissanayake R, Raj R, Kulasekeraraj A, Samarasinghe S, Gooneratne L. Donor-type bone marrow aplasia following hematopoietic stem cell transplantation in a child with a novel SAMD9L variant. Hematology. 2024;29(1):2337160. doi:10.1080/16078454.2024.2337160. PMID: 38597819.
45. Tan M, den Elzen N, Goode E, Prabahran A, Blombery P, Fox LC. Donor-derived somatic genetic rescue with prolonged hematopoietic stability following sibling allogeneic stem cell transplant in SAMD9 syndrome. Leuk Lymphoma. 2026;67(1):238-241. doi:10.1080/10428194.2025.2572426. PMID: 41173470.
46. Roucher-Boulez F, Mallet D, Chatron N, Dijoud F, Gorduza DB, Bretones P, et al. Reversion SAMD9 Mutations Modifying Phenotypic Expression of MIRAGE Syndrome and Allowing Inheritance in a Usually de novo Disorder. Front Endocrinol (Lausanne). 2019;10:625. doi:10.3389/fendo.2019.00625. PMID: 31572304.
47. Dellal H, Klifa R, Larcher L, Passet M, Celse T, Gouas L, et al. From a novel pathogenic SAMD9L variant to cohort-wide insights: Whole-genome sequencing highlights somatic genetic rescue and phenotypic heterogeneity. Br J Haematol. 2026;209(1):75-83. doi:10.1111/bjh.70563. PMID: 42154215.
48. Trottier AM, Feurstein S, Godley LA. Germline predisposition to myeloid neoplasms: Characteristics and management of high versus variable penetrance disorders. Best Pract Res Clin Haematol. 2024;37(1):101537. doi:10.1016/j.beha.2024.101537. PMID: 38490765.
49. Rentas S, Pillai V, Wertheim GB, Akgumus GT, Nichols KE, Deardorff MA, et al. Evolution of histomorphologic, cytogenetic, and genetic abnormalities in an untreated patient with MIRAGE syndrome. Cancer Genet. 2020;245:42-48. doi:10.1016/j.cancergen.2020.06.002. PMID: 32619790.
50. Shima H, Koehler K, Nomura Y, Sugimoto K, Satoh A, Ogata T, et al. Two patients with MIRAGE syndrome lacking haematological features: role of somatic second-site reversion SAMD9 mutations. J Med Genet. 2018;55(2):81-85. doi:10.1136/jmedgenet-2017-105020. PMID: 29175836.
51. Nagata Y, Narumi S, Guan Y, Przychodzen BP, Hirsch CM, Makishima H, et al. Germline loss-of-function SAMD9 and SAMD9L alterations in adult myelodysplastic syndromes. Blood. 2018;132(21):2309-2313. doi:10.1182/blood-2017-05-787390. PMID: 30322869.
52. Trapani L, Cognigni M, Maximova N, Valencic E, Fachin A, Tamaro G, et al. Case Report: Growth hormone deficiency and response to treatment in MIRAGE syndrome: expanding the endocrine phenotype. Front Endocrinol (Lausanne). 2026;17:1885153. doi:10.3389/fendo.2026.1885153. PMID: 42483680.
53. Huang Y, Fu J, Gan Z, Xiong F, Zhu H, Yang F. A De Novo Mutation (c.2423A>G) in SAMD9 Causing MIRAGE Syndrome With Intrauterine Growth Retardation and Renal Hypoplasia in a Chinese Family. Hum Mutat. 2026;2026:9470286. doi:10.1155/humu/9470286. PMID: 42006147.
54. Nakata Y, Hiratsuka A, Yoroidaka T, Takahashi W, Nishikawa I, Sakashita M, et al. Late acquisition of BCR::ABL1 during clonal evolution of SAMD9-associated MDS with phenotypic shift from AML to B-ALL. Int J Hematol. 2026;124(3):443-446. doi:10.1007/s12185-026-04248-5. PMID: 42533241.
55. Zhang S, Liu F, Huang X. Case Report: Hidden B-cell dysregulation in MIRAGE syndrome. Front Pediatr. 2026;14:1765896. doi:10.3389/fped.2026.1765896. PMID: 42038248.
56. Gao KM, Huang QYJ, Huang P, Muzaffar A, Beheshti R, Wood R, et al. Normal Immune Function in a Newborn With Early Identification of a SAMD9 Mutation Presenting With Growth Restriction, Thrombocytopenia, and Primary Adrenal Insufficiency. Clin Case Rep. 2025;13(11):e71453. doi:10.1002/ccr3.71453. PMID: 41268294.
57. Li S, Gao F, Lin X, Wang D, Gao S, Ding Y, et al. Case Report: Unveiling the enigma: a rare male neonatal case of MIRAGE syndrome with female external genital presentation and literature review. Front Pediatr. 2026;14:1808088. doi:10.3389/fped.2026.1808088. PMID: 42368282.
58. Garavatti E, Peretz RH, Cruz C, Baiandurova A, Maloney H, Mullin R, et al. Neurodevelopmental Symptoms Associated With MIRAGE Syndrome: A Case Report of Three Children and Review of the Literature. Pediatr Neurol. 2026;179:80-86. doi:10.1016/j.pediatrneurol.2026.03.003. PMID: 41962271.
59. Ishimori S, Imaide A, Yokota T, Narumi S, Yoshikawa N. A 1-year-old boy with MIRAGE syndrome and nephrotic syndrome, whose kidney histopathology revealed membranous nephropathy-like findings: a case report. Pediatr Nephrol. 2026;41(2):357-360. doi:10.1007/s00467-025-06971-3. PMID: 40986043.
60. García Colmenero S, Mira Padilla E, Lamas Fernández FJ. Lung Transplantation in a Rare Disease: A Novel Therapeutic Option for Lung Disease Associated With MIRAGE Syndrome. Arch Bronconeumol. 2026;62(7):518-519. doi:10.1016/j.arbres.2026.01.020. PMID: 41720649.
61. Thunström S, Axelsson M. Leukoencephalopathia, demyelinating peripheral neuropathy and dural ectasia explained by a not formerly described de novo mutation in the SAMD9L gene, ends 27 years of investigations - a case report. BMC Neurol. 2019;19(1):89. doi:10.1186/s12883-019-1319-1. PMID: 31053103.
62. Alm V, Öhberg C, Mitev P, Johansson A, Ludwigs G, Abdulkarim I, et al. Dentate nucleus hyperintensities as a clue to diagnosis in SAMD9L-related ataxia with gain-of-function variants. J Neurol Neurosurg Psychiatry. 2026:jnnp-2026-339495. doi:10.1136/jnnp-2026-339495. PMID: 42951356.
63. Boncukcuoğlu AE, Deniz R, Yürük Kıtay FD, Aydın Gümüş A, Bes C. Thrombotic microangiopathy in adult with SAMD9L-associated autoinflammatory disease: a case report of a novel SAMD9L variant. Rheumatology (Oxford). 2026;65(1):keaf561. doi:10.1093/rheumatology/keaf561. PMID: 41137672.
64. Spranger JW, Brill PW, Hall C, Nishimura G, Superti-Furga A, Unger S. Bone Dysplasias: An Atlas of Genetic Disorders of Skeletal Development. 4th ed. Oxford: Oxford University Press; 2018. p. 812.
65. de Jong JLO, Reed HD, Gaviria Garrido M, Koppayi AL, Ahmed Y, Atkinson I, et al. Low Incidence of Cryptic Genetic Marrow Failure in Pediatric Severe Aplastic Anemia Following Negative Clinical Workup. Blood. 2026:blood.2026034684. doi:10.1182/blood.2026034684. PMID: 42949905.
66. Maese LD, Wlodarski MW, Kim SY, Bertuch AA, Bougeard G, Chang VY, et al. Update on Recommendations for Surveillance for Children with Predisposition to Hematopoietic Malignancy. Clin Cancer Res. 2024;30(19):4286-4295. doi:10.1158/1078-0432.ccr-24-0685. PMID: 39078402.
67. St Martin EC, Mangaonkar AA, Moyer A, Perez Botero J, Ferrer A, Csizmar C, et al. Rapid germ line testing with hair root bulb-derived DNA for hereditary predisposition and inherited marrow failure states. Blood Adv. 2026;10(12):4193-4198. doi:10.1182/bloodadvances.2025018950. PMID: 41915893.
68. Vuelta E, Liquori A, Morín M, Soletto L, Marugal MÁ, Santiago-Balsera M, et al. Functional assessment of inherited myeloid neoplasm-associated SAMD9L germline variants via Monoallelic CRISPR modelling. Br J Haematol. 2026;209(4):1538-1548. doi:10.1111/bjh.70799. PMID: 42665571.
69. Achermann JC, Vilain EJ. NR0B1-Related Adrenal Hypoplasia Congenita. 2001 Nov 20 [updated 2018 Jan 25]. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews®. Seattle (WA): University of Washington, Seattle; 1993-. Available from: https://www.ncbi.nlm.nih.gov/books/NBK1431/
70. Orphanet. Familial glucocorticoid deficiency (ORPHA:361). Available from: https://www.orpha.net/en/disease/detail/361
71. Ahmed IA, Farooqi MS, Vander Lugt MT, Boklan J, Rose M, Friehling ED, et al. Outcomes of Hematopoietic Cell Transplantation in Patients with Germline SAMD9/SAMD9L Mutations. Biol Blood Marrow Transplant. 2019;25(11):2186-2196. doi:10.1016/j.bbmt.2019.07.007. PMID: 31306780.
72. Sarthy J, Zha J, Babushok D, Shenoy A, Fan JM, Wertheim G, et al. Poor outcome with hematopoietic stem cell transplantation for bone marrow failure and MDS with severe MIRAGE syndrome phenotype. Blood Adv. 2018;2(2):120-125. doi:10.1182/bloodadvances.2017012682. PMID: 29365320.
73. Fred Hutchinson Cancer Center. Hematopoietic cell transplantation using treosulfan-based conditioning for the treatment of bone marrow failure diseases. ClinicalTrials.gov NCT04965597. https://clinicaltrials.gov/study/NCT04965597
74. European Medicines Agency. Trecondi (treosulfan) — EPAR, summary of product characteristics (first authorisation on 20 June 2019; last renewal on 5 January 2024). https://www.ema.europa.eu/en/documents/product-information/trecondi-epar-product-information_en.pdf (accessed November 2026).
75. Marcu AD, Jercan CG, Bica AM, Serbanica AN, Radu LE, Avramescu I, et al. Conditioning regimens in pediatric myeloid malignancies undergoing allogeneic HSCT: a comparative single-center study. Front Oncol. 2025;15:1623636. doi:10.3389/fonc.2025.1623636. PMID: 40969262.
76. European Medicines Agency. Busulfan Fresenius Kabi (busulfan) 6 mg/ml concentrate for solution for infusion — EPAR, summary of product characteristics (first authorisation on 22 September 2014; last renewal on 20 June 2019). https://www.ema.europa.eu/en/documents/product-information/busulfan-fresenius-kabi-epar-product-information_en.pdf (accessed November 2026).
77. Clark A, Thomas S, Hamblin A, Talley P, Kulasekararaj A, Grinfeld J, et al. Management of patients with germline predisposition to haematological malignancies considered for allogeneic blood and marrow transplantation: Best practice consensus guidelines from the UK Cancer Genetics Group (UKCGG), CanGene-CanVar, NHS England Genomic Laboratory Hub (GLH) Haematological Malignancies Working Group and the British Society of Blood and Marrow Transplantation and cellular therapy (BSBMTCT). Br J Haematol. 2023;201(1):35-44. doi:10.1111/bjh.18682. PMID: 36786081.
78. Pandey AS, Mostafavi R, Ashcraft E, Wu C, Keerthi D, Lewis SK, et al. Outcomes of germ line testing for children with hematologic malignancies undergoing hematopoietic cell transplantation. Blood Adv. 2026;10(19):6489-6500. doi:10.1182/bloodadvances.2026020262. PMID: 42418371.
79. European Medicines Agency. Alkindi (hydrocortisone) granules in capsules for opening — EPAR, summary of product characteristics. https://www.ema.europa.eu/en/documents/product-information/alkindi-epar-product-information_en.pdf
80. Dairo K, Mehta N, Grischkan J, Kumar A, Sjulin C. Management of Severe Riga-Fede Disease in a Child With MIRAGE Syndrome. Cureus. 2025;17(11):e97970. doi:10.7759/cureus.97970. PMID: 41458652.
81. Ayala IN, Aziz S, Argudo JM, Yepez M, Camacho M, Ojeda D, et al. Use of Riluzole for the Treatment of Hereditary Ataxias: A Systematic Review. Brain Sci. 2022;12(8):1040. doi:10.3390/brainsci12081040. PMID: 36009103.
82. Sanchez GAM, Reinhardt A, Ramsey S, Wittkowski H, Hashkes PJ, Berkun Y, et al. JAK1/2 inhibition with baricitinib in the treatment of autoinflammatory interferonopathies. J Clin Invest. 2018;128(7):3041-3052. doi:10.1172/jci98814. PMID: 29649002.
83. European Medicines Agency. Olumiant (baricitinib) — EPAR, summary of product characteristics. https://www.ema.europa.eu/en/documents/product-information/olumiant-epar-product-information_en.pdf
84. European Medicines Agency. Rilutek (riluzole) — EPAR, summary of product characteristics. https://www.ema.europa.eu/en/documents/product-information/rilutek-epar-product-information_en.pdf
85. U.S. Food and Drug Administration. Riluzole tablets, USP — prescribing information. DailyMed SPL setid dd432c13-389e-46b3-8891-0cf1a58800ff, version of 15 July 2026. https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=dd432c13-389e-46b3-8891-0cf1a58800ff
86. INFARMED. Riluzol Aurovitas MG 50 mg comprimidos revestidos por película — SmPC approved on 09-02-2023 (med_id 45918); Riluzol Kabi MG (693084); Emylif 50 mg película orodispersível (678044). Infomed. https://extranet.infarmed.pt/INFOMED-fo/
87. U.S. Food and Drug Administration. Alkindi Sprinkle (hydrocortisone) oral granules — prescribing information. DailyMed setid 26c13a5f-7119-4c6a-bf10-fda4e07d7682, version of 21 April 2026. https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=26c13a5f-7119-4c6a-bf10-fda4e07d7682
88. U.S. Food and Drug Administration. Fludrocortisone acetate tablets, USP 0.1 mg — prescribing information. DailyMed setid 933e28ef-05e8-419e-a786-21eab67f33a2, version of 24 June 2026. https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=933e28ef-05e8-419e-a786-21eab67f33a2
89. U.S. Food and Drug Administration. Desferal (deferoxamine mesylate) for injection — prescribing information. DailyMed setid a7174843-5965-49fc-b842-f7eff7b48bbc, version of 18 June 2025. https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=a7174843-5965-49fc-b842-f7eff7b48bbc (accessed November 2026).
90. European Medicines Agency. Exjade (deferasirox) — EPAR, summary of product characteristics (first authorisation on 28 August 2006; last renewal on 18 April 2016). https://www.ema.europa.eu/en/documents/product-information/exjade-epar-product-information_en.pdf (accessed November 2026).
91. Agência Nacional de Vigilância Sanitária. Dados abertos — medicamentos registados (DADOS_ABERTOS_MEDICAMENTOS.csv). https://dados.anvisa.gov.br/dados/DADOS_ABERTOS_MEDICAMENTOS.csv