Hereditary Predisposition to Myeloid Neoplasms (MDS/AML)

Myeloid neoplasms with germline predisposition: RUNX1 familial platelet disorder, ANKRD26 and ETV6 thrombocytopenias, CEBPA familial acute myeloid leukaemia, DDX41 predisposition, GATA2 deficiency and SAMD9/SAMD9L syndromes, and their differential diagnosis with bone marrow failure syndromes, Li-Fraumeni and CHEK2.

Literature and regulatory information updated: 17 October 2026.

0. Reference Summary Table

This article groups the entities that the World Health Organization (WHO, 5th edition, 2022) and the International Consensus Classification (ICC, 2022) bring together as myeloid neoplasms with germline predisposition. The inherited bone marrow failure syndromes, Li-Fraumeni syndrome and the telomere disorders belong to the same category [1–3].

1. Introduction

1.1. From familial leukaemia to gene-defined predisposition syndromes

For decades, myelodysplastic syndromes (MDS) and acute myeloid leukaemia (AML) were regarded as acquired diseases, linked to the ageing of haematopoiesis, exposure to benzene, radiation or chemotherapy, and the accumulation of somatic mutations. Familial clustering was known, but the first gene was identified only at the end of the twentieth century. In 1999, Song and colleagues studied six families with autosomal dominant thrombocytopenia, qualitative platelet defects and a propensity to AML, mapped the locus to an 880 kb interval on 21q and found, in one RUNX1 allele (then called CBFA2 or AML1), nonsense variants or intragenic deletions in four families and missense variants at the conserved residues R166 and R201 in the other two; megakaryocytic colony formation was reduced, and the authors proposed a haploinsufficiency model that predisposes to the acquisition of additional mutations [5]. The designation FPD/AML, later familial platelet disorder with associated myeloid malignancy (FPDMM), corresponds to OMIM entry 601399 [4, 18].

In 2004, Smith and colleagues described a family with three members affected by AML who shared the same CEBPA variant, described in historical nomenclature as 212delC, absent from unaffected relatives; latencies to leukaemia were 10, 18 and 30 years, and one patient had a second CEBPA variant in the leukaemia only [62]. In 2011, four independent groups linked GATA2 to four presentations previously thought to be distinct: familial MDS/AML with the p.(Thr354Met) and p.(Thr355del) variants in the second zinc finger [63]; Emberger syndrome, that is, autosomal dominant primary lymphoedema with predisposition to AML [12]; MonoMAC syndrome, with monocytopenia, B and NK lymphopenia and mycobacterial infections [11]; and dendritic cell, monocyte, B and NK lymphoid (DCML) deficiency [13]. In the same year, Pippucci and colleagues showed that thrombocytopenia 2 (THC2) resulted from variants in a 19 base-pair stretch of the ANKRD26 5'-UTR in eight families [64], and Noris and colleagues characterised 78 patients from 21 families [65].

The next two entities emerged in 2015. Polprasert and colleagues described an adult familial AML syndrome caused by germline DDX41 variants, with a somatic second hit in the other allele in half of the carriers with a neoplasm [66]. Three groups described germline ETV6 variants in families with thrombocytopenia, red cell macrocytosis and haematological neoplasms, mainly B-cell acute lymphoblastic leukaemia (ALL) [67–69]. In 2016, Narumi and colleagues defined MIRAGE syndrome as caused by SAMD9 variants in 11 patients and observed MDS with loss of the chromosome 7 carrying the variant [70]; Chen and colleagues attributed ataxia-pancytopenia syndrome, described in 1978 as "myelocerebellar syndrome", to missense variants of SAMD9L [15, 71], and Tesi and colleagues demonstrated the gain of function of these variants [72].

1.2. Classifications: from 2016 to 2022

The 2016 revision of the WHO classification introduced the chapter of myeloid neoplasms with germline predisposition, with germline variants of ANKRD26, CEBPA, DDX41, GATA2 and RUNX1, and the 2017 European LeukemiaNet (ELN) created a category of the same name [8, 44, 73]. The 5th WHO edition (2022) adopted a scalable model combining phenotype and genotype: the qualifier "associated with germline [gene] variant" is added to the neoplasm type, for example "AML with germline RUNX1 pathogenic variant" or "MDS with low blasts associated with germline RUNX1 variant", and genetic counselling and family history become expected components of the evaluation of the index case [1]. The 2022 ICC broadened the title to haematological neoplasms, included TP53, SAMD9, SAMD9L, PAX5 and IKZF1 and organised the entities into four subgroups [2, 74]. ELN 2022 recommends considering germline predisposition in all patients with a haematological neoplasm, regardless of age, and qualifying the AML category with the germline disorder [48]. The 2024 WHO table of myeloid neoplasms, reproduced in the ESMO guide, retains the category "myeloid neoplasms with germline predisposition" among the secondary myeloid neoplasms and notes that the predisposition and the neoplasm itself may influence prognosis and response to therapy [3].

1.3. Definitions of the entities in scope

FPDMM is defined by the presence of a heterozygous pathogenic or likely pathogenic (P/LP) germline RUNX1 variant in a person with thrombocytopenia, a platelet function defect, bleeding or a haematological neoplasm, or in an at-risk relative [4]. ANKRD26-related thrombocytopenia is a non-syndromic congenital thrombocytopenia with normal-sized platelets, without pathognomonic features, which requires molecular confirmation [6]. ETV6-related thrombocytopenia is a non-syndromic disorder that combines thrombocytopenia and leukaemia risk, with no published formal diagnostic criteria [7]. CEBPA-associated familial AML is defined by a heterozygous P/LP germline CEBPA variant demonstrated in a sample free of leukaemic cells [8]. DDX41 predisposition is defined by a heterozygous P/LP germline DDX41 variant with findings suggestive of a myeloid or lymphoid neoplasm, cytopenias or macrocytosis [9]. GATA2 deficiency is a single disease described under three names (MonoMAC, DCML and Emberger syndrome), unified in 2011 [10]. The SAMD9 and SAMD9L syndromes result from heterozygous germline gain-of-function variants and include MIRAGE, ataxia-pancytopenia, monosomy 7 predisposition, SAMD9L-associated autoinflammatory disease and spinocerebellar ataxia 49 [14, 15, 58].

1.4. Related entities

Li-Fraumeni syndrome (TP53) and CHEK2-associated predisposition belong to the myeloid neoplasms with germline predisposition without pre-existing disease, but their tumour spectrum is dominated by solid tumours; subsection 2.9 summarises what is specific to the myeloid context [1, 48]. The inherited bone marrow failure syndromes with predisposition to MDS/AML (Fanconi anaemia, Shwachman-Diamond syndrome, severe congenital neutropenias, Diamond-Blackfan anaemia, telomere disorders), X-linked thrombocytopenia, Bloom syndrome and ataxia-telangiectasia are part of the differential diagnosis, and their discriminating features appear in section 6. ERCC6L2, MECOM and ERG are emerging genes with MDS risk and are part of the differential diagnosis [48, 75].

2. Molecular Genetics and Pathophysiology

2.1. Common model: germline variant, clonal haematopoiesis and second hit

All the entities in the article follow a sequence in which the germline variant alters haematopoiesis from birth and the neoplasm arises after additional somatic events. The mechanisms differ: loss of function with haploinsufficiency (RUNX1, GATA2, ETV6, DDX41), dominant-negative effect of some missense variants (RUNX1, ETV6, GATA2), transcriptional gain of function (ANKRD26) and toxic gain of function with selection of cells that eliminate the mutated allele (SAMD9, SAMD9L) [4, 6, 15, 44]. The second event may affect the wild-type allele of the same gene (CEBPA, DDX41, RUNX1) or consist of chromosomal alterations such as monosomy 7, or of ASXL1 or TP53 variants with multiple events; the presence of these events prompts consideration of early allo-HSCT [44]. ClinGen assigns a haploinsufficiency score of 3 (sufficient evidence) to RUNX1, GATA2, DDX41 and ETV6 [10].

2.2. RUNX1

RUNX1 encodes the DNA-binding subunit of core-binding factor; it associates with CBFβ, which does not bind DNA but allosterically enhances binding to the TGTGGT motif, and it is essential for definitive haematopoiesis and for T-lineage commitment [10]. Its direct targets include ALOX12, PF4, MYL9, PRKCQ, PLDN, RAB1B, TREML1, ITGA2, MPL, PAR4, NFE2, NOTCH4 and A4GALT, which explains the quantitative and qualitative platelet defect [4]. Most germline variants lie in the Runt homology domain; splice-site variants and variants in the transactivation domain are also described. Frameshift variants and large deletions act through haploinsufficiency, and some missense or nonsense variants act through a dominant-negative effect on DNA binding or transactivation [4]. Variants are usually private to each family, which requires sequencing of all exons and CNV analysis [44].

Malignant transformation results from the accumulation of somatic variants (ASXL1, CBL, CDC25C, FLT3, PHF6, SRSF2, WT1) and from loss of function of the second RUNX1 allele [4, 76]. About 30% of carriers have clonal haematopoiesis before leukaemia, with BCOR as the most frequent somatic gene, sometimes with several variants, without correlation with MDS/AML risk; somatic RUNX1 variants as a second hit are found only in the malignant clones [44, 48]. In the National Institutes of Health (NIH) natural history cohort, clonal haematopoiesis with a variant allele fraction (VAF) ≥1% was present in 3 of 65 children (4.6%) and 37 of 117 adults (32%) without a neoplasm [47]. Somatic loss of heterozygosity through uniparental disomy (UPD) of chromosome 21 in haematopoietic tissue, with preferential loss of the chromosome carrying the variant, constitutes revertant mosaicism that may produce false negatives in blood or marrow DNA [4, 77]. In a study of 40 carriers without a neoplasm, telomeres measured by flow-FISH were shortened in lymphocytes in 87.5% and in granulocytes in 95%, all below the age-adjusted 50th percentile, unrelated to somatic mutations, variant type, blood count or bleeding [78].

Braddock–Carey syndrome results from non-recurrent contiguous deletions of 21q22 that include RUNX1 and combines Pierre Robin sequence, agenesis of the corpus callosum, facial dysmorphism, developmental delay or intellectual disability, enamel hypoplasia and congenital thrombocytopenia [4]. Somatic RUNX1 variants are frequent in AML, MDS, chronic lymphocytic leukaemia, clonal cytopenia of undetermined significance (CCUS) and sporadic solid tumours, and are not inherited [4].

2.3. ANKRD26

ANKRD26 encodes a protein with N-terminal ankyrin repeats and C-terminal spectrin helices, expressed in the brain, gastrointestinal tract, liver, adipose tissue and haematopoietic and reproductive tissues, with a cellular function that remains to be clarified [6]. Almost all P/LP variants are single-nucleotide substitutions or small deletions in a 19- to 22-nucleotide stretch of the 5'-UTR that contains RUNX1 and FLI1 binding sites; these factors repress ANKRD26 during megakaryopoiesis, and their failure to bind causes persistent expression, defective proplatelet formation and increased TPO/MPL, JAK/STAT, PI3K and MAPK/ERK signalling [6, 44, 79]. The mechanism is transcriptional gain of function, not haploinsufficiency [10, 64]. One multigenerational family had a complex rearrangement detected by long-read sequencing, with fusion of the WAC promoter and exon 1 to exons 10–34 of ANKRD26 [80]. Germline variants that generate N-terminally truncated isoforms may predispose to myeloid neoplasia without thrombocytopenia [81]. Clonal evolution involves epigenetic regulators, transcription factors and cell cycle genes; in a cohort of 7 ANKRD26 carriers and 3 ETV6 carriers, clonal haematopoiesis was detected in 1 of 7 (SF3B1) and in none of the 3 [6, 44].

2.4. ETV6

ETV6 is a transcriptional repressor that binds the 5'-CCGGAAGT-3' motif, with an N-terminal self-associating pointed domain and a C-terminal DNA-binding ETS domain; it has two initiation codons and is essential for bone marrow haematopoiesis [7, 10]. Most germline variants are missense variants in the ETS domain, with loss of function; nonsense and frameshift variants and exon deletions are also described [7, 44]. The variants p.(Pro214Leu), in the internal linker domain, and p.(Arg369Gln) and p.(Arg399Cys), in the ETS domain, abolish DNA binding, alter subcellular localisation and reduce repression in a dominant-negative manner [67, 68]. Leukaemias in carriers usually lack a somatic second hit in ETV6, although two individuals had biallelic loss [7, 69]. Loss of repression leads to aberrant expression of inflammatory response genes [7]. Somatic ETV6 involvement is the most frequent genetic finding in paediatric B-ALL, with the ETV6::RUNX1 fusion in almost 25% of cases, unrelated to the inherited form [7].

2.5. CEBPA

CEBPA is a single-exon gene with a GC-rich coding region (75%). Two in-frame AUG codons give rise to the 42 kDa isoform, with two transactivation domains, a basic region and a leucine zipper (bZIP), and the 30 kDa isoform, which lacks the first transactivation domain [8]. Germline variants are mainly small insertions or deletions before codon 120 that truncate p42 and preserve p30, which inhibits p42 through a dominant-negative effect; less frequent are variants in the TAD2 or bZIP domains, with loss of function and reduced penetrance [8]. The second, somatic, variant is typically an in-frame insertion or deletion in the bZIP of the wild-type allele; the most frequent cooperating somatic variants are GATA2 (about 30%) and TET2 (about 20%) [44]. C/EBPα is required for the transition from the common myeloid progenitor to the granulocyte-monocyte progenitor and represses E2F1 to enforce proliferative arrest [10]. AML with biallelic (CEBPAdm) or monoallelic (CEBPAsm) CEBPA is defined in the leukaemic cells and does not indicate whether the variant is germline or somatic [8].

2.6. DDX41

DDX41 encodes a DEAD-box family RNA helicase with roles in messenger RNA splicing, sensing of cyclic nucleotides and double-stranded DNA through the STING pathway with type 1 interferon production, post-transcriptional regulation of translation, R-loop accumulation and processing of snoRNA and ribosomal RNA [9, 44]. About two-thirds of germline variants are truncating; splice-site, missense (mainly in the DEAD-box and helicase C domains) and start-loss variants also occur [9, 55]. The mechanism is that of a tumour suppressor, with a somatic second hit in the allele in trans in 55%–70% of carriers with a neoplasm; the somatic variant c.1574G>A p.(Arg525His) accounts for about 65% of second hits in all ethnic groups, and other somatic variants lie in the helicase domain [44]. The VAF of the clone with somatic DDX41 is higher than that of other driver genes, indicating an early event; ASXL1 (10%–20%), DNMT3A, TP53 and CUX1 are the other frequent somatic alterations [44]. In the ALFA-FILO series, the somatic second hit in DDX41 was present in 82% of AML [82]. In the international cohort of Makishima, co-mutation patterns differed from those of cases without DDX41, and the adverse effect of TP53, including multi-hit TP53, was almost abolished by the presence of DDX41 [49]. The germline biallelic form was described in two siblings with dysmorphism, intellectual disability and developmental delay, one of whom had blastic plasmacytoid dendritic cell neoplasm in childhood [9].

2.7. GATA2

GATA2 encodes a transcription factor with two zinc fingers, the N-terminal ZF1 and the C-terminal ZF2, the latter responsible for DNA binding at the 5'-AGATAG-3' motif [10, 44]. Germline variants cause loss of function: null alleles before or within ZF2 (haploinsufficiency), missense variants in ZF2 and variants in the intron 4 enhancer, with an E-box-GATA-ETS motif; up to 80% of cases are de novo, according to the review by Trottier, which cites another source [44]. Null alleles, whole-gene deletions, regulatory intronic variants and synonymous variants that cause selective loss of the mutant RNA converge on the same deficiency, according to a study of 911 patients [10, 56]. Haploinsufficiency causes progressive attrition of stem and progenitor cells, with hypocellular marrow, multilineage dysplasia, prominent erythropoiesis and fibrosis [10]. Somatic GATA2 variants arising in AML, frequently with biallelic CEBPA, cluster in ZF1 [44].

Clonal evolution includes monosomy 7 (30%–40%), trisomy 8 (15%–35%), der(1;7) (9%–20%) and trisomy 21; complex karyotype is very rare. The most frequent somatic variants are STAG2, ASXL1 (whose de novo truncating germline variants cause Bohring–Opitz syndrome), SETBP1 (whose germline degron variants cause Schinzel–Giedion syndrome) and RUNX1, and a second hit in the wild-type GATA2 allele is uncommon [44]. Three marrow states are described: normal marrow without acquired aberrations; hypoplastic marrow or low-grade MDS with predominant STAG2, without association with progression, unlike sporadic MDS, in which STAG2 is adverse; and MDS with excess blasts, AML or chronic myelomonocytic leukaemia (CMML), with SETBP1, RAS pathway and RUNX1 [44, 75]. Genotype–phenotype trends are partial: p.(Thr354Met) is associated with a predominance of myeloid neoplasia (44 of 53, 83%), p.(Arg398Trp) and p.(Arg398Gln) with immunodeficiency (23 of 26, 88%); complete haploinsufficiency may be required for lymphoedema, and non-coding regulatory variants (up to 10%) have reduced penetrance [10, 83]. In the NIH cohort, viral infections and lymphoedema were more frequent with null variants [52], and in the French–Belgian series leukaemias were more frequent with missense variants (14 of 34) than with nonsense or frameshift variants (2 of 28) [50].

2.8. SAMD9 and SAMD9L

SAMD9 and SAMD9L are contiguous paralogues at 7q21, interferon-inducible, with antiproliferative and antiviral functions; the wild-type protein requires viral activation, and gain-of-function variants make it constitutively active [15, 44]. SAMD9L has anticodon nuclease activity directed at phenylalanine transfer RNA, with translational arrest at the elongation phase and ribotoxic stress; overexpression of the variants blocks protein synthesis [15, 84]. More than 90% of germline variants are gain-of-function missense variants located in the middle and C-terminal regions of the protein, near the P-loop NTPase domain; in the autoinflammatory SAMD9L subgroup, truncating variants of a narrow region of the NTPase-like domain also have gain of function [15, 46, 58]. In the EWOG-MDS (European Working Group of MDS in Childhood) cohort, 94% of variants suppressed cell growth [46].

Cells that lose the pathogenic allele gain a selective advantage. There are three routes of elimination: total or partial loss of chromosome 7 with the mutated allele, which causes haploinsufficiency of EZH2, MLL3 and other 7q genes and increases MDS risk (maladaptive rescue); somatic inactivating variants in cis; and copy-neutral UPD 7q that duplicates the wild-type allele (adaptive rescues) [15, 71, 72]. Somatic genetic rescue was observed in 61% of EWOG-MDS patients [46] and in at least two-thirds in the 2025 registry [58]. Adaptive rescues explain spontaneous remissions and incomplete penetrance, but do not alter the neurological manifestations [15, 72]. An asymptomatic female carrier of a SAMD9 gain-of-function variant, who transmitted it to a son with MIRAGE, had a somatic loss-of-function variant in cis [85]. Biallelic loss-of-function variants of SAMD9 cause normophosphataemic familial tumoral calcinosis (OMIM 610455), an entity with no overlap with MIRAGE [14].

2.9. TP53 and CHEK2

In Li-Fraumeni syndrome, the lifetime cancer risk is ≥70% in men and ≥90% in women; ALL, MDS and AML account for only 2%–4% of cancers, but sensitivity to chemotherapy- and radiotherapy-induced carcinogenesis is high, with about 45% second cancers. Germline P/LP TP53 is found in about 6% of AML after cytotoxic therapy and less often in de novo AML, and loss of the wild-type allele with a complex karyotype is frequent [44]. Germline TP53 variants are rare (1:3555 to 1:5476), except for a lower-penetrance founder variant in southern Brazil, according to the 2026 British guideline, which cites another source [86]. As TP53 is one of the most frequent genes in clonal haematopoiesis, the rate of conversion of a somatic finding to germline is low, and the origin must be confirmed in a non-tumour sample and in several samples, given the possibility of mosaicism [86]. CHEK2 is classified by ELN 2022 among the multiple-cancer genes that include myeloid neoplasms, and the CHEK2 I200T variant (historical numbering on isoform NM_001005735.2, = NM_007194.4:c.470T>C p.(Ile157Thr)) is considered predominantly germline when detected in a tumour profile [48]. In a Chicago cohort, CHEK2 was the most frequent P/LP gene, related to enrichment in Eastern European ancestry, and the median age at diagnosis was 62 years [87].

2.10. Cited variants, normalised to the MANE transcript

The variants in the table were validated in VariantValidator (GRCh38, MANE Select transcript) on 17 October 2026 [16]. For RUNX1, the MANE transcript is NM_001754.5 (protein NP_001745.2, RUNX1c isoform). The update of the ClinGen rules for RUNX1 (version 3.1) states residue numbering on NP_001001890.1, but the listed residues (R107, K110, A134, R162, R166, S167, R169, G170, K194, T196, D198, R201, R204) match NP_001745.2; on NP_001001890.1, which lacks 27 N-terminal amino acids, the same R201Q variant corresponds to p.Arg174Gln (NM_001001890.3:c.521G>A) [16, 88]. The historical CEBPA nomenclature "212delC" has no transcript in the abstract of the founding article and was not converted [62].

2.11. Variant interpretation rules

The ClinGen Myeloid Malignancy Variant Curation Expert Panel (MM-VCEP) published rules for RUNX1 in 2019 (version 1), 2022 (version 2) and 2026 (version 3.1) [88, 92]. Version 3.1 applies PM1 at strong level to missense variants at the 13 Runt domain residues listed in subsection 2.10 (without PM5 or PS1 when PM1 applies), PM5 with Grantham score, PM2 at supporting level for frequency ≤0.00005 in gnomAD v4.1.0 (an empirical threshold set by the most frequent pathogenic variant, p.(Arg204Gln), with coverage ≥20× and filtering allele frequency by group), PM4 at strong level for in-frame insertions or deletions at these residues and for stop-loss, PP3 with REVEL ≥0.88 and BP7 with SpliceAI Δ ≤0.20 for intronic variants from +7 or −21; this resulted in 4 reclassifications from likely pathogenic to pathogenic, 13 from VUS to likely pathogenic and 84 from VUS to likely benign [88]. Functional assays of transactivation, DNA binding, cell fractionation and immunofluorescence on 26 RUNX1 VUS from the NIH cohort reclassified 17 variants and showed that the approved assays poorly capture the effect of C-terminal variants [93]. For GATA2, the MM-VCEP defined adult deficiency quantitatively from 339 people at 16 centres in 7 countries and UK Biobank controls, as any of 2903 combinations of 2 or 3 phenotypes with a likelihood ratio ≥1000, in preparation for gene-specific rules [94]. For DDX41 there were no international rules in 2024; the Nordic group advises against using a pathogenic somatic variant to reclassify a germline VUS [55], and a British centre with 239 patients showed that 45.2% of probable germline variants with a co-occurring somatic variant remained VUS under the current rules, and proposed its own rules that take this co-occurrence into account [95].

3. Clinical Spectrum and Manifestations by System

3.1. Penetrance and risk categories

The review by Trottier, Feurstein and Godley separates the syndromes with high malignant penetrance (CEBPA, GATA2, SAMD9/SAMD9L, TP53) from those with variable penetrance (ANKRD26, DDX41, ETV6, RUNX1 and the bone marrow failure syndromes) [44]. The 2024 update of the American Association for Cancer Research (AACR) recommendations for children stratifies MDS risk: above 15% for DDX41 (only in late adulthood), GATA2, Fanconi anaemia, Shwachman–Diamond syndrome, MBD4, ERCC6L2 and ERG; between 5% and 15% for RUNX1, SAMD9/SAMD9L, telomere biology disorders and severe congenital neutropenia; up to 5% for ANKRD26, ETV6, xeroderma pigmentosum, Diamond–Blackfan and MECOM [75]. The figures in the following subsections come from disease-ascertained cohorts and not from population samples; ascertainment bias tends to overestimate penetrance, and early HSCT may bias the reported risk [15, 44].

3.2. RUNX1-FPDMM

Abnormal bleeding or bruising occurs in about 90% of carriers: bruising without trauma, excessive bleeding after surgery, trauma or dental extraction, gingival bleeding, epistaxis and, in women, menorrhagia, miscarriage and bleeding before or after delivery; 20%–25% require platelet transfusions or antifibrinolytics, and severity varies within the same family [4]. Thrombocytopenia is mild to moderate (50–150 × 10^9/L), with normal-sized platelets; some carriers have a normal count, and severe thrombocytopenia usually appears only with MDS or AML [4]. Aggregation is abnormal in most, with absent or reduced response to arachidonic acid and collagen and absent or reduced second wave with ADP and adrenaline; about half have storage pool deficiency on electron microscopy [4]. Eczema or psoriasis, usually mild and of childhood onset, occurs in up to 50% of families [4].

In the prospective NIH cohort (enrolment from January 2019 to March 2025; 213 P/LP carriers analysed), platelets were 131 ± 49 × 10^3/µL in children and 122 ± 52 × 10^3/µL in adults, versus 314 ± 55 × 10^3/µL in paediatric controls and 262 ± 65 × 10^3/µL in adult controls, with no difference in mean platelet volume or immature platelet fraction; eosinophils were 372 ± 244/µL in carrier children versus 165 ± 100/µL in controls. The International Society on Thrombosis and Haemostasis bleeding assessment tool (ISTH-BAT) was abnormal in 8 of 25 children (32%), 13 of 30 men (43%) and 16 of 38 women (42%), with an estimated sensitivity of about 50% in adults and lower in children [47].

GeneReviews gives a lifetime risk of haematological neoplasm of 25%–50% in one passage and 20%–50% in another, with a median age of 33 years; myeloid neoplasms predominate, and T-ALL, B-ALL and lymphomas also occur, with about 25% of families having at least one member with a lymphoid neoplasm [4, 96]. In the NIH cohort, 31 of 213 carriers (14%) had a haematological neoplasm; age at diagnosis was bimodal, with peaks between 0 and 10 years and between 40 and 59 years (mean 32, median 36, range 1–74 years). The 13 paediatric neoplasms included MDS (3), AML (3), T-ALL (3), B-ALL (1), acute myelomonocytic leukaemia, juvenile myelomonocytic leukaemia (JMML) and mixed phenotype leukaemia (1 each); the 19 adult neoplasms were exclusively myeloid (MDS 7, AML 9, CMML 3). The estimated lifetime incidence was 42.5% (95% CI 25.8%–55.4%), and the cumulative incidence at 18 years 7.3% (3.5%–11%), 57.7 times that of the United States population (0.126%); a quarter of the neoplasms arose by 10 years and two-fifths before 18 years, with no apparent risk of solid tumours [47]. The review by Trottier gives a malignant penetrance of about 44%, up to 50%, with almost complete penetrance for thrombocytopenia [44]. Psychological distress is frequent: in an NIH study of 22 carriers aged 18 to 39 years, 50% had above-average anxiety and 64% symptoms of traumatic stress, associated with chronic uncertainty and lifelong monitoring [97].

3.3. ANKRD26-related thrombocytopenia

Thrombocytopenia is lifelong, usually between 50 and 150 × 10^9/L, sometimes below 10 × 10^9/L, with transient correction above 150 × 10^9/L during infections; platelets have normal size and granularity [6]. Penetrance of thrombocytopenia is complete, and the mean platelet count (48 × 10^9/L) is the lowest of the three thrombocytopenias with predisposition [44]. Most have normal haemostasis or mild bleeding, with a negative ISTH-BAT; a positive ISTH-BAT with platelets above 50 × 10^9/L suggests another diathesis [6]. Some carriers have erythrocytosis (haemoglobin up to 18.5 g/dL) or leucocytosis [6, 65]. The marrow shows increased numbers of small, hypolobulated megakaryocytes [6]. In the series of 78 patients from 21 Italian families, THC2 was found in 21 of 210 families in the inherited thrombocytopenia database, with glycoprotein Ia and α-granule deficiency and a higher than expected incidence of acute leukaemia [65]. The prevalence of AML or MDS among carriers was about 8%, with an AML risk estimated at 24 times that of the general population [6, 51]; the monosomy 7 GeneReview cites about 10% [73]. CML occurs in about 1% [44]. No paediatric cases of haematological neoplasm have been described [75].

3.4. ETV6-related thrombocytopenia

More than 150 individuals from about 30 families have been described, without consistent syndromic features [7]. Thrombocytopenia is present in more than 90% at diagnosis, mostly between 100 and 150 × 10^9/L, with normal haemostasis or mild bleeding; counts below 20 × 10^9/L are rare without MDS [7]. Erythrocyte mean corpuscular volume may be elevated [7, 68]. Between 20% and 30% develop a haematological neoplasm, with age at diagnosis between 2 and 82 years (mean 22, median 11 years); about two-thirds are B-ALL, and MDS, AML, mixed phenotype acute leukaemia, diffuse large B-cell lymphoma, polycythaemia vera and multiple myeloma are also described [7, 98]. ALL in children with germline ETV6 is more often hyperdiploid and arises at an older age than sporadic ALL [99]. At least 7 carriers had solid tumours before 50 years, 2 of them colorectal cancer, without a defined risk [7]. GeneReviews describes the risk as "high" in the summary and "moderate" in the body of the text [7].

3.5. CEBPA-associated familial AML

AML arises after the acquisition of a somatic CEBPA variant and other alterations (GATA2 in zinc finger 1, WT1, TET2, EZH2, NRAS) and presents as FAB M1, M2 or M4 AML, with Auer rods and aberrant CD7 expression on blasts, without greater central nervous system involvement [8]. Age at onset ranges from 1.8 to 60 years, against a median of 65 years in sporadic AML [8]. For N-terminal variants, the lifetime AML risk exceeds 80% (about 90% in the Trottier synthesis, median age 25 years); for variants downstream of codon 120 (TAD2, bZIP), present in 20%–30% of carriers, it is about 46%–50%, with a median age of 30 years (6–60) [8, 44]. Most present with overt AML without preceding cytopenias or myelodysplasia, although cytopenias may precede AML by a few months, as in a Uruguayan family [8]. After cure, new independent leukaemic episodes arise with a somatic CEBPA variant different from the original clone, often more than 3 years after diagnosis; the median time to first "relapse" was 27 months for 5' variants and 9 months for 3' variants [8, 44, 100]. In an analysis of 56 patients from 40 families, TAD2 variants were associated with later onset (40 versus 21 years), more cooperating somatic variants (64.3% versus 27.3%), a lower complete remission rate after induction (69.2% versus 95.3%) and worse overall survival [101].

3.6. DDX41 predisposition

DDX41-associated myeloid neoplasms typically arise in the sixth and seventh decades, with a median age of 66–68 years, similar to that of the sporadic form; high-risk MDS and AML predominate, with hypocellular marrow, prominent erythroid dysplasia (sometimes with erythroleukaemia morphology), normal karyotype in 59%–85% and low leucocyte counts [9, 44]. Unexplained mild cytopenias of one or more lineages or macrocytosis precede the neoplasm in 40%–66%, sometimes for years, and aplastic anaemia has been described [9]. In the international cohort of Makishima (9082 patients with myeloid neoplasms from 56 institutions in 7 countries, 346 with germline and/or somatic P/LP DDX41, 525 first-degree relatives), the risk was negligible up to 40 years and rose to 49% by 90 years; male predominance was 79.4% in patients with DDX41 versus 61.2% in the others (odds ratio 2.45), with no sex difference among healthy carriers; MDS with DDX41 progressed more rapidly to AML, but only with truncating variants [49]. In a French study of 63 families and 160 genotyped relatives (80 carriers), the cumulative risk at 70 years was 19.5% (95% CI 5.8%–62.5%), with a relative risk of 55 compared with the general population and cases before 50 years in women [102]. In the ALFA-FILO series of 191 AML with germline DDX41, 75% were men in their seventh decade, with a median leucocyte count of 2 × 10^9/L, median marrow blasts of 33% and normal cytogenetics in 75% [82].

Lymphoid neoplasms (early-onset follicular lymphoma, Hodgkin lymphoma, myeloma, monoclonal gammopathy, chronic lymphocytic leukaemia, ALL) and myeloproliferative neoplasms or CMML also occur; 9% of patients with MDS/AML had a prior haematological neoplasm [9]. Autoimmune diseases (sarcoidosis, rheumatoid arthritis, lupus, eosinophilic granulomatosis with polyangiitis, Graves' disease) and solid tumours (prostate, colorectal, melanoma; 6 of 33 with a haematological neoplasm had a prior solid tumour) have been described, without a demonstrated excess over the population [9]. Only 27%–39% of carriers with a myeloid neoplasm have a family history of haematological neoplasm [9], and the family history often appears sporadic because of elderly relatives, distant generations or deaths from other causes [10, 82]. In the table of phenotypes of constitutional bone marrow aplasias in the French PNDS (French National Diagnosis and Care Protocol) for bone marrow aplasias, published by the Haute Autorité de Santé (HAS) in 2023, DDX41 appears as an autosomal dominant predisposition, rarely biallelic, without extra-haematological features, with rare but described bone marrow aplasias, MDS or AML between 44 and 85 years, normal karyotype in about half and a somatic variant of the second allele in about half of cases [61].

3.7. GATA2 deficiency

GATA2 deficiency affects the haematological, immune, lymphatic, pulmonary, auditory, genitourinary and neurological systems [44]. In the NIH cohort (57 patients, 1992–2013), MDS occurred in 84%, AML in 14% and CMML in 8%; severe viral infections in 70%, disseminated mycobacterial infection in 53% and invasive fungal infection in 16%; impaired pulmonary diffusion in 79%, ventilatory impairment in 63%, alveolar proteinosis in 18% and pulmonary hypertension in 9%; warts in 53%, panniculitis in 30%, human papillomavirus (HPV)-associated tumours in 35%, venous thrombosis in 25%, lymphoedema in 11%, sensorineural deafness in 76%, miscarriages in 33% and hypothyroidism in 14% [52]. The non-tuberculous mycobacteria are mainly of the Mycobacterium avium complex, the origin of the name MonoMAC; progressive loss of dendritic cells, monocytes, B and NK lymphocytes (with absence of the CD56bright NK subset) and inversion of the CD4:CD8 ratio coexist with poor-quality humoral responses despite often normal immunoglobulins [10]. Primary lymphoedema affects 11%–20% of carriers, in one or both lower limbs, with frequent genital involvement; together with AML, with or without congenital sensorineural deafness, it constitutes Emberger syndrome [10, 103].

Neutropenia is often the first cytopenia; profound monocytopenia suggests the diagnosis, but monocytosis may be the initial sign in MDS [10, 45]. In the French–Belgian series (79 patients from 53 families, 2249 patient-years), the median age at first symptoms was 18.6 years (0–61), the probability of remaining asymptomatic at 40 years 8% and mortality by 40 years up to 35%; ALL, JMML, fatal progressive multifocal leucoencephalopathy due to JC virus and immune or inflammatory diseases were also described [50]. The penetrance of myeloid neoplasia is about 75%, with a median age of 20 years, and at least 80% according to the AACR [44, 75]. The marrow progresses from normocellular to hypoplastic, then hypercellular with dysplasia and, finally, MDS/AML [75]. The French PNDS (French National Diagnosis and Care Protocol) for bone marrow aplasias regards typical aplasia as rare in GATA2 deficiency: cytopenias predominate, mainly neutropenia and monocytopenia, with hypoplastic marrow and frequent moderate dysplasia; the cumulative incidence of MDS/AML is estimated at about 70% by 70 years, with diagnoses between 4 and more than 60 years, and atypical CMML and chronic myeloid leukaemia are described [61]. The acquired alterations cited are monosomy 7, trisomy 8, del(20q) and gain of 3q, and somatic ASXL1 and SETBP1 variants are considered of poor prognosis; the same document adds erythema nodosum and vascular malformations to the phenotype and includes prior thrombosis among the items to be sought in the history [61].

3.8. SAMD9 (MIRAGE)

MIRAGE (myelodysplasia, infection, restriction of growth, adrenal hypoplasia, genital phenotypes and enteropathy) was defined in 44 individuals, with myelodysplasia in 37, recurrent infection in 40, growth restriction in 43, adrenal hypoplasia in 34, atypical external genitalia in 33 of 34 46,XY individuals and gastrointestinal complications in 36 [14]. Severe cytopenias arise between birth and 2 years in about 60% and later in childhood in about 40%; before clonal evolution, the marrow is hypoplastic, with features of amegakaryocytic thrombocytopenia; thrombocytopenia is the most frequent cytopenia, and MDS is defined by monosomy 7 or del(7q) with multilineage dysplasia [14]. About 80% have primary adrenal insufficiency, with hyperpigmentation at birth, hyponatraemia, hyperkalaemia, hypoglycaemia and markedly elevated ACTH [14]. Chronic diarrhoea affects about 80%, usually after the start of enteral feeding, and about a third have dysphagia, reflux, achalasia or aspiration pneumonia [14]. Developmental delay is moderate to severe in most, with microcephaly, hydrocephalus, white matter changes, dysautonomia in 7 and renal dysfunction in 5 of 17 who lived beyond 3 years [14]. The median age at death was 3 years, about 60% from infection [14]. Some carriers have only myelodysplasia, with few or no other features [14, 58].

3.9. SAMD9L (ataxia-pancytopenia and related forms)

More than 150 individuals have been described, with marked inter- and intrafamilial variability and non-concordant haematological and neurological penetrance [15]. Most have haematological abnormalities in childhood: single-lineage cytopenias, hypocellular marrow failure that may mimic aplastic anaemia, and MDS, mainly with monosomy 7, del(7q) or copy-neutral UPD 7q; monosomy 7 may be transient, and multiple clones expand, contract or disappear [15, 104]. MDS and leukaemia are rare in adults, with the latest onset described at 42 years [15, 104]. Immunodeficiency, mainly B and NK deficiencies, affects about a quarter [15, 46]. Neurological involvement is present in most adults examined: horizontal and vertical nystagmus, brisk reflexes, clonus, Babinski sign, sensorimotor neuropathy and slowly progressive cerebellar ataxia, with onset between childhood and 62 years; imaging shows cerebellar atrophy, white matter changes and basal ganglia calcifications, which may precede the clinical signs [15]. SAMD9L-associated autoinflammatory disease, with de novo truncating variants in the NTPase-like domain, begins in the neonatal period or infancy with B-cell aplasia and hypogammaglobulinaemia, neutrophilic panniculitis, selective loss of Langerhans cells, marrow failure, basal ganglia calcifications and interstitial lung disease [15, 105]. Spinocerebellar ataxia 49 was described in a family from Menorca with purely neurological manifestations [15].

In the EWOG-MDS cohort (669 paediatric MDS), germline SAMD9/SAMD9L variants were present in 8%, mutually exclusive with GATA2 (7%); 90% had refractory cytopenia, 38% acquired monosomy 7, 57% constitutional anomalies and 28% immune dysfunction [46]. The clinical course separated stable disease (5 of 67), remission (5 of 67, diagnosis between 1.6 and 4.3 years) and high-risk or progressive disease (21 of 67), in which all were transplanted and 17 of 21 were alive despite monosomy 7 in 76% [46].

In the cohort of the French reference centre for bone marrow aplasias (179 patients, mostly children, median age 11 years, 21% younger than 2 years, with constitutional-appearing aplasia after exclusion of Fanconi anaemia and the classic forms), SAMD9/SAMD9L variants accounted for 9% of the total and 18% of the 48% with a confirmed genetic cause; the 2023 PNDS, which summarises this cohort, acknowledges that some cases, with rapidly resolving pancytopenias, may remain undiagnosed [61, 91]. The same document abandons the strict correspondence between MIRAGE and SAMD9 and between ataxia-pancytopenia and SAMD9L, because no extra-haematological sign is specific to one of the genes, and prefers the designation SAMD9/SAMD9L syndrome. To the phenotype already described it adds alveolar proteinosis, hypogonadism, ataxia that may appear only in adulthood or old age, white matter changes on magnetic resonance imaging in children, a pseudo-parkinsonian syndrome, behavioural disorders and mild dysmorphism; some patients have no extra-haematological signs when the aplasia is diagnosed, and these may appear later or be revealed by targeted investigations [61]. Somatic SETBP1, RUNX1, ETV6 and ASXL1 variants are described in the leukaemic clones [61]. In adults, the PNDS does not identify, on current data, an increased risk of MDS/AML or of spontaneous solid tumours, but notes the possible onset of hypogammaglobulinaemia with mainly B-cell immune deficiency, reduced marrow reserve (pauciclonal haematopoiesis or post-transplant state) and poor haematological tolerance of chemotherapy [61].

4. Clinical Recognition and Red Flags

4.1. When to suspect, by setting

The 2022 ELN lists the features that should prompt consideration of germline testing in a patient with a haematological neoplasm: a personal history of two or more cancers, one of them haematological, in any order; a relative within two generations with another haematological neoplasm, with a solid tumour by age 50 or with other haematological abnormalities; a deleterious variant on tumour profiling that may be germline, especially if it persists in remission; and a diagnosis earlier than average, for example MDS by age 40 [48]. The Nordic guidelines add signs of an inherited condition: excessive toxicity from chemotherapy or radiotherapy, multiple cancers, therapy-related leukaemia, poor stem cell mobilisation in a sibling donor candidate, consanguinity, skin or nail changes, unexplained liver disease, pulmonary fibrosis or alveolar proteinosis, short stature, microcephaly, skeletal or congenital anomalies, café-au-lait spots, hypopigmented macules, lymphoedema, immunodeficiency, atypical infections and excessive warts [53]. The absence of a family history does not exclude predisposition, because of de novo variants, gonadal mosaicism, genetic reversion and variable penetrance and expressivity [53].

4.2. Red flags by entity

Mild to moderate thrombocytopenia with normal-sized platelets, present since childhood, with an autosomal dominant family history of thrombocytopenia or myeloid neoplasia, points to RUNX1, ANKRD26 or ETV6 [6, 44]. A previous diagnosis of immune thrombocytopenia unresponsive to immunosuppression is an explicit red flag for ANKRD26 [6]. Bleeding out of proportion to the platelet count suggests RUNX1, and red cell macrocytosis suggests ETV6 [7, 73]. In AML, biallelic CEBPA, age under 50 years and a family history of AML suggest the familial form [8]. In older adults, hypocellular MDS with erythroid dysplasia, normal karyotype and cytopenias or macrocytosis of years' duration, especially in a man, or a DDX41 variant with a VAF close to 50% on tumour profiling, suggest DDX41 [9, 106]. Truncating DDX41 variants and CHEK2 I200T (= NM_007194.4:c.470T>C p.(Ile157Thr)) are predominantly germline and warrant testing even when detected only once [48].

Neutropenia with profound monocytopenia, B and NK lymphopenia, non-tuberculous mycobacterial infection, extensive warts or HPV dysplasia, lymphoedema, sensorineural deafness or alveolar proteinosis suggest GATA2 deficiency [10, 52]. Disseminated mycobacterial disease without monocytopenia or B and NK lymphopenia points instead to a defect of the IL-12/IFN-γ axis. Paediatric MDS, especially with monosomy 7, requires GATA2 testing [10, 45]. Intrauterine growth restriction, neonatal adrenal insufficiency, atypical genitalia in a 46,XY newborn, chronic diarrhoea and cytopenias point to MIRAGE; cerebellar ataxia with nystagmus, cytopenias or transient monosomy 7, or a neonatal autoinflammatory disease with B-cell aplasia, point to SAMD9L [14, 15]. On physical examination, short stature suggests Fanconi anaemia or MIRAGE, microcephaly Fanconi anaemia, LIG4 or ERCC6L2, café-au-lait spots Fanconi anaemia or neurofibromatosis 1, lymphoedema GATA2, ataxia SAMD9L and skeletal dysplasia Shwachman–Diamond syndrome [73].

In bone marrow aplasia, the French PNDS (French National Diagnosis and Care Protocol) adds as arguments for a constitutional form a blood count already abnormal for more than 6 months, a family history of cytopenias, MDS or leukaemia, fetal haemoglobin above 5% before any transfusion and, for GATA2 deficiency, prior thrombosis; conversely, a recent normal blood count, very rapidly worsening cytopenias, recent hepatitis, exposure to drugs or toxins and a PNH clone of at least 1% favour the acquired form [61]. In children, SAMD9/SAMD9L syndrome is suggested by age under 5 years, a family history, relatively acute onset, sometimes after an infection or vaccination (which may mimic acquired aplasia), monosomy 7 present from diagnosis and spontaneous, rapid improvement of the counts, in the absence of neonatal thrombocytopenia (congenital amegakaryocytosis), a Hoyeraal–Hreidarsson picture (telomeropathy) or signs of MECOM syndrome [61].

4.3. Emergency signs

A new or progressive cytopenia, blasts in peripheral blood, emergence of monosomy 7 or of a clone with a second hit (DDX41, RUNX1, CEBPA), severe systemic infections and severe bleeding require immediate haematological evaluation, with bone marrow examination [44, 53, 55]. In MIRAGE, severe dehydration with hypotension reflects potentially fatal adrenal insufficiency [14]. In children with monosomy 7 and cytopenias with multilineage dysplasia, urgent referral to a transplant centre should be made as soon as possible, because HSCT failure increases after transformation to AML [73]. Carriers of inherited thrombocytopenia should have a medical letter for school and an alert bracelet, because bruising may raise suspicion of maltreatment [4, 47]. There is no newborn screening for these entities.

5. Diagnosis: Criteria, Testing and Algorithms

5.1. WHO 2022 and ICC 2022 classification

The two 2022 classifications agree on the core genes but differ in structure. The WHO retains the heading "myeloid neoplasms associated with germline predisposition" with three subtypes; the ICC uses the heading "haematological neoplasms with germline predisposition", includes an ALL subgroup and also treats TP53, DDX41, RUNX1 and ETV6 as lymphoid predisposition [1, 2]. The ICC also provides a provisional category of genes with emerging evidence and retains the term refractory cytopenia of childhood, extended to the early manifestation of MDS in germline predisposition [2, 74].

Source: WHO 2022, Table 10 [1]; ICC 2022, Table 24 [2].

The WHO 2022 harmonised the cytopenia thresholds for CCUS and MDS (haemoglobin <13 g/dL in men and <12 g/dL in women, neutrophils <1.8 × 10^9/L, platelets <150 × 10^9/L), retained the 10% dysplasia threshold in all lineages and defined clonal haematopoiesis of indeterminate potential (CHIP) as somatic mutations in myeloid neoplasm genes with VAF ≥2% without haematological disease or unexplained cytopenia [1]. Hypoplastic MDS is a distinct type, and carriers of germline variants of GATA2, DDX41, Fanconi genes or telomerase complex genes may have hypoplastic marrow that does not respond to immunosuppression [1]. In childhood MDS with low blasts, which replaces refractory cytopenia of childhood in the WHO, exclusion of bone marrow failure syndromes and germline variants is a prerequisite; about 80% are hypocellular, and monosomy 7, del(7q) and complex karyotype confer a higher risk of AML [1]. The WHO 2022 includes biallelic forms and single bZIP mutations in AML with CEBPA mutation, and the ICC defines the favourable category by in-frame bZIP mutations [1, 2]. The ICC created the MDS/AML category (10%–19% blasts) in adults [2]. Morphological megakaryocytic dysplasia is common in the marrow of RUNX1, ANKRD26 or ETV6 carriers with isolated thrombocytopenia without MDS [2].

5.2. Criteria for ordering germline testing

The 2019 Nordic guidelines propose criteria for adults [53]: A1, MDS/AML with signs of an inherited condition diagnosed before 50 years; A2, two first- or second-degree relatives with MDS/AML, prolonged thrombocytopenia or signs of an inherited condition, one diagnosed before 50 years; A3, one patient with MDS/AML and two first- or second-degree relatives with solid tumours suggestive of CMMRD, Li-Fraumeni or BRCA2, one before 50 years; A4, three or more first- or second-degree relatives with myeloid neoplasia, prolonged thrombocytopenia or signs of an inherited condition, regardless of age; B, a somatic panel variant suspected of being germline (VAF of 40%–60% or above 90%), especially in RUNX1, GATA2, TP53, ETV6, CEBPA or truncating DDX41 variants; C, patients without A or B diagnosed before 50 years with monosomy 7, del(7q) or der(7). The 50-year threshold is arbitrary and under-represents DDX41 [53], so the 2024 Nordic update recommends including DDX41 in the work-up of all patients with myeloid neoplasms undergoing genetic investigation [55]. The 2022 ELN recommends considering predisposition in all patients with a haematological neoplasm, regardless of age, and testing as early as possible, because relatives are the preferred donors [48]. The NCCN recommends, according to the AACR, testing all patients with MDS and adults with AML before 50 years; the NCCN guideline was not accessible for direct reading [75]. In children, the MIPOGG tool helps to identify whom to refer [75].

5.3. Laboratory findings by entity

The blood count with film, reticulocytes and bone marrow examination with cytogenetics are the basis. In RUNX1, platelet aggregation shows the aspirin-like pattern, and the marrow is hypo- or normocellular with atypical megakaryocytes (small, hypolobulated nuclei) and eosinophilia, without overt dysplasia; unfamiliar pathologists may diagnose MDS in error, which should be considered only with additional cytopenias, overt dysplasia or expansion of myeloblasts [4]. In ETV6 the marrow may show small hyperchromatic megakaryocytes and dysplastic eosinophils without myelodysplasia [7]; in ANKRD26, increased, small, hypolobulated megakaryocytes, with a described baseline increase in marrow blasts, so general-population blast thresholds do not apply directly [6, 10]. In GATA2 deficiency, immunophenotyping shows dendritic cell deficiency, monocytopenia, loss of transitional B lymphocytes, absence of CD56bright NK cells and CD4:CD8 inversion; none of these tests confirms the diagnosis, and the pattern may be absent in childhood [10]. In MIRAGE, a deficient response to cosyntropin confirms adrenal insufficiency [14]. When monosomy 7 is suspected, the karyotype should analyse at least 20 metaphases from unstimulated samples, with confirmation when at least 3 of 20 cells lack chromosome 7; a normal karyotype does not exclude future monosomy, and FISH or microarray are preferable during follow-up [73]. Chromosomal fragility with mitomycin C or diepoxybutane and telomere length complete the investigation of bone marrow failure syndromes [73].

5.4. Choice of tissue and germline confirmation

In patients with active haematological neoplasia, DNA from blood, marrow and saliva is dominated by tumour cells and cannot be used to confirm a variant as germline [54]. The 2022 British consensus establishes a hierarchy (87% agreement): skin biopsy with cultured fibroblasts; skin biopsy with direct extraction; remission sample; hair root or nail; and, if the proband cannot be tested and the variant is at least class 4 with VAF ≥40%, predictive testing of an at-risk relative [54]. Mesenchymal stromal cells cultured from marrow or from the cytogenetic culture flask are an alternative [44]. Whole blood is sufficient to test healthy relatives [53]. Gene-specific rules differ: in DDX41, DNA from cultured fibroblasts is mandatory during active neoplasia or after allo-HSCT, and saliva or buccal cells are excluded [9]; in CEBPA, GeneReviews accepts buccal swab, saliva, skin biopsy and blood or marrow in complete remission [8]; in RUNX1, a buccal swab is possible but carries a risk of contamination, and fibroblasts are preferred [4]. The ELN considers germline origin confirmed when the variant is present in DNA from a tissue with a low somatic mutation rate with a compatible VAF (usually 30%–60%), or in two or more relatives [48].

VAF alone is not a reliable indicator [54]. A VAF below 40% for a RUNX1 variant in the tumour and below 30% for a DDX41 variant suggest, without proving, somatic origin [4, 9]. Somatic RUNX1 variants arise in about 10% of haematological neoplasms, and CEBPA variants in leukaemic cells in about 9% of AML, few of them germline [4, 8]. In AML with biallelic CEBPA, 7%–15% have a germline variant, and 5 of 71 (7%) in a cohort of AML with mutated CEBPA, only 2 with a family history [44, 107]. The 2026 British guideline allows a presumption of probable germline origin, for diagnosis and immediate management, of a class LP/P DDX41 variant with VAF ≥35% accompanied by a second variant of lower VAF (90% agreement), and recommends that the somatic report state the possible germline origin when the VAF exceeds 40% (81%) [86]. In patients informed before marrow genetic profiling, a "tumour-only" profile with a near-heterozygous variant requires confirmation in fibroblasts, and in a "tumour-normal" profile the germline sample should be reanalysed when somatic DDX41 is detected [55]. For TP53, VAF should not be considered in isolation, and age, personal and family history and the features of the neoplasm should be weighed, with multidisciplinary discussion with genetics before testing [86].

5.5. Testing methods and limitations

The work-up should detect single-nucleotide variants and CNV, by exome, genome or extended panel sequencing with in silico CNV calling, microarray or MLPA (multiplex ligation-dependent probe amplification) [53]. Somatic myeloid panels include RUNX1, CEBPA, ETV6, GATA2 and DDX41 and may reveal incidental germline variants, but do not cover all germline regions, so a normal result does not exclude predisposition [53, 108]. The proportions of detectable variants by method, which are not equivalent to clinical sensitivity, are as follows:

Germline CNV cannot be called definitively from somatic data from diseased blood or marrow, and incidental constitutional sex chromosome aneuploidies may emerge [86]. Germline JAK2, MPL, SH2B3 and CSF3R variants are associated with myeloproliferative indices, and monoallelic loss of function in MPL or CSF3R may be an incidental carrier state for a recessive disease [86]. Functional assays are useful for VUS in Fanconi anaemia, telomeropathies and Diamond–Blackfan [53].

5.6. Reporting, classification and variants of uncertain significance

The report should state the gene, the versioned transcript, the HGVS nomenclature, zygosity, VAF, the tissue tested and the class according to the ACMG/AMP rules or the gene-specific ClinGen specifications [54, 109]. Only P/LP variants are considered causal and followed in families; VUS are frequently reclassified [48]. The British transplant guideline subdivides VUS into "hot" (81.2%–90% probability of pathogenicity) and "warm" (67.5%–81.2%), which may be reported and tested after discussion at a genomic panel, and lower categories that are neither reported nor tested [109]. The British consensus requires classifying the variant by the ACMG, CanVIG-UK or ClinGen rules before offering confirmatory testing (96%), confirming in the proband before cascade testing except in emergency or death (99%) and discussing any VUS testing in a multidisciplinary team with genetics (96%–97%) [54]. For DDX41, the Nordic group recommends reassessing the classification 2 years after the initial investigation [55]. DDX41 is not on the ACMG list of reportable secondary findings, and management of incidental findings follows institutional policy and national legislation [55]. Genetic counselling should be offered to everyone investigated, even when no variant is detected [53].

5.7. Practical algorithm

For a patient with MDS, AML or persistent cytopenia, the sequence proposed by the sources is: (1) a three-generation pedigree with verification of cancer and cytopenia diagnoses, targeted physical examination and bleeding history [48, 53, 55]; (2) informing the patient in advance about possible inherited findings before somatic profiling [54]; (3) review of somatic variants with near-heterozygous or near-homozygous VAF and of predisposition genes [53]; (4) germline testing on cultured fibroblasts when there is a clinical or somatic criterion, requested early so as not to delay donor selection [48, 109]; (5) extended germline investigation in a DDX41 carrier with MDS before 50 years, to exclude another predisposition [55]; (6) targeted testing of related donors and family cascade testing after counselling [109]. In a patient with inherited thrombocytopenia with normal-sized platelets, the panel should include RUNX1, ETV6, ANKRD26 (with the 5'-UTR) and the differential diagnosis genes, and a baseline bone marrow examination with cytogenetics and molecular profiling is recommended [6, 44]. In a child with MDS, the panel should include GATA2, SAMD9 and SAMD9L, and the investigation of bone marrow failure syndromes is completed with chromosomal fragility and telomere length [10, 73].

5.8. Prenatal and preimplantation diagnosis

Once the P/LP variant has been identified in the family, prenatal diagnosis and preimplantation genetic testing are possible in all the entities in the article [4, 7–9, 14, 15]. For DDX41, being an adult-onset disease, GeneReviews refers to the National Society of Genetic Counselors position on late-onset diseases, and the Nordic guidelines refer to national legislation [9, 55]. Monosomy 7 is not expected in most fetal tissues, and testing for it does not replace testing for the variant [73].

5.9. Presentation as bone marrow aplasia: the French PNDS pathway

When one of these entities presents as bone marrow aplasia or hypoplasia, the 2023 HAS PNDS (French National Diagnosis and Care Protocol) defines the investigation. The diagnosis of aplasia rests on the blood count with reticulocytes and film, bone marrow aspirate and trephine biopsy (cellularity below 30% for age), which is not systematic in constitutional forms in children, because it requires general anaesthesia; macrocytosis is frequent in constitutional forms, and dysplasia does not exclude a genetic cause [61]. The marrow karyotype is complemented by interphase FISH for chromosomes 7 and 8, useful above all when the karyotype fails in a poor marrow; del(20q), trisomy 8 and loss of Y remain compatible with aplasia, and monosomy 7 from diagnosis, especially in a young child, points to SAMD9/SAMD9L [61]. Testing for somatic variants by NGS may be performed, but an isolated somatic variant, without other cytological or cytogenetic evidence of a myeloid haemopathy, does not change management, and the test should not delay treatment [61].

All patients undergo testing for a PNH clone, fetal haemoglobin before transfusion, alpha-fetoprotein, immunoglobulins and lymphocyte subsets (at least CD3, CD4, CD8 and CD19), autoantibodies, liver assessment, parvovirus B19 serology or PCR, echocardiography, abdominopelvic ultrasound and chest radiography; HLA typing is urgent below 60 years and is extended to siblings if the patient is younger than 40 years [61]. Below 18 years, chromosomal fragility testing and telomere length measurement by FlowFISH are systematic. The constitutional bone marrow failure gene panel (annex 3 of the PNDS, which includes ANKRD26, RUNX1, DDX41, GATA2, SAMD9 and SAMD9L alongside the Fanconi and telomere genes, SBDS, SRP54, SRP72, MPL, THPO, MECOM, ERCC6L2, LIG4 and DNAJC21) is recommended systematically before 10 years and case by case thereafter, preferably on skin biopsy fibroblasts, because reversion or clonal evolution may make a blood result falsely negative [61]. Without an argument for a constitutional cause, the investigation should not delay immunosuppression or transplantation by more than 6 weeks after diagnosis in a child or young adult; if the classic tests are negative and suspicion persists, or if immunosuppression fails in a child, genome sequencing on fibroblast DNA within the Plan France Médecine Génomique (PFMG) is discussed at a national meeting [61].

In the SAMD9/SAMD9L syndromes, fetal haemoglobin may be raised, the bone marrow aspirate shows aplasia or hypoplasia, sometimes with the dysmegakaryopoiesis typical of monosomy 7, and the latter should also be sought by FISH; other clonal anomalies raise concern about leukaemic evolution and justify a somatic myeloid panel [61]. Germline gain-of-function variants are often classified as of uncertain significance (class 3); blood may show somatic variants of the same gene, usually in cis, which reflect genetic rescue, and the germline variant may disappear from blood through monosomy 7 or acquired uniparental isodisomy, which requires analysis of skin fibroblasts [61].

6. Differential Diagnosis

6.1. Comparison table

6.2. Thrombocytopenias with and without predisposition

The three thrombocytopenias with myeloid predisposition have normal-sized platelets, which separates them from the macrothrombocytopenias due to ACTN1 or TUBB1, which do not predispose to myeloid neoplasia [6]. CYCS also causes non-syndromic thrombocytopenia with normal platelets, without myeloid predisposition; SRC is associated with bone pathology and risk of marrow fibrosis; THPO, with semidominant inheritance, progresses to bone marrow failure [6]. Among the three, the platelet count may be normal in RUNX1, which has more bleeding due to dense granule deficiency; ETV6 is associated with macrocytosis and lymphoid predisposition; ANKRD26 has the lowest platelet counts [6, 44]. Immune thrombocytopenia is the most frequent misdiagnosis; intrafamilial variability and the transient rise in platelets during inflammation make molecular testing more accurate than the count for distinguishing acquired causes such as drugs, viral infections or pregnancy [6]. Short telomeres are frequent in germline RUNX1 and do not, on their own, distinguish it from a telomeropathy [78].

6.3. Myeloid predisposition with and without organ disease

DDX41 is distinguished from the other adult predispositions by the absence of extra-haematological features and by age, and GeneReviews lists as differentials the telomere biology disorders (dysplastic nails, reticulate pigmentation, leucoplakia, pulmonary or hepatic fibrosis), 14q32 duplication and Li-Fraumeni syndrome [9]. CEBPA familial AML is molecularly similar to sporadic AML with biallelic CEBPA, and only testing on non-leukaemic DNA separates the two [8]. In a child with MDS and monosomy 7, GATA2 deficiency accounts for 35%–40% of all predisposition to monosomy 7, the SAMD9/SAMD9L syndromes for 10%–20%, Fanconi anaemia for 5%–15%, neurofibromatosis 1 for 5%–10%, TP53 for 2%–5%, and CEBPA, DDX41, Shwachman–Diamond, severe congenital neutropenia, ERCC6L2 and Bloom for 1%–5% each [73]. Transient monosomy 7 with spontaneous regression in a young child has been described only in SAMD9 and SAMD9L [73, 104]. MIRAGE is distinguished from triple A syndrome (AAAS), IMAGe syndrome (CDKN1C) and IMAGe-I (POLE) by the myelodysplasia, and from GATA2 deficiency by adrenal hypoplasia, growth restriction and diarrhoea [14]. Ataxia-pancytopenia is distinguished from X-linked sideroblastic anaemia with ataxia (ABCB7), which has hypochromic microcytic anaemia without marrow failure, and from ataxia-telangiectasia [15].

6.4. Hypoplastic MDS, aplastic anaemia and bone marrow failure syndromes

Hypoplastic MDS with GATA2, DDX41, Fanconi or telomerase genes does not respond to immunosuppression, which separates it from the immune form [1]. Monosomy 7 arises in 3.5% of paediatric aplastic anaemias treated with immunosuppression in a North American series and in 5%–10% of adult aplastic anaemias [73]. Biallelic ERCC6L2 carries an MDS/AML risk of 33%, and the presence of multi-hit TP53 prompts consideration of early HSCT [44]. In Shwachman–Diamond syndrome, somatic EIF6 is compensatory and biallelic TP53 maladaptive, with clonal haematopoiesis in 70% by 20 years [75]. In Fanconi anaemia, 1q gains do not correlate with transformation, unlike monosomy 7, del(7q), 3q26 (EVI1) gains and RUNX1 alterations [44]. MBD4 (recessive, 7 AML and 1 MDS in 11 patients aged 30 to 49 years), SH2B3, TYK2, USP9X and GATA1 with short isoform are emerging predispositions [75]. The French PNDS (French National Diagnosis and Care Protocol) sets the principles common to constitutional aplasias: no indication for immunosuppression; discussion of allogeneic transplantation, with early genetic diagnosis when this changes the conditioning or the indication itself (as in the SAMD9/SAMD9L syndromes, in which transplantation is not systematic); genetic counselling with family studies for donor selection; multidisciplinary care of the extra-haematological manifestations; and marrow surveillance for clonal evolution [61]. In a young child, the differential diagnosis of the SAMD9/SAMD9L syndromes is mainly with congenital amegakaryocytosis, Hoyeraal–Hreidarsson-type telomeropathy and MECOM syndrome [61].

7. Treatment and Risk Management Strategies

7.1. General principles and quality of evidence

There is no treatment directed at the cause of any of these entities, and no drug has an approved indication for the germline predisposition itself; the associated neoplasms are treated with the standard regimens for the neoplasm, and allo-HSCT is the only curative treatment of the predisposed haematopoiesis [10, 44]. The GeneReviews for RUNX1, DDX41, ANKRD26, CEBPA, SAMD9L and MIRAGE state that no specific published clinical practice guidelines exist and base their recommendations on the authors' experience [4, 6, 8, 9, 14, 15]. The available guidelines (Nordic 2019 and 2024, ELN 2022, AACR 2017 and 2024, British consensus statements of 2023 and 2026) are expert consensus statements without formal grading of evidence and without randomised trials [48, 53–55, 75, 86, 109, 112]. Follow-up should take place in a haematology centre with experience in predisposition, in collaboration with clinical genetics [53]. The recommended initial evaluation includes a blood count with manual differential, reticulocytes and film, baseline bone marrow examination with cytogenetics and a somatic myeloid panel as a baseline for comparison, early HLA typing in the high-penetrance entities and genetic counselling [8, 15, 44, 53].

7.2. Thrombocytopenias with predisposition (RUNX1, ETV6, ANKRD26): haemostasis

Most ANKRD26 and ETV6 carriers are asymptomatic and are kept under observation [6, 7]. A baseline platelet aggregation study is recommended, especially when there is a bleeding history [7, 44]. For mild to moderate bleeding, moderate- or high-risk surgery, dental treatment and delivery, local measures, antifibrinolytics (tranexamic acid, aminocaproic acid) and desmopressin are used; platelet transfusion is reserved for severe bleeding or high-risk procedures and is not first line, because of the risk of alloimmunisation, with particular caution in women of childbearing age [4, 6, 7, 44, 113]. Surgeons, dentists and obstetricians should be warned in advance [10]. Eczema is treated with emollients and topical corticosteroids [4]. Thrombopoietin analogues have been used preoperatively and for up to 6 weeks, with long-term safety not established [6]; the theoretical concern arises from the mechanism of ANKRD26 thrombocytopenia, which involves increased TPO/MPL signalling [10, 79].

In the phase II trial of Zaninetti and colleagues (NCT02422394; April 2015 to May 2017), 24 patients with MYH9-RD (9), ANKRD26-RT (9), X-linked thrombocytopenia or Wiskott–Aldrich syndrome (3), monoallelic Bernard–Soulier (2) and ITGB3 thrombocytopenia (1), with mean baseline platelets of 40.4 × 10^9/L, received eltrombopag 50 mg/day for 3 weeks; those who did not reach platelets above 100 × 10^9/L by day 21 moved to 75 mg/day for a further 3 weeks [57]. Of the 23 evaluable patients, 21 responded (91.3%; 95% CI 72.0–98.9): major response (>100 × 10^9/L) in 11 (47.8%) and minor response (at least twice the baseline value) in 10 (43.5%). In ANKRD26-RT, 7 of 8 evaluable patients responded, mainly with minor responses and a mean increase of 41.8 × 10^9/L; all patients with ANKRD26-RT or XLT/WAS except one needed to increase to 75 mg/day, and the authors propose 75 mg/day as the initial preoperative dose in these two diseases. Four patients with spontaneous bleeding entered a second 16-week part, starting at 25 mg/day with adjustment every 4 weeks according to bleeding and platelets, and achieved remission of mucosal bleeding, two of them on 25 mg/day. Adverse events were grade 1 in 5 patients (transient headache or bone pain in the first 2 to 3 days in 4); one patient with ANKRD26-RT discontinued because of a rise in creatinine attributed to urinary retention due to prostatic hyperplasia. The authors observed no changes in the other lineages or in morphology in ANKRD26-RT, but advise caution in thrombocytopenias that predispose to haematological neoplasia, especially with prolonged administration [57]. The 2026 editorial on thrombopoietin receptor agonists in inherited thrombocytopenias reinforces that short perioperative use raises less concern about clonal evolution than chronic exposure, and describes 15 elective procedures in 9 patients (including ANKRD26) with a rise in platelets in 14 and no perioperative bleeding or transfusion [114].

In ANKRD26 thrombocytopenia, aspirin for primary or secondary cardiovascular prevention and anticoagulation, when indicated, may be used with platelets above 50 × 10^9/L, weighing risk and benefit in severe thrombocytopenia [6]. Non-steroidal anti-inflammatory drugs (NSAIDs) and antiplatelet agents as analgesics should be avoided, as should activities with a high risk of trauma, such as contact sports [4, 6, 7]. Ongoing trials in RUNX1-FPDMM without neoplasia (imatinib to increase RUNX1 activity; low-dose sirolimus; autologous CD34+ cell mobilisation) are described in subsection 7.11 [115–118].

7.3. Dosing of the drugs mentioned in the sources

Tranexamic acid is authorised in the European Union through national procedures whose summaries of product characteristics were not consulted, and its figures come from the US labels. Desmopressin also has national authorisations: in France, injectable Minirin® is indicated in moderate or mild haemophilia A (FVIII >5%), in von Willebrand disease except severe forms and type 2B, in unexplained prolonged bleeding time and in complications of antiplatelet agents, and nasal Octim® in mild haemophilia A, in von Willebrand disease except types 2B and 3 and in haemophilia A carriers at bleeding risk; inherited platelet dysfunction is not named in either [122, 123].

7.4. Comparative regulatory status

7.5. Treatment of neoplasms by entity

In neoplasms associated with RUNX1, ETV6 and ANKRD26, standard therapy for the neoplasm applies, with standard-conditioning HSCT after remission induction [7, 44]. In the NIH cohort, RUNX1-associated AML had outcomes comparable to AML with somatic RUNX1 (adverse risk in ELN 2022), and the authors recommend consolidation HSCT in all when possible, also to prevent second neoplasms [47]. GeneReviews considers that FPDMM is not curable with chemotherapy alone and that HSCT is almost always required [4]. There is no evidence that the ETV6 variant alters the response of ALL to standard therapy [7].

In CEBPA familial AML, management is the same as for sporadic CEBPA AML: induction with cytarabine and an anthracycline and consolidation with cytarabine, with or without HSCT according to clinical, cytogenetic and molecular risk, within a clinical trial whenever possible; relapses are treated with salvage chemotherapy followed by allo-HSCT [8, 148]. The initial evaluation includes HLA and serology (hepatitis A, B and C, cytomegalovirus, HIV), information on fertility preservation (oocyte cryopreservation is rarely feasible in newly diagnosed AML), a pregnancy test, lumbar puncture if there are neurological symptoms, cardiac assessment before anthracyclines and targeted testing of at-risk relatives [8]. The cytarabine and daunorubicin labels indicate them in combination for induction of acute non-lymphocytic leukaemia, and the liposomal formulation of daunorubicin and cytarabine (Vyxeos®) has an indication restricted to therapy-related AML or AML with myelodysplasia-related changes, without extension to CEBPA familial AML [10]. In ELN 2022, favourable prognosis is associated with in-frame bZIP insertions or deletions [44, 48].

In AML with DDX41, the ALFA-FILO series compared 86 patients treated with intensive chemotherapy in 5 trials with 1604 without DDX41: complete remission in 94% versus 69% (P <0.0001), relapse at 1 year of 15% versus 44% but at 3 years of 82% versus 75% (censoring at HSCT), and longer restricted mean survival (difference of 13.6 months at 5 years) [82]. ELN 2024 classifies mutated DDX41 as favourable risk in patients treated with less intensive therapy (hypomethylating agent monotherapy, hypomethylating agent with venetoclax, azacitidine with ivosidenib), with median overall survival above 24 months; 2-year survival was 60.1% with hypomethylating agent monotherapy and 91.1% with a hypomethylating agent and venetoclax, according to the studies cited by the ELN; DDX41 was germline in 61% in a guadecitabine trial (n = 604) [106]. This classification does not apply to patients previously treated with hypomethylating agents [106]. In MDS, the IPSS-M tends to assign a higher score to patients with DDX41 and did not discriminate prognosis in a retrospective study cited by the Nordic group, so it should be used with caution [55]; in the Makishima cohort, outcomes were not affected by the IPSS-R or the IPSS-M [49]. Lenalidomide produced responses in observational series [9, 66].

In GATA2 deficiency, chemotherapy alone has limited efficacy and high infectious toxicity [50]. Supportive management includes early HPV vaccination, mycobacterial prophylaxis with azithromycin, antimycobacterial treatment until HSCT and for about 1 year afterwards if there is active infection, immunoglobulin replacement if there is a humoral deficiency, gynaecological surveillance of HPV lesions, regular pulmonary function testing and avoidance of corticosteroids and immunosuppressants [10, 44]. In the SAMD9 and SAMD9L syndromes, support includes transfusions, antibiotic prophylaxis in severe neutropenia, immunoglobulin in hypogammaglobulinaemia, immunomodulators for skin inflammation (corticosteroids, JAK inhibitors, as SAMD9L is interferon-inducible), ataxia rehabilitation, speech therapy and feeding adaptation; there are no data for erythropoietin or granulocyte colony-stimulating factor (G-CSF), which may increase the risk of MDS/AML [15]. In MIRAGE, hydrocortisone and fludrocortisone, nutrition, an elemental diet for diarrhoea, duodenal feeding for recurrent aspiration pneumonia, gonadectomy of dysgenetic gonads, an interdisciplinary decision on sex of rearing, artificial tears and control of ambient temperature are added [14]. In MDS/AML with germline SAMD9L, high-dose cytarabine should be limited because of the risk of worsening ataxia [44].

The French PNDS (French National Diagnosis and Care Protocol) gives the SAMD9/SAMD9L syndromes their own management, because spontaneous correction of the aplasia is possible and because monosomy 7, usually an indication for urgent transplantation, may here be a transient event that restores haematopoiesis for a time; as there are no data on the proportion of patients who correct, nor criteria that predict correction or transformation, each case is discussed at a multidisciplinary meeting, and the search for a family or unrelated donor begins at diagnosis [61]. Surveillance combines blood counts (weekly or more often in a patient dependent on platelet transfusions, up to monthly in moderate pancytopenia), marrow cytology every 3 months while monosomy 7 is present, marrow karyotype and molecular study with a twofold aim: to detect somatic variants of leukaemic progression and to follow, by quantification, the revertant clones with somatic SAMD9 or SAMD9L variants [61]. Three courses are distinguished. In rapid improvement of the counts, which may be slow and incomplete and with monosomy 7 persisting for more than a year, continuous improvement, the absence of new chromosomal abnormalities, a fall of monosomy 7 on FISH, the absence of myeloid variants and revertant clones with rising VAF are reassuring. Signs of leukaemic transformation indicate allogeneic transplantation. In the intermediate situation, with neither correction nor worrying signs, transplantation is discussed at a meeting, especially if pancytopenia is severe, with transfusion dependence and infectious risk [61]. Support is that of any constitutional aplasia, with transfusion sparing to avoid alloimmunisation; growth factors are not used to correct the cytopenias, although G-CSF may be used occasionally in a severe infection, and thrombopoietin receptor agonists are not recommended, for lack of data [61].

In the initial extra-haematological evaluation of the SAMD9/SAMD9L syndromes, the PNDS requires at least abdominopelvic ultrasound (search for adrenal hypoplasia, of low diagnostic value), echocardiography, endocrine assessment with testing for adrenal insufficiency as soon as the syndrome is suspected, because of the risk of acute crisis, neurological and neurosensory assessment with brain magnetic resonance imaging and psychological assessment [61]. In children, paediatric endocrine follow-up is systematic, because of short stature of syndromic or iatrogenic cause (corticosteroids, conditioning), delayed puberty, which may require hormone replacement, and the possible indication for growth hormone; the enteropathy may require parenteral nutrition, and neurological follow-up continues into adulthood, because of the risk of late-onset ataxia or pseudo-parkinsonian syndrome. Hypogammaglobulinaemia in adults is managed according to the recommendations for common variable immunodeficiency [61].

7.6. Allogeneic transplantation: indication and timing

The Nordic guidelines regard all patients of appropriate age with a myeloid neoplasm arising on a germline predisposition as potential allo-HSCT candidates, except AML with germline CEBPA, and recommend discussing each case at a transplant panel, including international experts [53]. In predispositions without pre-existing disease, there is no absolute indication for HSCT in first complete remission (CR1) in CEBPA AML, because remissions with conventional chemotherapy are long, but the risk of new leukaemic clones allows it to be considered later; in the thrombocytopenias with predisposition there is no indication for HSCT until a myeloid neoplasm appears; in organ dysfunction (GATA2), HSCT may be considered before transformation, to avoid the mortality of the other manifestations [53]. The table summarises the positions by entity and the divergences.

In GATA2 deficiency, allo-HSCT eradicates the clone, restores haematopoiesis, eliminates the underlying infection and improves alveolar proteinosis, but does not reverse the extra-haematopoietic manifestations, and HPV may persist [10]. The indications described include recurrent, disfiguring or severe infections, organ dysfunction, MDS with cytogenetic abnormalities, high-risk somatic mutations, transfusion dependence and myeloid progression; a myeloablative regimen with busulfan and post-transplant cyclophosphamide in 59 patients achieved overall survival of 85% and event-free survival of 82%, according to the GeneGuidance synthesis of Rajput and Arnold [10, 150]. In the EWOG-MDS cohort, GATA2 deficiency had no impact on overall survival or HSCT outcome, but the high risk of progression justifies timely HSCT [45]. In the paediatric SAMD9/SAMD9L syndromes, 5-year overall survival was 84% (SAMD9) and 93% (SAMD9L) in 42 patients, dependent on karyotype (95% with normal karyotype versus 76% with monosomy 7), and 93% versus 77% in the 29 transplanted [58]. Of the 41 published SAMD9L patients who underwent HSCT, 36 survived, followed for up to 16 years; haematological reconstitution does not improve the neurological or inflammatory manifestations [15]. In MIRAGE, of 8 transplanted patients, all had severe complications (adrenal crisis, systemic infections, hepatic veno-occlusive disease, respiratory failure) and 3 died, so non-curative supportive care may be preferable in some cases [14, 151, 152].

7.7. Selection of a related donor

The inadvertent use of carrier related donors has led to donor-derived MDS/AML, graft failure and poor outcomes [2, 53]. ELN 2022 considers the use of cells from donors carrying deleterious RUNX1 and CEBPA variants "prohibitive", with a lack of data for most other genes [48]. In all autosomal dominant predispositions, a donor without the variant is strongly recommended, and all potential related donors should be tested during eligibility screening [44]. The European Society for Blood and Marrow Transplantation (EBMT) survey estimated donor-derived leukaemia at 80.5 per 100,000 transplants, with a cumulative incidence of 0.067%, 0.132% and 0.363% at 5, 10 and 25 years, probably underestimated, according to the British transplant guideline [109].

The 2023 British guideline (UK Cancer Genetics Group, CanGene-CanVar, NHS England and British Society of Blood and Marrow Transplantation and Cellular Therapy) reached consensus on 14 of 16 statements [109]. All patients who need HSCT should be assessed for an inherited cause (100%); when one is suspected, the unrelated donor search and confirmatory donor testing proceed in parallel with the assessment of related donors (92%); potential donors for a proband with a P/LP variant should have urgent access to counselling (96%); donors may decline testing (100%); urgent donor testing may proceed in parallel with germline confirmation in the proband (83%); matched relatives without the variant take priority over carriers (87%); a carrier is not usually considered as a donor, except when options are limited (93%); if all matched relatives are carriers or decline testing, the choice between an unrelated donor and a carrier is made by the multidisciplinary team gene by gene (95%–98%); and identified carriers have a rapid appointment with clinical and psychosocial support (93%) [109]. The guideline recommends separating donor and recipient care, according to the JACIE, ISCT-Europe and EBMT standards, and the Human Tissue Act for donors without capacity, including minors, and the creation of a national database of donor leukaemia and graft failure when carriers are used; the safety of G-CSF mobilisation in an occult carrier is unknown [109]. Unrelated donors are screened only by questionnaire, and there is some evidence of over-representation of predisposition variants in cord blood transplants [109].

7.8. DDX41: graft-versus-host disease and donor-derived leukaemia

Data on GVHD after allo-HSCT in DDX41 carriers diverge. In the retrospective cohort from Chicago, Adelaide, Pennsylvania and Moffitt (472 patients with myeloid neoplasms tested, 161 transplanted), stage 3–4 acute GVHD occurred in 38% of DDX41 carriers versus 0% (CHEK2), 12% (other P/LP) and 9% (no variant) (P = 0.002), with severe gastrointestinal GVHD in one-third; DDX41 was an independent predictor of acute GVHD of any stage (OR 4.76; 95% CI 1.27–17.9) and of stage II–IV (OR 4.84; 1.54–15.1). With post-transplant cyclophosphamide, 0 of 5 had severe acute GVHD versus 8 of 15 (53%) with other prophylaxis (P = 0.03); without it, 7 of 15 died, 4 from severe GVHD; only 1 donor shared the familial variant [87]. On the basis of these data, Trottier and colleagues recommend post-transplant cyclophosphamide in DDX41 carriers, even with a donor without the variant [44]. In the French cohort of the Société Francophone de Greffe de Moelle et de Thérapie Cellulaire (SFGM-TC; 83 patients, HSCT between 2007 and 2022, median age at HSCT 61.6 years, 84.3% men, reduced-intensity conditioning in 74.7%), 2-year overall survival was 73.9% (95% CI 62.8–82.1), leukaemia-free survival 68.9%, relapse 10.2% and non-relapse mortality 21%; grade 3–4 acute GVHD at day 100 was 14.6% and extensive chronic GVHD at 2 years 9.6% [156]. In the matched analysis of AML in CR1 (32 with DDX41 versus 77 without), 6-year overall survival was 56.8% versus 54.2% (HR 0.96; 0.45–1.93), with no excess GVHD, but with a median time to relapse of 50.1 versus 7.5 months, reflecting late relapses [156]. The authors attribute the divergence from Saygin to the higher proportion of HSCT in CR1 in the French cohort; both studies are retrospective, with non-uniform conditioning and prophylaxis [87, 156].

In the French cohort, 7 of the 11 patients who relapsed after HSCT from a related donor had persistent 100% or 95% donor chimerism, that is, donor-derived leukaemia; all received grafts from an HLA-identical male sibling with a normal blood count at donation, and the DDX41 variant was confirmed in the blood of 5 of the 7 donors; the median time to donor leukaemia was 86 months (6–116). Donor leukaemia occurred in 8.4% of the cohort and in 19.4% of recipients of related-donor grafts; two of the donors followed developed MDS 11 and 4 years after donation [156]. The authors recommend screening for DDX41 in all patients with MDS/AML who are candidates for related-donor HSCT and testing donors before donation [156]. Similar cases were described by Berger and by Kobayashi, and the case in section 10 illustrates the sequence [89, 157, 158].

7.9. Conditioning and post-transplant follow-up

In Fanconi anaemia and the telomeropathies, increased sensitivity to chemotherapy requires specific reduced-dose conditioning; in Fanconi anaemia total body irradiation is avoided, because of the risk of secondary solid tumours; in GATA2 deficiency before transformation, reduced-intensity conditioning is possible [44, 53]. Several monosomy 7 predisposition syndromes have increased morbidity from prolonged cytopenias after intensive chemotherapy or prior immunodeficiency, so direct HSCT without cytoreductive chemotherapy is the standard approach in many [73, 159]. An unrecognised inherited bone marrow failure syndrome is associated with worse HSCT outcomes, so genetic diagnosis should precede transplantation [44]. Allo-HSCT increases the risk of secondary tumours, and surveillance should be closer in the subgroup with organ dysfunction [53]. In childhood AML with monosomy 7, high-risk protocols with HSCT from the best donor in CR1 are the rule [73, 160].

In the SAMD9/SAMD9L syndromes, the French PNDS (French National Diagnosis and Care Protocol) considers transplant outcomes similar to the usual ones, except in children with severe extra-haematological manifestations (MIRAGE), with no special sensitivity to conditioning or increased risk of GVHD described; it recommends standard myeloablative or reduced-toxicity conditioning (busulfan with fludarabine, or thiotepa, treosulfan and fludarabine, always with serotherapy; doses in the table in subsection 7.3), without total body irradiation, and reserves reduced-intensity conditioning, discussed at a national meeting, for the severe comorbidities of MIRAGE and for prolonged severe aplasia without myelodysplasia [61]. GVHD prophylaxis combines a calcineurin inhibitor with short-course methotrexate, or with mycophenolate mofetil with cord blood; bone marrow is the preferred source, to reduce extensive chronic GVHD, a rule that the PNDS extends to all bone marrow aplasias, as there is no benefit from a graft-versus-leukaemia effect; alternative donors (unrelated with compatibility below 10/10 or haploidentical) are discussed at a meeting, especially in MDS, acute transformation or prolonged severe aplasia without a tendency to correction [61]. No spontaneous excess of solid tumours has been described in these syndromes, but the risk may exist in transplanted patients [61].

7.10. Trials and studies with results

The studies in the table do not allow indirect comparison: the eltrombopag trial is single-arm, with a predominance of diseases without myeloid predisposition, and does not assess long-term clonal risk; the briquilimab trial has two arms with different donors and only 2 evaluable patients in arm B, and the survival and GVHD outcomes have not yet been published; the DDX41 studies are retrospective or derived from trials not designed for the question, with different populations and prophylaxis, and the relative reductions in relapse risk should be read together with the long-term absolute rates, which converge [57, 82, 87, 129, 156]. No randomised trials or published negative trials were found in this area; their absence from the searches does not show that none exist.

7.11. Developments in the past 12 months and research

Still ongoing are the National Human Genome Research Institute RUNX1 natural history study (NCT03854318, since March 2019, target of 1000 participants), the NIAID GATA2 deficiency natural history study (NCT01905826, n = 600), the NCI phase II allo-HSCT trial in GATA2/MonoMAC (NCT01861106, n = 144, with fludarabine 40 mg/m² from days −6 to −3 or 30 mg/m² from days −6 to −2 and busulfan with a test dose of 0.8 mg/kg) and the autologous CD34+ mobilisation study in RUNX1-FPD (NCT06414889, n = 4) [118, 167–169]. Briquilimab, imatinib and sirolimus are experimental in this setting.

7.12. Agents and exposures to avoid

Carriers of thrombocytopenia with predisposition should avoid NSAIDs and antiplatelet agents as analgesics and contact sports [4, 6, 7]. In SAMD9L syndrome, NSAIDs, anticoagulants and thrombolytics are contraindicated in severe thrombocytopenia, and alcohol and centrally acting sedatives worsen gait and coordination [15]. In GATA2 deficiency, corticosteroids and immunosuppressants should be avoided [10]. Smoking, chemical exposure, unnecessary radiation and, in FPDMM, obesity may increase the risk of neoplasia [4, 9]. There are no data on erythropoietin or G-CSF in the SAMD9/SAMD9L syndromes, and their use may increase the risk of MDS/AML [15]. A carrier related donor should be avoided in all entities (subsection 7.7).

7.13. Pregnancy, delivery and the perioperative period

In FPDMM, platelets and bleeding should be monitored in pregnancy; severe bleeding is possible before, during or after delivery, mild thrombocytopenia without an aggregation defect is unlikely to affect pregnancy, and a count below 50 × 10^9/L may limit neuraxial anaesthesia, with an individual decision on transfusion among obstetrician, haematologist and anaesthetist [4]. The ANKRD26 GeneReview cites a threshold of 70 × 10^9/L for epidural analgesia, attributed to the 2021 Society for Obstetric Anesthesia and Perinatology (SOAP) consensus, and allows transfusion or thrombopoietin agonists case by case [6]; that consensus, however, expressly excludes inherited thrombocytopenias from its scope, so the threshold has not been validated for these diseases [170]. In ETV6, platelets should be monitored in pregnancy and before delivery, with transfusion before invasive procedures (epidural, caesarean section) case by case [7]. For delivery and surgery, Trottier recommends antifibrinolytics and/or desmopressin and transfusion only for severe bleeding [44]. In DDX41, onset is usually after reproductive age; with a family history of early onset, a blood count is performed before pregnancy, and new abnormalities in pregnancy require urgent referral [9]. In SAMD9L, anaemia, thrombocytopenia or neutropenia may increase the risk of complications, and there is no information on the effect of pregnancy [15]. In children with RUNX1, the baseline bone marrow examination may be postponed until about 3 years to avoid anaesthesia, unless there is clinical suspicion, and bleeding has neonatal, perioperative, trauma and suspected maltreatment implications [47]. In the SAMD9/SAMD9L syndromes, surgery needed for organ manifestations should, whenever possible, be performed before thrombocytopenia worsens [61]. Before allogeneic transplantation for constitutional bone marrow aplasia, the French PNDS (French National Diagnosis and Care Protocol) provides for a reproductive medicine consultation for fertility preservation, even with mildly gonadotoxic conditioning, weighing the bleeding and infectious risk of the procedures and the delay to transplantation: before puberty only cryopreservation of gonadal tissue is possible, and in adolescents and adults sperm cryopreservation or oocyte vitrification after stimulation is used, which thrombocytopenia or neutropenia may contraindicate [61]. Hypogonadism in the SAMD9/SAMD9L syndromes, urogenital malformations and transplant sequelae reduce fertility, and these patients should have access to an assisted reproduction clinic [61].

7.14. Transition to adult care

Transition should be planned, with surveillance maintained, and adolescents and young adults with RUNX1-FPD benefit from psychological support, peer groups and genetic counselling [75, 97]. In CEBPA AML, adolescents and young adults should be followed by dedicated teams [8]. For bone marrow aplasias, the French PNDS (French National Diagnosis and Care Protocol) proposes a three-phase transition programme: preparation in paediatrics, with empowerment of the adolescent, part of the consultation without the parents, early choice of the adult physician and a written summary of the record; transfer, if possible with a joint consultation and alternating consultations in the first year, support from a liaison nurse and a flexible age between 16 and 20 years, during a stable phase of the disease; and reception in the adult department with an extended consultation [61]. Therapeutic education includes the reference centre's booklets for patients and families (updated in 2022), the national post-transplant follow-up booklet of the Société Francophone de Greffe de Moelle et de Thérapie Cellulaire (SFGM-TC), nursing and clinical pharmacy consultations and the days organised by the associations [61].

7.15. French PNDS for bone marrow aplasias: position and divergences

The HAS PNDS (French National Diagnosis and Care Protocol) on acquired and constitutional bone marrow aplasias (version of 1 March 2023), prepared by the French reference centre and the MaRIH filière, covers the entities in this article only when they present as constitutional bone marrow aplasia and devotes a dedicated annex to the SAMD9/SAMD9L syndromes [61]. There are seven divergences from the other sources. (1) In the SAMD9/SAMD9L syndromes, the PNDS sets neither an age nor a waiting period, unlike the 5-year threshold for surveillance of monosomy 7 used by Sahoo, GeneReviews and Trottier, and bases the decision on the course of the counts, the clone and the revertant clones [14, 44, 58, 61]. (2) With monosomy 7, marrow cytology is quarterly, at the lower limit of Sahoo's 3–6 months and closer than the annual bone marrow examination of the other sources; after normalisation, the PNDS proposes a blood count every 3 months and marrow every 12–18 months, against a blood count every 4–6 months (Trottier, MIRAGE GeneReview) or 6–12 months (SAMD9L GeneReview) [14, 15, 44, 58, 61, 75]. (3) The PNDS allows occasional G-CSF in severe infection, whereas the SAMD9L GeneReview notes the lack of data and a possible risk of MDS/AML with growth factors [15, 61]. (4) The PNDS attributes the autoinflammatory disease with excessive interferon response to "acquired" gain-of-function SAMD9L variants, whereas GeneReviews describes it as caused by de novo germline variants; the distinction between de novo and somatic changes counselling, and this article follows GeneReviews [15, 61]. (5) The cumulative incidence of MDS/AML in GATA2 deficiency is given as about 70% by 70 years, below Trottier's about 75% and the AACR's at least 80%, a difference that reflects different cohorts and outcomes [44, 61, 75]. (6) In the conditioning doses, the PNDS fludarabine (40 mg/m²/day over 4 days, 160 mg/m²) exceeds that of the regimen labelled with treosulfan (30 mg/m²/day over 5 days, 150 mg/m²); the treosulfan body surface area cut-offs cited by the PNDS (0.4 and 1 m²) differ from those of the current SmPC (0.4 and 0.9 m²), and the transplant section of the same document, in reproducing the SmPC, writes the doses in g/kg/day (10, 12 and 14 g/kg/day) and the first cut-off as 0.5 m², whereas the SmPC and the SAMD9/SAMD9L annex give them in g/m²/day; thiotepa is expressed as 8 mg/m², a value with no counterpart in the SmPC, whereas paediatric treosulfan, fludarabine and thiotepa regimens use thiotepa in mg/kg (for example, 2 × 5 mg/kg on D-2, total 10 mg/kg). For busulfan and fludarabine, the PNDS dose is the reference and the SmPC dose the comparison; for treosulfan the unit (g/m²/day) and cut-offs of the current SmPC apply, and for thiotepa the mg/kg dose of the centre's protocol applies, within the SmPC maxima [61, 144–146]. (7) The PNDS states that treosulfan is not authorised in non-malignant paediatric disease, whereas the current European SmPC covers children older than 1 month in malignant and non-malignant disease [61, 144].

8. Reproductive Options and Family Counselling

8.1. Inheritance and recurrence risks

All the entities in the article are autosomal dominant, and each child of a carrier has a 50% risk of inheriting the variant [4, 6–9, 14, 15]. The proportion of de novo cases differs: it is unknown in FPDMM and ETV6; in ANKRD26, the vast majority are inherited, with a single "de novo" case reported without confirmation of parentage; in CEBPA, the de novo form is rare; in DDX41, all probands with tested parents had inherited the variant at the time of GeneReviews; in GATA2 deficiency, up to 80% of cases are de novo; in the SAMD9 and SAMD9L syndromes, the de novo form is frequent, and all cases of SAMD9L-associated autoinflammatory disease described were de novo [4, 6–9, 14, 15, 44].

When the variant is not detected in the parents, the risk to siblings is estimated at about 1% (germline mosaicism) for DDX41, ETV6, ANKRD26 and CEBPA, and slightly above the population risk in the SAMD9 and SAMD9L syndromes [6–9, 14, 15]. In FPDMM and in the SAMD9/SAMD9L and monosomy 7 predisposition syndromes, a negative result in a parent tested on blood may be false, owing to somatic loss of heterozygosity (UPD 21 in RUNX1; UPD 7q or monosomy 7 in SAMD9/SAMD9L), so parents should be tested on non-haematopoietic DNA [4, 10, 15, 73]. Parents with normal platelet counts on several occasions have a low risk of being carriers of ANKRD26, whose penetrance for thrombocytopenia is complete [6]. In DDX41, if the partner is also a carrier, each pregnancy has a 25% probability of a biallelic state, with intellectual disability, developmental delay and haematological neoplasm described, and partners may consider testing [9]. Individuals with typical MIRAGE have not reproduced, because of severity or gonadal dysfunction [14]. The French PNDS (French National Diagnosis and Care Protocol) stresses that family studies may reveal in a carrier parent a suggestive phenotype or pauciclonal haematopoiesis, a sign of a past aplasia, recognised or not, that has corrected; complete correction by acquired uniparental isodisomy may hide the variant in the parent's blood and lead to the index case's variant being wrongly considered de novo [61].

The risk of neoplasia for a carrier sibling depends on the entity: 20%–50% in FPDMM, with marked intrafamilial variability; 50% to more than 80% in CEBPA, depending on the location of the variant; reduced and sex-influenced penetrance in DDX41; expected mild thrombocytopenia and leukaemia of reduced penetrance in ETV6; thrombocytopenia of complete penetrance with variable severity and risk of transformation in ANKRD26 [4, 6–9]. In counselling on DDX41, the Nordic group recommends conveying that it is a moderate-risk gene, with a penetrance for myeloid neoplasia of up to about 10% at 60 years and about 50% at 90, higher in men, and with a typical age at onset at the end of the seventh decade; truncating variants may have higher penetrance, but the recommended counselling is the same for all types [49, 55].

8.2. Cascade testing and testing of minors

At-risk asymptomatic relatives should be tested to start surveillance and, when they are potential donors, to guide donor selection [4, 6–8]. Predictive testing requires prior counselling by clinical genetics, with a rapid timeframe for potential donors (95%–96% consensus in the United Kingdom), and only P/LP variants are used for predictive testing [54, 55]. A family member with AML should not be assumed to be a carrier without testing [8]. Without an identified variant in a family with several affected members, all first-degree relatives should have a blood count, repeated annually if normal, and bone marrow examination with FISH or cytogenetics if there are cytopenias or macrocytosis [73]. Relatives with a negative test on non-haematological DNA are no longer at increased risk and are discharged from surveillance [73].

The age for predictive testing differs between entities. In DDX41, predictive testing is not recommended before 18 years, except when there are childhood-onset neoplasms in the family (very rare) or when the minor is a potential donor; the Nordic group offers testing to adult first-degree relatives from 50 years or 10 years before the earliest case in the family, with caution in second-degree relatives, and the British consensus does not test before adulthood (96%) [9, 54, 55]. In CEBPA, ANKRD26, ETV6, GATA2 and RUNX1, testing of children is individualised by genotype and family history (81%), with counselling by clinical genetics (93%) [54]. In RUNX1, the NIH recommends universal and early cascade testing, regardless of age and symptoms, because of the implications of bleeding in the newborn, in the perioperative period, in trauma and in suspected maltreatment [47]. In ANKRD26, GeneReviews weighs the autonomy of the minor against the risk of unrecognised thrombocytopenia, with surgical delay, misdiagnosis of immune thrombocytopenia and unnecessary treatments, and proposes a baseline blood count to support shared decision-making [6]. In CEBPA, testing of minors presupposes prior education, assent of older children and adolescents and an agreed mode of communication [8]. In SAMD9L, testing of minors is appropriate to guide management [15]. The AACR proposes cascade testing at the age at which risk increases, even without signs, and deferring DDX41 until adulthood [75].

8.3. Prenatal and preimplantation diagnosis

Prenatal diagnosis and preimplantation genetic testing are possible after identification of the familial variant, and referral should preferably take place before pregnancy, with information on the risk of transmission [4, 8, 53]. In adult-onset diseases (DDX41), these options follow national legislation and the position of the National Society of Genetic Counselors, and DDX41 testing for family planning is not recommended in people with MDS/AML without a molecular diagnosis of DDX41 [9, 55]. Storage of DNA from probands without a molecular diagnosis is recommended, for future analysis [73]. In the SAMD9/SAMD9L syndromes, the French PNDS (French National Diagnosis and Care Protocol) considers genetic counselling indispensable before assisted reproduction with preimplantation diagnosis, which presupposes a variant classified as pathogenic and analysable in the DNA of a single embryo cell; double selection of an unaffected embryo that is also HLA-compatible with an affected sibling is no longer practised in France [61].

8.4. Counselling content and psychosocial support

Counselling should include a three-generation pedigree with verification of diagnoses, the risk of transmission, penetrance and the uncertainty of the figures, warning symptoms (fatigue, fever, weight loss, infections, bleeding, skin changes), the surveillance plan and the implications for stem cell donation [9, 54, 55]. Donors should know that they may decline testing and decide when to learn the result; dedicated genetic counselling is the standard, and, where it is not available, it should be provided by trained haematologists external to the transplant programme [109]. All P/LP carriers should be referred to clinical genetics and encouraged to take part in registries [54, 55]. In England, the National Inherited Cancer Predisposition Register, launched on 1 July 2025, includes about 120 predisposition genes and receives enrolments through clinical genetics [86]. Psychological distress in young adults with RUNX1-FPD justifies psychological support and peer groups [97].

9. Special Considerations: Population Diversity and Epidemiology

9.1. What the available figures measure

Frequency data for these entities come from three types of study, which are not interchangeable. Hospital series of tested patients with MDS/AML measure the fraction of neoplasms attributable to a germline variant, influenced by the selection criteria; family cohorts measure penetrance in ascertained carriers, with a bias towards families with many affected members; and genome-first studies in population cohorts measure carrier frequency and risk without clinical ascertainment [15, 42]. The Nordic guidelines estimate that 5%–15% of adults and 4%–13% of children with MDS or AML have germline variants in susceptibility genes, and Hendricks cites about 10% of myeloid neoplasms with germline susceptibility [42, 53]. In MDS, P/LP variants occur in 25% of paediatric patients and in 19% of patients aged 18 to 40 years, according to the studies cited by Saygin [87]. In children, P/LP germline variants are found in about 15%–30% of MDS or persistent cytopenias and in 2%–19% of paediatric AML [75].

9.2. United States

In the Chicago cohort (2010–2021), a deleterious germline variant was found in 28% of patients with myeloid neoplasms tested because of clinical suspicion (80 of 289), with CHEK2 as the most frequent gene, in relation to enrichment for Eastern European ancestry; the median age at diagnosis was 67 years in DDX41, 62 in CHEK2 and 61 for the other P/LP variants [87]. In a published cohort of 391 unselected AML, P/LP CHEK2 was present in 2.0% and DDX41 in 1.8% [87]. In the genome-first approach (DiscovEHR, n = 170,503; UK Biobank, n = 469,595) for ANKRD26, CEBPA, DDX41, MECOM, SRP72, ETV6, RUNX1 and GATA2, heterozygotes had a risk of myeloid neoplasm with an odds ratio of 4.6 (95% CI 2.1–9.7) and 6.0 (4.3–8.2), and worse overall survival (HR 1.8 and 1.4); DDX41 was the most frequent gene, with an OR of 5.7 (3.9–8.3) in the UK Biobank [42]. The prevalence of ETV6 was estimated at 2–4 per 100,000 adults [7, 42]. In the United States, about 10,000 MDS and 13,300 AML are diagnosed per year according to the 2021 DDX41 GeneReview, and about 22,720 AML per year according to the 2026 CEBPA GeneReview [8, 9]. The NIH RUNX1 cohort analysed 213 P/LP carriers [47].

9.3. Europe and populations with founder variants

In Europe, DDX41 founder variants have a geographical distribution: c.3G>A p.(Met1?) and c.415_418dup p.(Asp140GlyfsTer2) are the most frequent in North America, the United Kingdom and France, c.121C>T p.(Gln41Ter) in Germany and c.653G>A p.(Gly218Asp) in Italy, whereas c.1496dup p.(Ala500CysfsTer9) predominates in Japan, Korea and China and c.62_63del p.(Ser21ThrfsTer7) in Thailand [9, 44]. In the French ALFA-FILO series, germline DDX41 was present in 5% of intensively treated AML [82], and in the French family study the risk at 70 years was 19.5% [102]. In Japan, DDX41 risk alleles were 10 times more frequent in patients (n = 4461) than in the general population (n = 20,238), and DDX41 accounted for about 80% of the known germline predisposition to adult myeloid neoplasms [49]. In Adelaide (Australia), DDX41 was the most frequent P/LP gene (20% deleterious variants in 175 patients tested) [87].

The German EWOG-MDS cohort prospectively studied 426 primary and 82 secondary MDS in children and adolescents over 15 years: GATA2 was present in 15% of advanced MDS, 7% of all primary MDS, 37% of MDS with monosomy 7 and 72% of adolescents with monosomy 7, and absent in therapy-related MDS and acquired aplastic anaemia [45]. In the same network, SAMD9/SAMD9L were present in 8% of 669 paediatric MDS [46]. In large multicentre cohorts of children and young adults, monosomy 7 occurs in about 20% of MDS and 5% of AML, up to 41% in single-centre cohorts of primary paediatric MDS, with an identifiable germline predisposition in most [73]. In France and Belgium, the series of 79 patients with GATA2 deficiency from 53 families described the natural history [50], and an Italian national study identified new symptoms and genotype–phenotype associations, in a disease with about 500 patients described worldwide between 2011 and 2023 [10, 171]. In Italy, THC2 accounted for 21 of 210 families in the inherited thrombocytopenia database, and ANKRD26 for 15% of index cases, with an incidence of inherited thrombocytopenia estimated at about 2.7 per 100,000 [41, 65]. Spinocerebellar ataxia 49 was described in a family from Menorca [15]. In the United Kingdom, germline DDX41 is present in more than 3% of adult MDS/AML [86].

9.4. Portugal and Brazil

The sources consulted do not describe series from Portugal or Brazil for these entities. The only reference to the Brazilian population is the lower-penetrance TP53 founder variant in southern Brazil, mentioned by the 2026 British guideline on the basis of another source [86]. The absence of data in these sources does not demonstrate the absence of published data.

9.5. Sex, age and special populations

Male predominance is characteristic of DDX41: 75% of patients in the ALFA-FILO series and 84.3% of French transplant recipients were men, and the odds ratio for male sex was 2.45 in the Makishima cohort, with no difference among healthy carriers, which indicates higher penetrance in men and not a higher frequency of the variant [49, 82, 156]. In children, the GATA2 and SAMD9/SAMD9L syndromes are the most frequent germline causes of MDS, and ANKRD26 and DDX41 rarely cause neoplasia before adulthood [45, 46, 75]. In adult survivors of childhood cancer (more than 12,000), P/LP variants were found in 4.4% of those who had AML, 3.3% ALL, 2.9% non-Hodgkin lymphoma and 2% Hodgkin lymphoma [75]. Up to 20% of presumed therapy-related neoplasms may have a germline predisposition, and mutations in PPM1D and in DNA damage response genes may require germline evaluation [1, 73].

9.6. European networks and patient organisations

ERN EuroBloodNet (rare haematological diseases) is the European Reference Network for these entities, and ERN RITA (immunodeficiency, autoinflammatory and autoimmune diseases) covers the immunodeficiency of GATA2 deficiency and the autoinflammation of SAMD9L [59, 60]. Patient organisations and resources include the RUNX1 Research Program, founded by a patient family and a partner in the NIH natural history study, the Platelet Disorder Support Association, the Aplastic Anemia & MDS International Foundation, the MDS Foundation, the Leukemia & Lymphoma Society, Leukaemia Care, MPN Voice and MDS-UK in the United Kingdom, and the National Ataxia Foundation for SAMD9L ataxia [4, 9, 15, 86, 172–174]. No organisation specific to GATA2 deficiency was found; the German Krebs-Prädisposition portal suggests the registry of cancer predisposition syndromes [175].

In France, constitutional bone marrow aplasias, including the aplastic presentations of GATA2 deficiency and the SAMD9/SAMD9L syndromes, are followed by the reference centre for acquired and constitutional bone marrow aplasias (Hôpital Saint-Louis, Paris), of the MaRIH rare immuno-haematological diseases filière, with a national multidisciplinary meeting for treatment and transplant decisions [61]. Genetic panels are sent to reference laboratories (aplasias at Hôpital Saint-Louis, telomeropathies at Hôpital Bichat, Diamond–Blackfan and telomere length at Hôpital Robert-Debré), and genome sequencing is performed on the PFMG platforms; bone marrow aplasia is eligible for the long-term condition scheme (ALD), and patients receive an emergency letter prepared by the reference centre with the filière [61]. The PNDS lists the codes ORPHA:228423, ORPHA:3226 and ORPHA:319465 for GATA2 syndrome and ORPHA:494433 for MIRAGE, with no code of its own, to date, for SAMD9/SAMD9L syndrome as a whole, and includes the association HPN France – Aplasie médullaire among the patient resources [61].

10. Illustrative Clinical Case

10.1. Donor cell leukaemia from germline DDX41 shared with the sibling donor

Rolles and colleagues (RWTH Aachen, Germany) described a Caucasian man who, at 59 years, had MDS with excess blasts (EB2), neutropenia, thrombocytopenia and multiple infections [89]. He received, without prior induction, conditioning with fludarabine and busulfan and allo-HSCT from his HLA-identical brother. Engraftment was timely and donor chimerism complete, but mild leucopenia and thrombocytopenia persisted (CTCAE grade 2–3). Ten years later, thrombocytopenia progressed from grade 3 to grade 4, 1% blasts appeared in the blood, and the marrow showed AML with 22% blasts, normal karyotype and 100% donor chimerism, that is, donor cell leukaemia. The DNMT3A variant NM_022552.5:c.2249C>G p.(Pro750Arg), with a VAF of 12%, was absent from the patient's marrow before transplantation but present in the donor's blood at the time of donation, which indicated transfer of clonal haematopoiesis with the graft [89].

The patient underwent reinduction with daunorubicin and cytarabine (7+3 regimen), achieved complete remission and received a second HSCT from a matched unrelated donor, with fludarabine, treosulfan and antithymocyte globulin, and 6.7 × 10^6 CD34+ cells/kg; myeloid engraftment occurred on day +23, platelets exceeded 20 × 10^9/L on day +29, and there was grade I cutaneous acute GVHD. At 73 years, 42 months after the second HSCT, he remained in complete remission without relevant chronic GVHD [89]. Exome sequencing, with hair root as the germline sample and blood at leukaemia as the tumour sample, identified the heterozygous germline variant DDX41 NM_016222.4:c.992_994del p.(Lys331del), confirmed by Sanger sequencing in the hair root and in the blood of the brother, asymptomatic at the time, and a somatic second hit DDX41 c.1574G>A p.(Arg525His) with a VAF of 7.2%. There were no other siblings and no known tumour predisposition in the family [89].

Figure 1 of the article, which was consulted, shows the timeline of the disease (panel A), the pedigree (panel B) and the germline and somatic DDX41 variants and the DNMT3A variant of this case, compared with the cases published by Berger, Kobayashi and Gibson (panel C): https://pmc.ncbi.nlm.nih.gov/articles/PMC10170132/figure/Fig1/ [89, 157, 158].

10.2. Relation to the guidelines

The case brings together the elements that motivated the current recommendations: MDS with onset at 59 years in a man, without a family history, who did not meet the age criterion of the 2019 Nordic guidelines; the choice of an HLA-identical brother without germline testing; and donor cell leukaemia a decade later, with a somatic second hit in the DDX41 allele [53, 89]. The 2024 Nordic guidelines exclude DDX41 carriers as donors and recommend including DDX41 in the genetic work-up of all myeloid neoplasms, and the 2023 British guideline requires assessment for an inherited cause in all HSCT candidates and testing of related donors before donation [55, 109]. The conditioning and the second HSCT described are individual decisions in the report, not a recommended regimen. The authors propose genetic screening of patients and related donors and a central registry of family donors in Germany; they estimate that about 2% of post-HSCT relapses of AML/MDS are donor cell leukaemias [89].

11. Prognosis and Long-Term Follow-Up

11.1. Prognosis by entity

In FPDMM, the lifetime incidence of haematological neoplasm was 42.5% in the NIH cohort, GeneReviews gives 20%–50%, and the probability of a second neoplasm was estimated in the order of 10% [4, 47]. RUNX1-associated AML has outcomes comparable to those of AML with somatic RUNX1, of adverse risk in ELN 2022 [47]. In ANKRD26, the risk of MDS/AML is about 8%, and in ETV6 a haematological neoplasm arises in 20%–30%, with no evidence that the variant alters the response to therapy [6, 7].

In CEBPA familial AML, 10-year overall survival was 67% in ten pedigrees, against 54% in young adults with sporadic CEBPAdm and 29% with sporadic CEBPAsm [8, 100]. Germline variants involving the p30 isoform were associated with median overall survival above 25 years, against 11 years for the others (P = 0.013) [8, 176], and TAD2 variants with worse overall survival (OR for 2-year survival 0.06; 95% CI 0.01–0.34) in a univariate analysis of pooled cases [101]. New leukaemic episodes after cure require lifelong surveillance, and molecular testing of CEBPA in leukaemic cells at recurrence distinguishes relapse from a new episode [8].

In DDX41, MDS/AML has a favourable prognosis compared with sporadic patients of the same age, independent of the risk systems [44, 49], and ELN 2024 classifies DDX41 as favourable with less intensive therapy [106]. Late relapse after HSCT is a distinctive pattern, and DDX41 was the most frequent mutation in leukaemic cells in a study of 25 very late relapses cited by Loschi [156]. In the UK Biobank population, all-cause mortality of DDX41 carriers was higher than that of non-carriers (HR 1.35; 95% CI 1.1–1.7) [42].

In GATA2 deficiency, mortality by 40 years reached 35% in the French–Belgian series, and survival at 40 years after symptom onset ranges from 60% to 80% across series [10, 50]. Monosomy 7 with progressive MDS/AML has a poor prognosis without aggressive chemotherapy and HSCT, and, even with treatment, almost half of the patients in whom monosomy 7 is the main abnormality die of MDS/AML or complications [73]. In the SAMD9/SAMD9L syndromes, prognosis depends on karyotype more than on the variant: 5-year overall survival of 95% with normal karyotype and 76% with monosomy 7 [58]. In MIRAGE, the median age at death was 3 years, with the oldest described alive at 20 years [14]. In SAMD9L, mortality is increased by haematopoietic and immunological disease, and HSCT does not halt neurological progression [15].

The markers of progression that prompt consideration of early HSCT are common to the entities: second hits in the unaffected allele (CEBPA, DDX41, RUNX1), monosomy 7, ASXL1, SETBP1, RAS pathway and multi-hit TP53 [44]. An isolated clone is not, by itself, an indication for action; the gene, VAF, number of variants and dynamics are weighed [53]. In GATA2 deficiency, STAG2 in hypoplastic marrow is not associated with progression [44].

11.2. Principles of surveillance and divergences

Surveillance recommendations are expert opinion, with little evidence, and apply to P/LP carriers and to those who meet clinical criteria without a variant, but generally not to VUS carriers [53]. The start is individualised according to the typical age of each entity, and surveillance includes regular physical examination and education about symptoms [53, 75]. An abnormal blood count should be repeated within a few weeks, with exclusion of other causes, and leads to bone marrow examination; in the thrombocytopenias with predisposition, even a slight fall in platelets warrants investigation [53]. In children, worrying signs or worsening cytopenias warrant a blood count every 2–4 weeks until resolution, or bone marrow examination if they persist, and with stable cytopenias the frequency may be reduced, but not below annually [75].

There are documented divergences. Routine bone marrow examination with a stable blood count is not recommended by the Nordic guidelines, unlike the American recommendations of annual marrow in high-risk diseases [53]; the NIH proposes annual bone marrow biopsy with next-generation sequencing (NGS) in FPDMM, because of discordance between blood and marrow in up to a quarter of non-malignant states [47], and the AACR recommends annual marrow in children with GATA2 and SAMD9/SAMD9L and every 1–3 years with RUNX1, ETV6 and ANKRD26 [75]. Surveillance for DDX41 starts at 50 years (or 10 years before the earliest case) in the Nordic guidelines and the French penetrance study, at 40 years in the American proposal of Trottier and colleagues, and the 2026 British guideline restricts surveillance to research programmes [44, 55, 86, 102]. In the United Kingdom there is no consensus on whether, with which test and how often to monitor carriers without a haematological phenotype, and in practice a minority of centres perform an annual blood count [54]. The Nordic guidelines consider an annual myeloid NGS panel on blood potentially useful, especially in GATA2 and RUNX1 [53]. The French PNDS (French National Diagnosis and Care Protocol) for bone marrow aplasias adds, for the SAMD9/SAMD9L syndromes, a specific schedule in the active phase (weekly to monthly blood count and marrow cytology every 3 months with monosomy 7) and, after normalisation, a quarterly blood count and marrow every 12–18 months, with lifelong follow-up even without transplantation (subsection 7.15) [61].

11.3. Surveillance table

After HSCT, surveillance for secondary tumours is closer in the subgroup with organ dysfunction, and carriers should be referred to the hereditary cancer clinic for solid tumour screening [53].

12. Summary and Concluding Remarks

Myeloid neoplasms with germline predisposition have been recognised by the WHO and the ICC since 2022 as a category in which the germline gene qualifies the diagnosis of the neoplasm. The entities fall into three profiles: predisposition without prior disease (CEBPA, DDX41, TP53), thrombocytopenias with normal-sized platelets (RUNX1, ANKRD26, ETV6) and multisystem diseases (GATA2, SAMD9, SAMD9L and the bone marrow failure syndromes) [1, 2]. Risk ranges from about 8% MDS/AML in ANKRD26 to more than 80% AML for N-terminal CEBPA variants, and age at onset from early childhood (SAMD9/SAMD9L, GATA2, part of RUNX1) to the seventh decade (DDX41) [6, 8, 46, 47, 49].

DDX41 is the most frequent germline predisposition in adults, with 2%–6% of MDS/AML, male predominance and an often silent family history, which justifies including it in the work-up of all myeloid neoplasms regardless of age [44, 48, 55]. Germline confirmation requires non-haematopoietic DNA, preferably cultured fibroblasts, and the technical limitations are gene-specific: the 5'-UTR of ANKRD26, the non-coding regions and synonymous variants of GATA2, the GC richness of CEBPA and loss of the mutated allele through UPD or monosomy 7 in RUNX1, SAMD9 and SAMD9L [6, 8, 15, 54, 56].

Allo-HSCT is the only curative treatment, and its preventive indication has advocates only in high-penetrance diseases, with documented divergences in CEBPA and GATA2 [44, 53]. Donor selection is the decision with the greatest practical consequence: related donors should be tested, and carriers excluded whenever an alternative exists, because donor-derived leukaemia affected 19.4% of related-donor transplants in a French DDX41 cohort [48, 109, 156]. Surveillance of carriers rests on consensus statements without grading of evidence, with divergences concerning routine bone marrow examination and the starting age in DDX41, and the ongoing natural history studies (RUNX1, GATA2, DDX41) should provide the missing prospective data [47, 53, 86, 165, 167, 168].

13. References

1. 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. PMC9252913.
2. 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. PMC9479031.
3. Hörst K, Kühn C, Haferlach T. Genomics in Haematological Malignancies. In: ESMO Handbook — Cancer Genomics. Lugano: ESMO Press. p. 107–117 (Table 1, p. 109).
4. Deuitch N, Broadbridge E, Cunningham L, Liu P. RUNX1 Familial Platelet Disorder with Associated Myeloid Malignancies. 2021 Mar 4 [revised 2025 Dec 4]. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993-2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK568319/
5. Song WJ, Sullivan MG, Legare RD, Hutchings S, Tan X, Kufrin D, et al. Haploinsufficiency of CBFA2 causes familial thrombocytopenia with propensity to develop acute myelogenous leukaemia. Nat Genet. 1999;23(2):166-175. doi:10.1038/13793. PMID: 10508512.
6. Perez Botero J, Dugan SN. ANKRD26-Related Thrombocytopenia. 2018 Jun 21 [updated 2026 Sep 15]. In: Adam MP et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993-2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK507664/
7. Pencheva B, Di Paola J, Porter CC. ETV6-Related Thrombocytopenia and Predisposition to Leukemia. 2020 Nov 19 [updated 2025 Jun 26]. In: Adam MP et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993-2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK564234/
8. Drazer M, Tawana K. CEBPA-Associated Familial Acute Myeloid Leukemia. 2010 Oct 21 [updated 2026 Sep 3]. In: Adam MP et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993-2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK47457/
9. Churpek JE, Smith-Simmer K. DDX41-Associated Familial Myelodysplastic Syndrome and Acute Myeloid Leukemia. 2021 Oct 28. In: Adam MP et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993-2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK574843/
10. GeneGuidance. Genetic curation database: gene records for RUNX1, GATA2, DDX41, ETV6, ANKRD26, CEBPA, SAMD9 and SAMD9L [local database]. Version of 2026 Sep 26 (identifiers, clinical curation, ClinGen, labels and cited references). Accessed 2026 Oct 17.
11. Hsu AP, Sampaio EP, Khan J, Calvo KR, Lemieux JE, Patel SY, et al. Mutations in GATA2 are associated with the autosomal dominant and sporadic monocytopenia and mycobacterial infection (MonoMAC) syndrome. Blood. 2011;118(10):2653-2655. doi:10.1182/blood-2011-05-356352. PMID: 21670465. PMC3172785.
12. Ostergaard P, Simpson MA, Connell FC, Steward CG, Brice G, Woollard WJ, et al. Mutations in GATA2 cause primary lymphedema associated with a predisposition to acute myeloid leukemia (Emberger syndrome). Nat Genet. 2011;43(10):929-931. doi:10.1038/ng.923. PMID: 21892158.
13. Dickinson RE, Griffin H, Bigley V, Reynard LN, Hussain R, Haniffa M, et al. Exome sequencing identifies GATA-2 mutation as the cause of dendritic cell, monocyte, B and NK lymphoid deficiency. Blood. 2011;118(10):2656-2658. doi:10.1182/blood-2011-06-360313. PMID: 21765025. PMC5137783.
14. Tanase-Nakao K, Olson TS, Narumi S. MIRAGE Syndrome. 2020 Nov 25. In: Adam MP et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993-2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK564655/
15. Raskind WH, Keel S, Allenspach EJ. SAMD9L-Related Syndrome. 2017 Jun 1 [updated 2026 Aug 20]. In: Adam MP et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993-2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK435692/
16. 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. PMID: 28967166. PMC5765404. Tool: https://variantvalidator.org (GRCh38, accessed 2026-10-17).
17. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 468–470.
18. Online Mendelian Inheritance in Man (OMIM). #601399 PLATELET DISORDER, FAMILIAL, WITH ASSOCIATED MYELOID MALIGNANCY; FPDMM. Baltimore (MD): Johns Hopkins University. Available from: https://www.omim.org/entry/601399. Accessed 2026 Oct 17.
19. Online Mendelian Inheritance in Man (OMIM). *151385 RUNT-RELATED TRANSCRIPTION FACTOR 1; RUNX1. Baltimore (MD): Johns Hopkins University. Available from: https://www.omim.org/entry/151385. Accessed 2026 Oct 17.
20. Online Mendelian Inheritance in Man (OMIM). #188000 THROMBOCYTOPENIA 2; THC2. Baltimore (MD): Johns Hopkins University. Available from: https://www.omim.org/entry/188000. Accessed 2026 Oct 17.
21. Online Mendelian Inheritance in Man (OMIM). #616216 THROMBOCYTOPENIA 5; THC5. Baltimore (MD): Johns Hopkins University. Available from: https://www.omim.org/entry/616216. Accessed 2026 Oct 17.
22. Online Mendelian Inheritance in Man (OMIM). #601626 LEUKEMIA, ACUTE MYELOID; AML. Baltimore (MD): Johns Hopkins University. Available from: https://www.omim.org/entry/601626. Accessed 2026 Oct 17.
23. Online Mendelian Inheritance in Man (OMIM). #616871 MYELOPROLIFERATIVE/LYMPHOPROLIFERATIVE NEOPLASMS, FAMILIAL (MULTIPLE TYPES), SUSCEPTIBILITY TO; MPLPF. Baltimore (MD): Johns Hopkins University. Available from: https://www.omim.org/entry/616871. Accessed 2026 Oct 17.
24. Online Mendelian Inheritance in Man (OMIM). *608170 DEAD-BOX HELICASE 41; DDX41. Baltimore (MD): Johns Hopkins University. Available from: https://www.omim.org/entry/608170. Accessed 2026 Oct 17.
25. Online Mendelian Inheritance in Man (OMIM). #614172 IMMUNODEFICIENCY 21; IMD21. Baltimore (MD): Johns Hopkins University. Available from: https://www.omim.org/entry/614172. Accessed 2026 Oct 17.
26. Online Mendelian Inheritance in Man (OMIM). #614286 MYELODYSPLASTIC SYNDROME; MDS. Baltimore (MD): Johns Hopkins University. Available from: https://www.omim.org/entry/614286. Accessed 2026 Oct 17.
27. Online Mendelian Inheritance in Man (OMIM). #617053 MIRAGE SYNDROME; MIRAGE. Baltimore (MD): Johns Hopkins University. Available from: https://www.omim.org/entry/617053. Accessed 2026 Oct 17.
28. Online Mendelian Inheritance in Man (OMIM). #619041 MONOSOMY 7 MYELODYSPLASIA AND LEUKEMIA SYNDROME 2; M7MLS2. Baltimore (MD): Johns Hopkins University. Available from: https://www.omim.org/entry/619041. Accessed 2026 Oct 17.
29. Online Mendelian Inheritance in Man (OMIM). #159550 ATAXIA-PANCYTOPENIA SYNDROME; ATXPC. Baltimore (MD): Johns Hopkins University. Available from: https://www.omim.org/entry/159550. Accessed 2026 Oct 17.
30. Online Mendelian Inheritance in Man (OMIM). #252270 MONOSOMY 7 MYELODYSPLASIA AND LEUKEMIA SYNDROME 1; M7MLS1. Baltimore (MD): Johns Hopkins University. Available from: https://www.omim.org/entry/252270. Accessed 2026 Oct 17.
31. Orphanet. Familial platelet disorder with associated myeloid malignancy. ORPHA:71290. https://www.orpha.net/en/disease/detail/71290
32. Orphanet. Autosomal isolated thrombocytopenia with normal platelets size. ORPHA:168629. https://www.orpha.net/en/disease/detail/168629
33. Orphanet. Inherited acute myeloid leukemia. ORPHA:319465. https://www.orpha.net/en/disease/detail/319465
34. Orphanet. DDX41-related hematologic malignancy predisposition syndrome. ORPHA:488647. https://www.orpha.net/en/disease/detail/488647
35. Orphanet. GATA2 deficiency spectrum. ORPHA:228423. https://www.orpha.net/en/disease/detail/228423
36. Orphanet. Deafness-lymphedema-leukemia syndrome. ORPHA:3226. https://www.orpha.net/en/disease/detail/3226
37. Orphanet. MIRAGE syndrome. ORPHA:494433. https://www.orpha.net/en/disease/detail/494433
38. Orphanet. Ataxia-pancytopenia syndrome. ORPHA:2585. https://www.orpha.net/en/disease/detail/2585
39. Orphanet. SAMD9L-associated autoinflammatory syndrome. ORPHA:619367. https://www.orpha.net/en/disease/detail/619367
40. Orphanet. Spinocerebellar ataxia type 49. ORPHA:631106. https://www.orpha.net/en/disease/detail/631106
41. Balduini CL, Pecci A, Noris P. Inherited thrombocytopenias: the evolving spectrum. Hamostaseologie. 2012;32(4):259-270. doi:10.5482/ha12050001. PMID: 22972471.
42. Hendricks RM, Kim J, Haley JS, Ramos ML, Mirshahi UL, Carey DJ, et al. Genome-first determination of the prevalence and penetrance of eight germline myeloid malignancy predisposition genes: a study of two population-based cohorts. Leukemia. 2025;39(2):400-411. doi:10.1038/s41375-024-02436-y. PMID: 39501104. PMC11794151.
43. Pabst T, Eyholzer M, Haefliger S, Schardt J, Mueller BU. Somatic CEBPA mutations are a frequent second event in families with germline CEBPA mutations and familial acute myeloid leukemia. J Clin Oncol. 2008;26(31):5088-5093. doi:10.1200/jco.2008.16.5563. PMID: 18768433.
44. 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. PMC13170486.
45. Wlodarski MW, Hirabayashi S, Pastor V, Starý J, Hasle H, Masetti R, et al. Prevalence, clinical characteristics, and prognosis of GATA2-related myelodysplastic syndromes in children and adolescents. Blood. 2016;127(11):1387-97; quiz 1518. doi:10.1182/blood-2015-09-669937. PMID: 26702063.
46. 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. PMC9330547.
47. Kajdic A, Deuitch NT, Bresciani E, Chong S, Craft K, Davis J, et al. Hematologic malignancies in pediatric patients with RUNX1-familial platelet disorder with associated myeloid malignancy. Blood Adv. 2026;10(9):3075-3085. doi:10.1182/bloodadvances.2025017808. PMID: 41616279. PMC13141479.
48. 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.
49. Makishima H, Saiki R, Nannya Y, Korotev S, Gurnari C, Takeda J, et al. Germ line DDX41 mutations define a unique subtype of myeloid neoplasms. Blood. 2023;141(5):534-549. doi:10.1182/blood.2022018221. PMID: 36322930. PMC10935555.
50. Donadieu J, Lamant M, Fieschi C, de Fontbrune FS, Caye A, Ouachee M, et al. Natural history of GATA2 deficiency in a survey of 79 French and Belgian patients. Haematologica. 2018;103(8):1278-1287. doi:10.3324/haematol.2017.181909. PMID: 29724903. PMC6068047.
51. Noris P, Favier R, Alessi MC, Geddis AE, Kunishima S, Heller PG, et al. ANKRD26-related thrombocytopenia and myeloid malignancies. Blood. 2013;122(11):1987-1989. doi:10.1182/blood-2013-04-499319. PMID: 24030261.
52. Spinner MA, Sanchez LA, Hsu AP, Shaw PA, Zerbe CS, Calvo KR, et al. GATA2 deficiency: a protean disorder of hematopoiesis, lymphatics, and immunity. Blood. 2014;123(6):809-821. doi:10.1182/blood-2013-07-515528. PMID: 24227816. PMC3916876.
53. Baliakas P, Tesi B, Wartiovaara-Kautto U, Stray-Pedersen A, Friis LS, Dybedal I, et al. Nordic Guidelines for Germline Predisposition to Myeloid Neoplasms in Adults: Recommendations for Genetic Diagnosis, Clinical Management and Follow-up. Hemasphere. 2019;3(6):e321. doi:10.1097/hs9.0000000000000321. PMID: 31976490. PMC6924562.
54. Speight B, Hanson H, Turnbull C, Hardy S, Drummond J, Khorashad J, et al. Germline predisposition to haematological malignancies: Best practice consensus guidelines from the UK Cancer Genetics Group (UKCGG), CanGene-CanVar and the NHS England Haematological Oncology Working Group. Br J Haematol. 2023;201(1):25-34. doi:10.1111/bjh.18675. PMID: 36744544.
55. Baliakas P, Tesi B, Cammenga J, Stray-Pedersen A, Jahnukainen K, Andersen MK, et al. How to manage patients with germline DDX41 variants: Recommendations from the Nordic working group on germline predisposition for myeloid neoplasms. Hemasphere. 2024;8(8):e145. doi:10.1002/hem3.145. PMID: 39139355. PMC11320078.
56. Kozyra EJ, Pastor VB, Lefkopoulos S, Sahoo SS, Busch H, Voss RK, et al. Synonymous GATA2 mutations result in selective loss of mutated RNA and are common in patients with GATA2 deficiency. Leukemia. 2020;34(10):2673-2687. doi:10.1038/s41375-020-0899-5. PMID: 32555368. PMC7515837.
57. Zaninetti C, Gresele P, Bertomoro A, Klersy C, De Candia E, Veneri D, et al. Eltrombopag for the treatment of inherited thrombocytopenias: a phase II clinical trial. Haematologica. 2020;105(3):820-828. doi:10.3324/haematol.2019.223966. PMID: 31273088. PMC7049343.
58. 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. PMC11826520.
59. European Reference Network on Rare Hematological Diseases (ERN-EuroBloodNet). https://eurobloodnet.eu/
60. European Reference Network on Rare Immunodeficiency, Autoinflammatory and Autoimmune Diseases (ERN RITA). https://ern-rita.org/
61. 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 on 2023-05-04; reference centre for acquired and constitutional bone marrow aplasias, MaRIH filière). Available from: https://www.has-sante.fr/jcms/p_3431517
62. Smith ML, Cavenagh JD, Lister TA, Fitzgibbon J. Mutation of CEBPA in familial acute myeloid leukemia. N Engl J Med. 2004;351(23):2403-2407. doi:10.1056/nejmoa041331. PMID: 15575056.
63. Hahn CN, Chong CE, Carmichael CL, Wilkins EJ, Brautigan PJ, Li XC, et al. Heritable GATA2 mutations associated with familial myelodysplastic syndrome and acute myeloid leukemia. Nat Genet. 2011;43(10):1012-1017. doi:10.1038/ng.913. PMID: 21892162. PMC3184204.
64. Pippucci T, Savoia A, Perrotta S, Pujol-Moix N, Noris P, Castegnaro G, et al. Mutations in the 5' UTR of ANKRD26, the ankirin repeat domain 26 gene, cause an autosomal-dominant form of inherited thrombocytopenia, THC2. Am J Hum Genet. 2011;88(1):115-120. doi:10.1016/j.ajhg.2010.12.006. PMID: 21211618. PMC3014357.
65. Noris P, Perrotta S, Seri M, Pecci A, Gnan C, Loffredo G, et al. Mutations in ANKRD26 are responsible for a frequent form of inherited thrombocytopenia: analysis of 78 patients from 21 families. Blood. 2011;117(24):6673-6680. doi:10.1182/blood-2011-02-336537. PMID: 21467542.
66. Polprasert C, Schulze I, Sekeres MA, Makishima H, Przychodzen B, Hosono N, et al. Inherited and Somatic Defects in DDX41 in Myeloid Neoplasms. Cancer Cell. 2015;27(5):658-670. doi:10.1016/j.ccell.2015.03.017. PMID: 25920683. PMC8713504.
67. Zhang MY, Churpek JE, Keel SB, Walsh T, Lee MK, Loeb KR, et al. Germline ETV6 mutations in familial thrombocytopenia and hematologic malignancy. Nat Genet. 2015;47(2):180-185. doi:10.1038/ng.3177. PMID: 25581430. PMC4540357.
68. Noetzli L, Lo RW, Lee-Sherick AB, Callaghan M, Noris P, Savoia A, et al. Germline mutations in ETV6 are associated with thrombocytopenia, red cell macrocytosis and predisposition to lymphoblastic leukemia. Nat Genet. 2015;47(5):535-538. doi:10.1038/ng.3253. PMID: 25807284. PMC4631613.
69. Topka S, Vijai J, Walsh MF, Jacobs L, Maria A, Villano D, et al. Germline ETV6 Mutations Confer Susceptibility to Acute Lymphoblastic Leukemia and Thrombocytopenia. PLoS Genet. 2015;11(6):e1005262. doi:10.1371/journal.pgen.1005262. PMID: 26102509. PMC4477877.
70. 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.
71. 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. PMC4908176.
72. 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. PMC5399482.
73. Olson TS, Dickerson KE, Nakano TA, Wlodarski M. Monosomy 7 Predisposition Syndromes Overview. 2021 Jun 10. In: Adam MP et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993-2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK571103/
74. Rudelius M, Weinberg OK, Niemeyer CM, Shimamura A, Calvo KR. The International Consensus Classification (ICC) of hematologic neoplasms with germline predisposition, pediatric myelodysplastic syndrome, and juvenile myelomonocytic leukemia. Virchows Arch. 2023;482(1):113-130. doi:10.1007/s00428-022-03447-9. PMID: 36445482.
75. 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. PMC11444884.
76. Antony-Debré I, Duployez N, Bucci M, Geffroy S, Micol JB, Renneville A, et al. Somatic mutations associated with leukemic progression of familial platelet disorder with predisposition to acute myeloid leukemia. Leukemia. 2016;30(4):999-1002. doi:10.1038/leu.2015.236. PMID: 26316320.
77. Glembotsky AC, Marin Oyarzún CP, De Luca G, Marzac C, Auger N, Goette NP, et al. First description of revertant mosaicism in familial platelet disorder with predisposition to acute myelogenous leukemia: correlation with the clinical phenotype. Haematologica. 2020;105(10):e535. doi:10.3324/haematol.2020.253070. PMID: 33054100. PMC7556663.
78. Lawal RA, Gutierrez-Rodrigues F, Young DJ, Bresciani E, Groarke EM, Deuitch NT, et al. Telomere attrition is common in patients with germline RUNX1 pathogenic variants. Haematologica. 2026;111(6):2118-2123. doi:10.3324/haematol.2025.288522. PMID: 41504231. PMC13231119.
79. Bluteau D, Balduini A, Balayn N, Currao M, Nurden P, Deswarte C, et al. Thrombocytopenia-associated mutations in the ANKRD26 regulatory region induce MAPK hyperactivation. J Clin Invest. 2014;124(2):580-591. doi:10.1172/jci71861. PMID: 24430186. PMC3904625.
80. Wahlster L, Verboon JM, Ludwig LS, Black SC, Luo W, Garg K, et al. Familial thrombocytopenia due to a complex structural variant resulting in a WAC-ANKRD26 fusion transcript. J Exp Med. 2021;218(6):e20210444. doi:10.1084/jem.20210444. PMID: 33857290. PMC8056752.
81. Marconi C, Canobbio I, Bozzi V, Pippucci T, Simonetti G, Melazzini F, et al. 5'UTR point substitutions and N-terminal truncating mutations of ANKRD26 in acute myeloid leukemia. J Hematol Oncol. 2017;10(1):18. doi:10.1186/s13045-016-0382-y. PMID: 28100250. PMC5242010.
82. Duployez N, Largeaud L, Duchmann M, Kim R, Rieunier J, Lambert J, et al. Prognostic impact of DDX41 germline mutations in intensively treated acute myeloid leukemia patients: an ALFA-FILO study. Blood. 2022;140(7):756-768. doi:10.1182/blood.2021015328. PMID: 35443031. PMC9389637.
83. Santiago M, Liquori A, Such E, Zúñiga Á, Cervera J. The Clinical Spectrum, Diagnosis, and Management of GATA2 Deficiency. Cancers (Basel). 2023;15(5):1590. doi:10.3390/cancers15051590. PMID: 36900380. PMC10000430.
84. 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. PMC7970252.
85. 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. PMC6749008.
86. Speight B, Hamblin A, Talley P, Tsoulaki O, Robinson R, Dean P, et al. Somatic and germline genetic testing pathways in haematological malignancies: Best practice consensus guidelines from the 2025 national meeting organised by UK Cancer Genetics Group (UKCGG), CanGene-CanVar and the NHS England Haematological Oncology Working Group. Br J Haematol. 2026;208(4):1211-1222. doi:10.1111/bjh.70404. PMID: 41773566. PMC13071493.
87. Saygin C, Roloff G, Hahn CN, Chhetri R, Gill S, Elmariah H, et al. Allogeneic hematopoietic stem cell transplant outcomes in adults with inherited myeloid malignancies. Blood Adv. 2023;7(4):549-554. doi:10.1182/bloodadvances.2022008172. PMID: 36001442. PMC9979761.
88. Hattig Z, Carlelycke TE, Ruiz Martin A, Walker T, Speck NA, Liu P, et al. Update of germ line RUNX1 variant curation rules: version 3.1. Blood Adv. 2026;10(12):4395-4398. doi:10.1182/bloodadvances.2026020060. PMID: 42018664. PMC13285826.
89. Rolles B, Meyer R, Begemann M, Elbracht M, Jost E, Stelljes M, et al. DDX41 germline variants causing donor cell leukemia indicate a need for further genetic workup in the context of hematopoietic stem cell transplantation. Blood Cancer J. 2023;13(1):73. doi:10.1038/s41408-023-00846-2. PMID: 37160870. PMC10170132.
90. Wolff JH, Skov TW, Dorset SR, Sjelborg J, Ravendran S, Revenfeld ALS, et al. Prime editing of human hematopoietic stem cells for correction of GATA2 deficiency. Sci Transl Med. 2026;18(869):eaec7031. doi:10.1126/scitranslmed.aec7031. PMID: 42814803.
91. 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.
92. Luo X, Feurstein S, Mohan S, Porter CC, Jackson SA, Keel S, et al. ClinGen Myeloid Malignancy Variant Curation Expert Panel recommendations for germline RUNX1 variants. Blood Adv. 2019;3(20):2962-2979. doi:10.1182/bloodadvances.2019000644. PMID: 31648317. PMC6849945.
93. Menezes AC, Deuitch NT, Pintuff A, Zhen T, Wincovitch S, Ozturk K, et al. Functional analysis of germline RUNX1 variants identified in individuals with suspected familial platelet disorder. Blood Adv. 2026;10(19):6407-6421. doi:10.1182/bloodadvances.2026019984. PMID: 42308221. PMC13633912.
94. Mohan S, Carlelycke TE, Koppayi AL, Walker T, Tallis E, Trottier AM, et al. A quantitative definition of the clinical manifestations of GATA2 deficiency in adults. Blood. 2026:blood.2026034942. doi:10.1182/blood.2026034942. PMID: 42561126.
95. George A, Rolf E, Domeradzka M, Ribeiro S, Mills B, Hallett A, et al. Leveraging paired germline and somatic analysis to improve the classification of DDX41 variants. Br J Haematol. 2026;208(5):1584-1595. doi:10.1111/bjh.70411. PMID: 41873058. PMC13176513.
96. Brown AL, Arts P, Carmichael CL, Babic M, Dobbins J, Chong CE, et al. RUNX1-mutated families show phenotype heterogeneity and a somatic mutation profile unique to germline predisposed AML. Blood Adv. 2020;4(6):1131-1144. doi:10.1182/bloodadvances.2019000901. PMID: 32208489. PMC7094007.
97. Kearns O, Deuitch NT, Craft K, Young DJ, Liu PP, Wiener L. Psychosocial aspects of RUNX1-familial platelet disorder in adolescents and young adults: "The fear of knowing it could happen sometime down the road". J Genet Couns. 2026;35(4):e70253. doi:10.1002/jgc4.70253. PMID: 42383489. PMC13321132.
98. Di Paola J, Porter CC. ETV6-related thrombocytopenia and leukemia predisposition. Blood. 2019;134(8):663-667. doi:10.1182/blood.2019852418. PMID: 31248877. PMC6706811.
99. Moriyama T, Metzger ML, Wu G, Nishii R, Qian M, Devidas M, et al. Germline genetic variation in ETV6 and risk of childhood acute lymphoblastic leukaemia: a systematic genetic study. Lancet Oncol. 2015;16(16):1659-1666. doi:10.1016/s1470-2045(15)00369-1. PMID: 26522332. PMC4684709.
100. Tawana K, Wang J, Renneville A, Bödör C, Hills R, Loveday C, et al. Disease evolution and outcomes in familial AML with germline CEBPA mutations. Blood. 2015;126(10):1214-1223. doi:10.1182/blood-2015-05-647172. PMID: 26162409.
101. Yang J, Duan X, Zhu M, Wu Y. Location matters: topography of germline CEBPA variants predicts variable outcomes in familial acute myeloid leukaemia-a rare disease perspective. J Med Genet. 2026;63(9):596-604. doi:10.1136/jmg-2026-111543. PMID: 42209215.
102. Villy MC, Drouet Y, Larcher L, Vial Y, Dauriat B, Veyrune L, et al. Myeloid neoplasms risks for germline DDX41 pathogenic variants carriers. Eur J Hum Genet. 2025. doi:10.1038/s41431-025-01952-4. PMID: 41044237.
103. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 260–264.
104. 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. PMC5830370.
105. 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. PMC7108905.
106. Döhner H, DiNardo CD, Appelbaum FR, Craddock C, Dombret H, Ebert BL, et al. Genetic risk classification for adults with AML receiving less-intensive therapies: the 2024 ELN recommendations. Blood. 2024;144(21):2169-2173. doi:10.1182/blood.2024025409. PMID: 39133932.
107. Taskesen E, Bullinger L, Corbacioglu A, Sanders MA, Erpelinck CA, Wouters BJ, et al. Prognostic impact, concurrent genetic mutations, and gene expression features of AML with CEBPA mutations in a cohort of 1182 cytogenetically normal AML patients: further evidence for CEBPA double mutant AML as a distinctive disease entity. Blood. 2011;117(8):2469-2475. doi:10.1182/blood-2010-09-307280. PMID: 21177436.
108. Hörst K, Kühn C, Haferlach T. Genomics in Haematological Malignancies. In: ESMO Handbook — Cancer Genomics. Lugano: ESMO Press. p. 112–116.
109. 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.
110. Carapito R, Konantz M, Paillard C, Miao Z, Pichot A, Leduc MS, et al. Mutations in signal recognition particle SRP54 cause syndromic neutropenia with Shwachman-Diamond-like features. J Clin Invest. 2017;127(11):4090-103. doi:10.1172/JCI92876
111. Bellanné-Chantelot C, Schmaltz-Panneau B, Marty C, Fenneteau O, Callebaut I, Clauin S, et al. Mutations in the SRP54 gene cause severe congenital neutropenia as well as Shwachman-Diamond-like syndrome. Blood. 2018;132(12):1318-31. doi:10.1182/blood-2017-12-820308
112. Porter CC, Druley TE, Erez A, Kuiper RP, Onel K, Schiffman JD, et al. Recommendations for Surveillance for Children with Leukemia-Predisposing Conditions. Clin Cancer Res. 2017;23(11):e14-e22. doi:10.1158/1078-0432.ccr-17-0428. PMID: 28572263.
113. Dupuis A, Gachet C. Inherited platelet disorders : Management of the bleeding risk. Transfus Clin Biol. 2018;25(3):228-235. doi:10.1016/j.tracli.2018.07.003. PMID: 30077511.
114. Zaninetti C. Thrombopoietin receptor agonists in inherited thrombocytopenias: an evolving mosaic of evidence. Res Pract Thromb Haemost. 2026;10(6):106907. doi:10.1016/j.rpth.2026.106907. PMID: 42703607. PMC13546829.
115. ClinicalTrials.gov. NCT06090669 — Phase Ib Study of Imatinib to Increase RUNX1 Activity in Participants With Germline RUNX1 Deficiency (NCI). https://clinicaltrials.gov/study/NCT06090669
116. ClinicalTrials.gov. NCT06261060 — Low-Dose Sirolimus to Increase Hematopoietic Function in Patients With RUNX1 Familial Platelet Disorder (MD Anderson). https://clinicaltrials.gov/study/NCT06261060
117. ClinicalTrials.gov. NCT07844694 — Low-Dose Sirolimus… RUNX1 Familial Platelet Disorder – Part II (OHSU). https://clinicaltrials.gov/study/NCT07844694
118. ClinicalTrials.gov. NCT06414889 — Safety and Feasibility of Autologous CD34+ Hematopoietic Stem Cells Mobilization and Apheresis in Participants With RUNX1 Familial Platelet Disorder (MD Anderson). https://clinicaltrials.gov/study/NCT06414889
119. Pfizer. CYKLOKAPRON (tranexamic acid injection) prescribing information. DailyMed setid 6e89a7d9-4da4-42aa-b7f8-c602c24eefe5; current label 2025-09-03. https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=6e89a7d9-4da4-42aa-b7f8-c602c24eefe5
120. Amneal Pharmaceuticals. Aminocaproic acid tablets USP (generic of AMICAR®) prescribing information. DailyMed setid 4a22e559-2dac-42d7-af8d-eb0ebbc2f76d; current label 2026-07-28. https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=4a22e559-2dac-42d7-af8d-eb0ebbc2f76d (accessed 26/10/2026).
121. Ferring. DDAVP Injection (desmopressin acetate) prescribing information. DailyMed setid 651f6fee-a2c7-431b-8d5d-58b156c72244; current label 2022-09-28. https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=651f6fee-a2c7-431b-8d5d-58b156c72244
122. Agence nationale de sécurité du médicament et des produits de santé. MINIRIN 4 microgrammes/mL, solution injectable (desmopressine): résumé des caractéristiques du produit, rubriques 4.1–4.3 [updated 2026 May 21; accessed 2026 Nov 7]. Base de données publique des médicaments. Available from: https://base-donnees-publique.medicaments.gouv.fr/medicament/67576220/extrait
123. Agence nationale de sécurité du médicament et des produits de santé (ANSM). OCTIM 150 microgrammes/dose, solution pour pulvérisation nasale (desmopressine) — Résumé des caractéristiques du produit, sections 4.1–4.3 [updated 21 Oct 2025; accessed 7 Nov 2026]. Base de données publique des médicaments. Available from: https://base-donnees-publique.medicaments.gouv.fr/medicament/61396361/extrait
124. Novartis. PROMACTA (eltrombopag) prescribing information. NDA 022291 (initial approval 2008-11-20). DailyMed setid 7714a0ed-34bb-46e6-a0a5-b363908b22c2; current label 2025-12-18. https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=7714a0ed-34bb-46e6-a0a5-b363908b22c2
125. AbbVie. VENCLEXTA (venetoclax) prescribing information. NDA 208573 (initial approval 2016-04-11). DailyMed setid b118a40d-6b56-cee3-10f6-ded821a97018; current label 2026-05-20. https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=b118a40d-6b56-cee3-10f6-ded821a97018
126. European Medicines Agency. Venclyxto (venetoclax) — EPAR. Authorisation 04/12/2016; revision 23 (10/07/2026). https://www.ema.europa.eu/en/medicines/human/EPAR/venclyxto
127. Bristol-Myers Squibb. VIDAZA (azacitidine) prescribing information. NDA 050794 (initial approval 2004-05-19). DailyMed setid 3495a71a-cc04-4776-851f-f185956f32af; current label 2024-01-29. https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=3495a71a-cc04-4776-851f-f185956f32af
128. European Medicines Agency. Vidaza (azacitidine) — EPAR. Authorisation 17/12/2008; revision 28 (05/11/2024). https://www.ema.europa.eu/en/medicines/human/EPAR/vidaza
129. ClinicalTrials.gov. NCT05907746 — A Phase II Study of Allogeneic Hematopoietic Stem Cell Transplantation With Briquilimab-Based Conditioning in Participants With GATA2 Deficiency (NCI). https://clinicaltrials.gov/study/NCT05907746
130. Cyclophosphamide for injection USP prescribing information. DailyMed setid 5bad5bfc-12c0-46a2-8355-1028f108ecbf; current label 2026-10-15. https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=5bad5bfc-12c0-46a2-8355-1028f108ecbf (accessed 26/10/2026).
131. 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. PMC9405857.
132. Ascend Laboratories. Riluzole tablets (generic of RILUTEK®) prescribing information. DailyMed setid dd432c13-389e-46b3-8891-0cf1a58800ff; current label 2026-07-15. https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=dd432c13-389e-46b3-8891-0cf1a58800ff (accessed 26/10/2026).
133. European Medicines Agency. Rilutek (riluzole) — EPAR, product information (SmPC, section 4.2). https://www.ema.europa.eu/en/documents/product-information/rilutek-epar-product-information_en.pdf (accessed 26/10/2026).
134. European Medicines Agency. Alkindi (hydrocortisone) — EPAR, product information (SmPC, section 4.2). https://www.ema.europa.eu/en/documents/product-information/alkindi-epar-product-information_en.pdf (accessed 26/10/2026).
135. Eton Pharmaceuticals. ALKINDI SPRINKLE (hydrocortisone) oral granules prescribing information. DailyMed setid 26c13a5f-7119-4c6a-bf10-fda4e07d7682; current label 2026-04-21. https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=26c13a5f-7119-4c6a-bf10-fda4e07d7682 (accessed 26/10/2026).
136. European Medicines Agency. Plenadren (hydrocortisone, modified release) — EPAR, product information (SmPC, section 4.2). https://www.ema.europa.eu/en/documents/product-information/plenadren-epar-product-information_en.pdf (accessed 26/10/2026).
137. Major Pharmaceuticals. Fludrocortisone acetate tablets USP 0.1 mg prescribing information. DailyMed setid 933e28ef-05e8-419e-a786-21eab67f33a2; current label 2026-06-24. https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=933e28ef-05e8-419e-a786-21eab67f33a2 (accessed 26/10/2026).
138. Laboratórios Pfizer. Zithromax 500 mg comprimidos revestidos por película (azithromycin) — summary of product characteristics. INFARMED, Infomed, med_id 9466. https://extranet.infarmed.pt/INFOMED-fo/download-ficheiro.xhtml?med_id=9466&tipo_doc=rcm; and Pfizer. ZITHROMAX (azithromycin) prescribing information, FDA label of 2026-07-24 (set ID 3b631aa1-2d46-40bc-a614-d698301ea4f9), https://api.fda.gov/drug/label.json?search=openfda.brand_name:ZITHROMAX (accessed 26/10/2026).
139. CSL Behring. PRIVIGEN (immune globulin intravenous [human], 10% liquid) prescribing information. DailyMed setid 7e5649da-75be-4a42-8eeb-4aeba562c401; current label 2026-05-04. https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=7e5649da-75be-4a42-8eeb-4aeba562c401 (accessed 26/10/2026).
140. European Medicines Agency. Privigen (human normal immunoglobulin, IV) — EPAR, product information (SmPC, section 4.2). https://www.ema.europa.eu/en/documents/product-information/privigen-epar-product-information_en.pdf (accessed 26/10/2026).
141. Celgene (Bristol-Myers Squibb). REVLIMID (lenalidomide) capsules prescribing information. DailyMed setid 5fa97bf5-28a2-48f1-8955-f56012d296be; current label 2023-03-24. https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=5fa97bf5-28a2-48f1-8955-f56012d296be (accessed 26/10/2026).
142. European Medicines Agency. Revlimid (lenalidomide) — EPAR, product information (SmPC, sections 4.1 and 4.2). https://www.ema.europa.eu/en/documents/product-information/revlimid-epar-product-information_en.pdf (accessed 26/10/2026).
143. European Medicines Agency. Busulfan Fresenius Kabi: EPAR – Product Information (Summary of Product Characteristics). Authorised 2014 Sep 22; last EC decision 2025 Jul 1. Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/busulfan-fresenius-kabi
144. European Medicines Agency. Trecondi (treosulfan): EPAR — product information [Internet]. Amsterdam: EMA; authorisation 2019-06-20; last renewal 2024-01-05; text version consulted dated December 2025. Available from: https://www.ema.europa.eu/en/documents/product-information/trecondi-epar-product-information_en.pdf
145. European Medicines Agency. Tepadina (thiotepa): EPAR — product information [Internet]. Amsterdam: EMA; authorisation 2010-03-15; last renewal 2014-11-17 [accessed 6 Nov 2026]. Available from: https://www.ema.europa.eu/en/documents/product-information/tepadina-epar-product-information_en.pdf
146. Kałwak K, Mielcarek M, Patrick K, Styczyński J, Bader P, Corbacioglu S, et al. Treosulfan-fludarabine-thiotepa-based conditioning treatment before allogeneic hematopoietic stem cell transplantation for pediatric patients with hematological malignancies. Bone Marrow Transplant. 2020;55(10):1996-2007. doi:10.1038/s41409-020-0869-6.
147. European Medicines Agency. Revolade (eltrombopag) — EPAR. Authorisation 11/03/2010; revision 39 (26/02/2026). https://www.ema.europa.eu/en/medicines/human/EPAR/revolade
148. Döhner H, Estey E, Grimwade D, Amadori S, Appelbaum FR, Büchner T, et al. Diagnosis and management of AML in adults: 2017 ELN recommendations from an international expert panel. Blood. 2017;129(4):424-447. doi:10.1182/blood-2016-08-733196. PMID: 27895058. PMC5291965.
149. Olson TS, Ericson K, Antin JH, Babbitt L, Babich M, Deuitch NT, et al. Preemptive hematopoietic stem cell transplantation in RUNX1 familial platelet disorder: a shared decision-making framework. Haematologica. 2026;111(9):2911-2923. doi:10.3324/haematol.2025.300403. PMID: 41924923. PMC13530962.
150. Rajput RV, Arnold DE. GATA2 Deficiency: Predisposition to Myeloid Malignancy and Hematopoietic Cell Transplantation. Curr Hematol Malig Rep. 2023;18(4):89-97. doi:10.1007/s11899-023-00695-7. PMID: 37247092.
151. 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. PMC5787871.
152. 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. PMC7110513.
153. Owen CJ, Toze CL, Koochin A, Forrest DL, Smith CA, Stevens JM, et al. Five new pedigrees with inherited RUNX1 mutations causing familial platelet disorder with propensity to myeloid malignancy. Blood. 2008;112(12):4639-4645. doi:10.1182/blood-2008-05-156745. PMID: 18723428.
154. Godley LA. Inherited predisposition to acute myeloid leukemia. Semin Hematol. 2014;51(4):306-321. doi:10.1053/j.seminhematol.2014.08.001. PMID: 25311743.
155. 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.
156. Loschi M, Villy MC, Galimard JE, Auboiroux MM, Gachard N, Perani A, et al. Allogeneic hematopoietic stem cell transplantation in patients with germ line DDX41 mutated myeloid malignancies. Blood Adv. 2025;9(24):6314-6325. doi:10.1182/bloodadvances.2025017084. PMID: 40924929. PMC12744251.
157. Berger G, van den Berg E, Sikkema-Raddatz B, Abbott KM, Sinke RJ, Bungener LB, et al. Re-emergence of acute myeloid leukemia in donor cells following allogeneic transplantation in a family with a germline DDX41 mutation. Leukemia. 2017;31(2):520-522. doi:10.1038/leu.2016.310. PMID: 27795557.
158. Kobayashi S, Kobayashi A, Osawa Y, Nagao S, Takano K, Okada Y, et al. Donor cell leukemia arising from preleukemic clones with a novel germline DDX41 mutation after allogenic hematopoietic stem cell transplantation. Leukemia. 2017;31(4):1020-1022. doi:10.1038/leu.2017.44. PMID: 28194039.
159. Peffault de Latour R, Soulier J. How I treat MDS and AML in Fanconi anemia. Blood. 2016;127(24):2971-2979. doi:10.1182/blood-2016-01-583625. PMID: 27020090.
160. Hasle H, Alonzo TA, Auvrignon A, Behar C, Chang M, Creutzig U, et al. Monosomy 7 and deletion 7q in children and adolescents with acute myeloid leukemia: an international retrospective study. Blood. 2007;109(11):4641-4647. doi:10.1182/blood-2006-10-051342. PMID: 17299091.
161. ClinicalTrials.gov. NCT07112287 — MyeloGen: Germline Testing for Predisposition to Myeloid Malignancies. https://clinicaltrials.gov/study/NCT07112287
162. Mohammadhosseini M, Sahasrabudhe A, Ostrowski LH, McClatchy J, Knight R, Schabel K, et al. TNF-α signaling drives myeloid skewing and clonal expansion of stem and progenitor cells in RUNX1-familial platelet disorder. Haematologica. 2026;111(4):1436-1442. doi:10.3324/haematol.2025.288410. PMID: 41195520. PMC13040192.
163. Ghalamkari S, Hahn CN, Lau A, Scott HS, Homan CC, Brown AL. Targeting RUNX1 Germline Variants: Agents Under Investigation. Curr Hematol Malig Rep. 2025;20(1):23. doi:10.1007/s11899-025-00767-w. PMID: 41398138.
164. Ren Y, Zhang R, Chen X, Zhang X, Wan Y, Li X, et al. Successful allogeneic hematopoietic stem cell transplantation in children with myeloid malignancies secondary to GATA2 deficiency using decitabine-containing busulfan conditioning. Bone Marrow Transplant. 2026. doi:10.1038/s41409-026-02978-y. PMID: 42471475.
165. ClinicalTrials.gov. NCT06022016 — Study of Families With an Hemopathies Predisposition Related to the DDX41 Gene (LUCID; Institut Claudius Regaud). https://clinicaltrials.gov/study/NCT06022016
166. ClinicalTrials.gov. NCT07019155 — Longitudinal Study of Individuals and Families With Aberrations in DDX41 or Similar Cancer Predisposition Variants (NCI). https://clinicaltrials.gov/study/NCT07019155
167. ClinicalTrials.gov. NCT03854318 — Longitudinal Studies of Patients and Families With FPDMM Caused by RUNX1 Germline Variants or FPDMM-Like Conditions (NHGRI). https://clinicaltrials.gov/study/NCT03854318
168. ClinicalTrials.gov. NCT01905826 — The Natural History of GATA2 Deficiency and Related Disorders (NIAID). https://clinicaltrials.gov/study/NCT01905826
169. ClinicalTrials.gov. NCT01861106 — Allogeneic Hematopoietic Stem Cell Transplant for Patients With Mutations in GATA2 or the MonoMAC Syndrome (NCI). https://clinicaltrials.gov/study/NCT01861106
170. Bauer ME, Arendt K, Beilin Y, Gernsheimer T, Perez Botero J, James AH, et al. The Society for Obstetric Anesthesia and Perinatology Interdisciplinary Consensus Statement on Neuraxial Procedures in Obstetric Patients With Thrombocytopenia. Anesth Analg. 2021;132(6):1531-1544. doi:10.1213/ane.0000000000005355. PMID: 33861047. PMC13109829.
171. Roncareggi S, Girardi K, Fioredda F, Pedace L, Arcuri L, Badolato R, et al. A Nationwide Study of GATA2 Deficiency in Italy Reveals Novel Symptoms and Genotype-phenotype Association. J Clin Immunol. 2023;43(8):2192-2207. doi:10.1007/s10875-023-01583-8. PMID: 37837580.
172. RUNX1 Research Program. https://runx1-fpd.org
173. MDS Foundation. https://www.mds-foundation.org/
174. Leukaemia Care (United Kingdom). https://www.leukaemiacare.org.uk/
175. Krebs-Prädisposition. GATA2 deficiency — compendium. https://www.krebs-praedisposition.de/en/kompendium/gata2-deficiency/
176. Pan L, Li Y, Gao H, Lai X, Cai Y, Chen Z, et al. Clinical features and management of germline CEBPA-mutated carriers. Leuk Res. 2024;138:107453. doi:10.1016/j.leukres.2024.107453. PMID: 38442594.