The genetic landscape of haematological malignancies: an evidence-based review

Introduction

Twenty years ago most blood cancers were named by what they looked like under the microscope. Many are now named by their genes. Both 2022 classifications, WHO-HAEM5 and the International Consensus Classification (ICC), define much of acute myeloid leukaemia by a fusion or a mutation, and in some types the genetic finding makes the diagnosis whatever the blast count.^1,2 Genes also set risk and choose treatment. A FLT3 or IDH1 mutation opens the way to a targeted drug, a TP53 mutation changes the prognosis of almost every disease in this review, and a mutation used as a marker after treatment finds disease the microscope misses and predicts relapse.³

The review starts with the kinds of genetic lesion and the pathways they act on, because the same few mechanisms recur across diseases that look unrelated. It then covers the precursor states and inherited predisposition, and takes the diseases in turn: the myeloid neoplasms, the lymphoblastic leukaemias and chronic lymphocytic leukaemia, the lymphomas and myeloma. It ends with clonal evolution, the lesions that have an approved drug, and the tests used in practice.

The kind of evidence is named with each claim: a large sequencing cohort, a randomised trial, a classification or guideline with its body and year, or laboratory work. Mutation frequencies differ between cohorts, so each figure is tied to the cohort it came from. Where WHO-HAEM5 and the ICC disagree, both are given. Thresholds such as a 2% variant allele fraction or a 20% blast count are conventions, and are named as such.

Kinds of genetic lesion

A haematological cancer is defined by the lesions in it, not only by the cell it arose in. The same handful of mechanisms recur across acute leukaemias, lymphomas, myeloproliferative neoplasms and myeloma, which is why a gene named once in this review reappears through the disease-specific sections that follow.

A review of sequencing across common cancer types found around 140 genes that, when altered, can act as drivers, and that a typical tumour carries only two to eight driver-gene mutations; everything else is a passenger mutation, carried along in the same clone without conferring any selective advantage.⁴ Exome sequencing across 3083 tumour-normal pairs found somatic mutation frequencies varying more than 1000-fold across cancer types, with the lowest frequencies in haematological and paediatric tumours and the highest in cancers driven by a known mutagen — melanoma and lung cancer both exceeded 100 mutations per megabase, against a median of 0·37 per megabase in AML, whose own patients still spanned three orders of magnitude (0·01–10 per megabase).⁵ A pan-cancer analysis of nearly five million mutations across 7042 tumours found more than 20 distinct mutational processes, some — a signature attributed to APOBEC cytidine-deaminase activity among them — recurring across many cancer types, and others confined to a single one.⁶ The same 3083-tumour study showed that as sequencing cohorts grow, naive statistical analysis produces long lists of implausible "driver" genes purely from background mutational heterogeneity, and built a correction (MutSigCV) specifically to separate true drivers from this noise.⁵

The variant allele fraction (VAF) — the proportion of sequencing reads at a position that carry the variant — is the working proxy for clone size in a bone marrow or blood sample. A mutation with a high VAF is present in most of the malignant (or pre-malignant) population; a low VAF marks a small subclone. VAF is read alongside the pattern of the mutation itself. Genes such as JAK2, FLT3, NRAS and IDH1/IDH2 are activated by point mutations that recur at the same codon — a hotspot — because only that specific amino-acid change produces the gain of function the clone is selected for. Genes such as TET2, ASXL1, RUNX1 and TP53 are inactivated by mutations scattered across the coding sequence — nonsense, frameshift or missense — because almost any change that disrupts the protein has the same loss-of-function effect. The two patterns point to opposite biology: gain of function usually means an activated enzyme that a drug can be built against; loss of function usually means a missing brake that is much harder to restore pharmacologically.

Balanced chromosomal translocations act by two distinct mechanisms, and haematological cancers use both. In one, an oncogene is left intact but moved next to the strong regulatory elements of an immunoglobulin (IG) or T-cell receptor (TCR) locus, so that a normal protein is produced in the wrong amount at the wrong time: MYC next to an IG enhancer in Burkitt lymphoma, BCL2 next to an IG enhancer in follicular lymphoma, CCND1 next to an IG enhancer in mantle cell lymphoma and in a subset of myeloma. In the other, the breakpoints fall inside two different genes and fuse them into a single novel protein with new or constitutively active function: BCR::ABL1 in chronic myeloid leukaemia (CML) and a subset of B-cell acute lymphoblastic leukaemia (B-ALL), and PML::RARA in acute promyelocytic leukaemia (APL). The nomenclature convention used throughout this review, a double colon between the two gene names, marks a fusion gene rather than a locus.

The lymphoid translocations are not a random hazard of cell division. B and T lymphocytes deliberately break their own DNA as part of normal development — V(D)J recombination by the RAG1/RAG2 recombinase to assemble an antigen receptor, and, in B cells, class-switch recombination and somatic hypermutation driven by activation-induced cytidine deaminase (AID) to diversify and refine the antibody response. Both processes are mechanisms for controlled DNA breakage and repair, and both are recognised sources of the translocations and off-target mutations seen in B-cell lymphomas: a review of the mechanism traces recurrent lymphoma translocations, including those juxtaposing MYC, BCL2 and BCL6 with immunoglobulin loci, to mistakes in this same physiological machinery.⁷ It is the price of a system built to break and rejoin DNA on purpose.

Copy-number change and aneuploidy — the gain or loss of whole chromosomes, chromosome arms or smaller segments — alter the dose of many genes at once rather than the sequence of any one of them. Trisomies drive hyperdiploidy in myeloma; deletions of 5q, 7q or 20q are recurrent in myelodysplastic syndrome (MDS); deletion of 17p removes one copy of TP53 across several diseases. A single catastrophic event can produce dozens of rearrangements on one or a few chromosomes in one step — chromothripsis — rather than the stepwise accumulation classical models assumed; the mechanism was first demonstrated in a chronic lymphocytic leukaemia (CLL) genome, where one chromosome had been shattered and reassembled with 42 rearrangements in what appeared to be a single event.⁸ A complex karyotype (conventionally, three or more unrelated abnormalities) is itself a marker of this kind of genomic catastrophe and is strongly associated with TP53 mutation.

A further class of lesion changes chromatin and DNA methylation rather than any single target gene's sequence: mutations in DNMT3A, TET2, IDH1/IDH2, ASXL1 and EZH2 reset the epigenetic landscape of the whole genome. They are treated as a group under DNA methylation and chromatin, below, because they recur across AML, MDS and the myeloproliferative neoplasms (MPN) rather than marking any one of them. Germline variants — inherited rather than acquired in the blood — are a separate class (see Germline predisposition to myeloid neoplasms). A growing list of genes (RUNX1, GATA2, DDX41, CEBPA, ETV6, the TP53-linked Li-Fraumeni syndrome and the Fanconi and telomere-biology disorders among them) predispose to AML and MDS through a germline first hit, on top of which acquired mutations accumulate later.

A mutational signature is the fingerprint a mutagenic process leaves in the pattern of bases it changes. It is extracted by computer from large sequencing datasets, not read off any single mutation. In myeloma, signatures of APOBEC and of AID activity both contribute, in amounts that change as the disease evolves (see Multiple myeloma). Treatment also shapes the genome, mostly by selecting clones that were already there (see Clonal evolution and co-mutation).

The pathways

The genes above cluster into a smaller number of pathways, and the same pathway is often hit by different mechanisms in different diseases — a point mutation in one, a translocation in another. Grouping by pathway rather than by gene is the more useful map, because it is usually the pathway, not the individual gene, that a drug is built against.

RAS–MAPK and receptor tyrosine kinases

Receptor tyrosine kinases and the RAS–MAPK cascade downstream of them are activated in AML through several different genes that converge on the same signal: FLT3 internal tandem duplications or tyrosine-kinase-domain mutations, KIT mutations, NRAS or KRAS hotspot mutations, PTPN11, NF1 and CBL. A large genomic classification of adult AML grouped these together as one of the recurrent driver classes and showed they frequently co-occur with mutations from the other pathways below rather than occurring alone.⁹ The same kinase logic applies in Philadelphia-chromosome-like (Ph-like) B-ALL, where rearrangements of ABL1, ABL2, CSF1R or PDGFRB produce fusion kinases that dasatinib inhibits in laboratory models (see B-lymphoblastic leukaemia).¹⁰

JAK–STAT

A single acquired point mutation, JAK2 V617F, was reported independently by four groups in 2005 as the driver of most classic MPN (frequencies are given under Myeloproliferative neoplasms).^11–14 In the minority of MPN without a JAK2 mutation, somatic mutations in calreticulin (CALR) were identified in 2013 as an alternative route to the same constitutive JAK–STAT signalling, again reported by two independent groups in the same issue of the same journal.^15,16 MPL mutation is a third, rarer route into the same pathway in essential thrombocythaemia and myelofibrosis. The same axis is reached from further upstream in Ph-like ALL, through rearrangements of CRLF2, EPOR, JAK2 or TYK2 that activate STAT5.¹⁰ Direct mutation of a STAT gene itself, rather than upstream rearrangement, defines a separate and rarer route: somatic STAT3 mutation was the first genetic lesion described in T-cell large granular lymphocytic leukaemia.¹⁷ Loss of the negative regulator SOCS1 removes a brake on the same pathway in several lymphoid cancers, Hodgkin lymphoma among them (see Hodgkin lymphoma).

DNA methylation and chromatin

DNMT3A, encoding a DNA methyltransferase, was found mutated in 22·1% (62 of 281) of a cohort of adults with de novo AML, most often at a single hotspot codon.¹⁸ TET2, whose product oxidises 5-methylcytosine as the first step of active DNA demethylation, is recurrently deleted or mutated across MDS, MPN and AML.¹⁹ IDH1 and IDH2 mutations act by a different route to the same downstream effect: the mutant enzyme acquires a new, neomorphic activity that converts α-ketoglutarate to the oncometabolite 2-hydroxyglutarate (2-HG) instead of the normal reaction.^20,21 Because 2-HG competitively inhibits α-ketoglutarate-dependent enzymes, including TET2 itself and the histone demethylases, IDH-mutant leukaemia cells display a hypermethylation phenotype that converges on the same chromatin state produced directly by TET2 loss, despite the mutations lying in different genes.²² ASXL1, a polycomb-group gene, was first found mutated in myelodysplastic syndrome and chronic myelomonocytic leukaemia;²³ EZH2, the catalytic subunit of polycomb repressive complex 2, is inactivated by mutation in 12% of myelodysplastic/myeloproliferative neoplasms and 13% of myelofibrosis in the founding series.²⁴ In germinal-centre B-cell lymphomas the effect runs the other way. A recurrent change at Tyr641, in the enzyme's SET domain, is found in 21·7% of germinal-centre-type diffuse large B-cell lymphomas and 7·2% of follicular lymphomas, and not in the activated B-cell subtype.²⁵ The change increases the enzyme's histone-methyltransferase activity: it is a gain of function.²⁶ In follicular lymphoma and diffuse large B-cell lymphoma, sequencing of tumour and matched normal DNA found histone-modifying genes as a frequent target of mutation as a class: KMT2D (also called MLL2) was mutated in 32% of DLBCL and 89% of FL cases in the discovery cohort, and MEF2B — a gene that cooperates with the histone acetyltransferases CREBBP and EP300 — in 11·4% and 13·4% respectively.²⁷

RNA splicing

Recurrent mutations in the core RNA-splicing machinery were reported in 2011 as a previously unrecognised driver class in myelodysplasia, affecting SF3B1, SRSF2, U2AF1 and ZRSR2.²⁸ SF3B1 mutation in particular defines the subtype of MDS with ring sideroblasts, identified in the same wave of sequencing studies as one of the most common single mutations in the disease and closely linked to that specific morphological feature.²⁹ A consistent finding across this gene group, confirmed in the founding papers, is that the splicing-factor mutations are close to mutually exclusive with one another within a single patient — consistent with each disrupting the same core spliceosome function, so that a second hit adds little further selective advantage.

Cohesin

The cohesin complex, which holds sister chromatids together and also regulates transcription and DNA repair independently of its role in cell division, is recurrently disrupted by mutation or deletion of its components — STAG2, RAD21, SMC1A and SMC3 — in 12·1% (19 of 157) of AML in the series that first described this class, with the same mutual-exclusivity pattern seen in the splicing genes.³⁰

TP53 and the DNA-damage response

In the myeloid neoplasms TP53 is often inactivated by two hits rather than one, and the number of hits matters: in MDS, disease with both alleles lost ("multi-hit" TP53) is unstable and aggressive, while disease with a single mutant allele behaves much like TP53-wild-type MDS (see Myelodysplastic neoplasms).³¹ ATM, mutated or deleted in a subset of CLL, and PPM1D, a phosphatase that dampens the p53 response and is recurrently mutated after cytotoxic chemotherapy, act on the same DNA-damage-response axis; how chemotherapy selects PPM1D- and TP53-mutant clones is covered under Clonal haematopoiesis.

Lineage transcription factors

A further group of genes acts by disrupting the transcription-factor programme that specifies a blood lineage, rather than by activating a kinase or resetting chromatin. RUNX1, CEBPA and GATA2 are mutated somatically in AML and, in a smaller number of families, carried as germline variants that predispose to it — heritable GATA2 mutation was first described as a cause of familial MDS/AML in 2011.³² In B-lineage ALL, PAX5, IKZF1 and ETV6 are the recurrent targets. IKZF1, encoding the transcription factor Ikaros, is deleted in 83·7% of BCR::ABL1-positive B-ALL but essentially never in chronic-phase CML carrying the same fusion; the deletion was also identified as an acquired event at the point where chronic-phase CML transforms into lymphoid blast crisis, indicating that loss of Ikaros function is specifically involved in the lymphoid, rather than the myeloid, evolution of BCR::ABL1-positive disease.³³

B-cell receptor and NF-κB signalling

A subset of diffuse large B-cell lymphoma (DLBCL), the activated B-cell-like (ABC) subtype, depends on constitutive, "chronic active" signalling through the B-cell receptor complex (CD79A/CD79B) into NF-κB, discovered through an RNA-interference screen that showed ABC-DLBCL cells specifically require components of this pathway to survive.³⁴ Downstream, activating mutations in the scaffold protein CARD11 were found in 7 of 73 ABC-DLBCL biopsies (9·6%), all within the domain that nucleates the signalling complex.³⁵ Negative regulators of the same pathway are lost in parallel: mutations affecting TNFAIP3 (A20) and other NF-κB inhibitors were found in more than half of ABC-DLBCL and a smaller fraction of the germinal-centre B-cell (GCB) subtype.³⁶ MYD88 L265P reaches NF-κB by a different route, through the Toll-like receptor adaptor and IRAK kinases rather than the B-cell receptor. It is recurrent in ABC-DLBCL³⁷ and found in the great majority of Waldenström macroglobulinaemia.³⁸

NOTCH

The first whole-genome sequencing study of CLL, combining four fully sequenced genomes with targeted analysis of 363 further patients, identified NOTCH1 as one of four recurrently mutated genes (with XPO1, MYD88 and KLHL6), and found NOTCH1 and XPO1 mutations concentrated in patients with unmutated immunoglobulin genes — the molecular subgroup of CLL already known to behave more aggressively.³⁹ In splenic marginal zone lymphoma, the coding-genome sequencing study that established the disease's mutational landscape found NOTCH2, rather than NOTCH1, as the recurrently activated member of the pathway.⁴⁰ FBXW7, which normally targets activated NOTCH1 for degradation, is mutated in a smaller but overlapping group of the same diseases, with the same net effect of prolonging NOTCH signalling.

MYC

MYC dysregulation, most often by translocation to an immunoglobulin locus as set out above, is the defining lesion of Burkitt lymphoma and recurs as a secondary event in other B-cell lymphomas and in myeloma (see Multiple myeloma).

Apoptosis and BCL2 dependence

Many B-cell and myeloid cancers survive by over-expressing anti-apoptotic BCL2-family proteins, whether or not BCL2 itself is translocated. This dependence was the rationale for developing venetoclax, a BH3-mimetic small molecule designed to bind BCL2 directly and release the pro-apoptotic proteins it sequesters, restoring the cell's ability to undergo apoptosis; the compound was shown to kill BCL2-dependent tumour cells while sparing platelets, which depend on the related protein BCL-XL instead.⁴¹ Its clinical uses are listed under Targetable lesions and approved agents.

Clonal haematopoiesis: the precursor state

Most myeloid cancers do not begin at diagnosis; they begin years earlier, as a single blood stem cell that acquires a mutation and slowly outgrows its neighbours. When that clone is large enough to detect by sequencing but the person has no cytopenia and no cancer, the finding is called clonal haematopoiesis of indeterminate potential, CHIP. When the same clone is found alongside an otherwise unexplained cytopenia that does not meet the criteria for a myeloid neoplasm, it is called clonal cytopenia of undetermined significance, CCUS. Both are defined by a mutation in a gene recurrently mutated in myeloid cancer, at a variant allele fraction of at least 2% of blood cells — a working convention close to the detection limit of standard sequencing, not a biological threshold below which risk disappears.⁴²

Age is the dominant driver. In a whole-exome analysis of 17,182 people not selected for any blood disorder, detectable clonal mutations were rare before 40 but rose steeply with age: they were found in 9·5% of people aged 70–79, 11·7% of those 80–89, and 18·4% of those 90 and older.⁴³ A second whole-exome study of 12,380 people found clonal haematopoiesis in 10% of those over 65 but only 1% of those under 50.⁴⁴ In both cohorts, the great majority of variants fell in three genes: DNMT3A, TET2 and ASXL1.^43,44 Less common but still recurrent are PPM1D, TP53, the spliceosome genes (SF3B1, SRSF2, U2AF1) and JAK2.⁴⁵

Clonal haematopoiesis is not a bystander finding. In the 17,182-person cohort, carrying a detectable mutation carried a hazard ratio of 11·1 (95% CI 3·9–32·6) for subsequent haematological cancer, and a hazard ratio of 1·4 (1·1–1·8) for death from any cause.⁴³ In the second cohort the hazard ratio for haematological cancer was 12·9 (5·8–28·7), and roughly 42% of the cancers that arose in the cohort occurred in people who already carried the mutation more than six months before diagnosis.⁴⁴ Most carriers of clonal haematopoiesis, however, never develop a myeloid neoplasm; the practical question is which carriers should worry.

A clonal haematopoiesis risk score (CHRS) was built to answer that question. It was derived and validated in 438,890 UK Biobank participants and combines the specific gene mutated, the number of mutations, the variant allele fraction, age, red-cell indices, and whether the clone is accompanied by a cytopenia (CHIP vs CCUS). It separates carriers into three groups with very different 10-year probabilities of progressing to a myeloid neoplasm, from under 1% to 85%, and in independent clinical cohorts most events occurred in the small high-risk group⁴² (Table 3).

Clonal haematopoiesis also changes shape after cancer treatment. Radiation, platinum agents and topoisomerase II inhibitors do not select at random: they preferentially expand clones carrying mutations in the DNA-damage-response genes TP53, PPM1D and CHEK2, whereas smoking instead enriches for ASXL1-mutant clones. Sequential sampling in patients who went on to develop a therapy-related myeloid neoplasm showed that the same DNA-damage-response clone detected years earlier was already present at diagnosis.⁴⁶ This is why a myeloid neoplasm that follows chemotherapy for a solid tumour or lymphoma so often carries a TP53 or PPM1D mutation: the clone was usually there before the second cancer was diagnosed, selected by the first cancer's treatment.

Clonal haematopoiesis has a life outside haematology as well. In four case-control studies totalling 4726 people with coronary heart disease and 3529 controls, CHIP carriers had 1·9 times the risk of coronary heart disease (95% CI 1·4–2·7), and 4·0 times the risk of early-onset myocardial infarction (2·4–6·7), than non-carriers. DNMT3A, TET2, ASXL1 and JAK2 mutations were each individually associated with coronary disease, and engrafting atherosclerosis-prone mice with Tet2-deficient bone marrow produced larger atherosclerotic lesions, supporting a causal rather than purely correlative link.⁴⁵ A patient found to have CHIP incidentally therefore carries two things to track: a small but real haematological risk, and a cardiovascular risk factor not yet part of routine risk calculators.

Germline predisposition to myeloid neoplasms

Both current classifications of myeloid disease carry a dedicated category for cancers arising on a germline predisposition, reflecting a shift from treating adult-onset leukaemia as sporadic by default. The practical reason to look for a germline cause is not only genetic counselling: if the plan is an allogeneic transplant, a sibling or relative donor who carries the same germline variant can silently transmit the predisposition, and in some genes may already have subclinical disease. A 2017 review of clinical surveillance in these syndromes recommended expert referral before donor selection, because presentations can be subtle enough to miss on a routine family history.⁴⁷

DDX41 is the single most frequent cause identified so far. In the largest series assembled, 346 patients with a DDX41 pathogenic or likely pathogenic variant were drawn from 9082 patients with a myeloid neoplasm, and germline DDX41 variants explained roughly 80% of known germline predisposition to myeloid neoplasms diagnosed in adulthood. Risk was negligible before age 40 but rose to 49% by age 90, and carriers were disproportionately male (variant enrichment 20·7-fold in Japanese men with a myeloid neoplasm against the general population, versus 5·0-fold in women).⁴⁸ The disease is unusual among germline syndromes for how late and ordinary it looks: no childhood clues, just an adult with MDS or AML carrying a second, somatic hit in the other DDX41 allele — found in half of patients with a germline variant in the original description,⁴⁹ most often p.R525H, which was also the most frequent somatic DDX41 variant in a Korean series of 33 germline carriers.⁵⁰ Unusually for MDS, a co-occurring TP53 mutation in a DDX41-mutated case did not carry its usual dismal prognosis, and neither the revised nor the molecular IPSS predicted outcome in these patients.⁴⁸ The germline lesion appears to change the biology of everything that follows it.

RUNX1 germline mutation causes familial platelet disorder with predisposition to AML (FPD/AML): an autosomal dominant syndrome of mild thrombocytopenia and qualitative platelet dysfunction from birth, with leukaemia developing later in a substantial minority. The original linkage studies found heterozygous nonsense mutations, an intragenic deletion, or missense mutations at two specific conserved residues (R166 and R201) all converging on the same mechanism: haploinsufficiency of the transcription factor RUNX1 (then called CBFA2). Haploinsufficiency reduced megakaryocyte colony formation, and the authors proposed that it predisposes to the further mutations that cause leukaemia.⁵¹

Germline CEBPA mutation causes a purer familial AML syndrome, without an antecedent platelet defect. The original family carried an identical frameshift mutation (212delC) in three affected relatives, with latency from birth to overt leukaemia of 10, 18 and 30 years; unaffected relatives did not carry the variant. One of the three had a second CEBPA mutation, found only at the diagnosis of leukaemia.⁵²

GATA2 deficiency spans a wider phenotype: familial MDS and AML, but also MonoMAC syndrome (monocytopenia with susceptibility to disseminated non-tuberculous mycobacterial infection), lymphoedema, and sensorineural deafness in some kindreds. The founding description found a recurrent heterozygous missense mutation in the second zinc finger of GATA2 segregating with multigenerational MDS-AML in three families, and a nearby deletion in a fourth.³² Because GATA2-deficient marrow can look like acquired aplastic anaemia or hypocellular MDS, it is easy to miss until a sibling donor is worked up and found to carry the same lesion.

ETV6 germline mutation causes a dominant syndrome of thrombocytopenia with susceptibility to a broad range of blood cancers, not only myeloid. The founding families carried missense variants clustering in the ETS DNA-binding domain and an internal linker domain, at residues that are themselves recurrent sites of acquired somatic mutation in leukaemia — a germline lesion sitting on the same hotspot tumours later find independently.⁵³ ANKRD26-related thrombocytopenia (THC2) is mechanistically distinct: the causal mutations cluster in a short, conserved sequence in the gene's 5′ untranslated region and appear to increase rather than abolish ANKRD26 expression, with leukaemic transformation a recognised long-term risk.⁵⁴

SAMD9 and SAMD9L cause germline bone marrow failure with monosomy 7, by a mechanism unique among these syndromes: the mutations are gain-of-function and growth-restricting, so a cell that loses the mutant allele — through monosomy 7, or, in SAMD9L disease, copy-neutral loss of heterozygosity — gains a proliferative advantage and can found the leukaemic clone. Germline SAMD9 mutation causes MIRAGE syndrome (myelodysplasia, infection, restriction of growth, adrenal hypoplasia, genital phenotypes and enteropathy), first described in 11 patients, two of whom developed MDS with loss of the mutated chromosome 7.⁵⁵ Germline SAMD9L mutation instead causes ataxia-pancytopenia syndrome, cosegregating in two independent kindreds, again followed by acquired loss of the mutant allele over time.⁵⁶

Two further groups belong in the same differential even though they are not, strictly, myeloid-gene syndromes. The telomere biology disorders — dyskeratosis congenita and its milder allelic variants — arise from pathogenic germline variants in telomere maintenance genes; very short telomeres cause a life-threatening spectrum of bone marrow failure, pulmonary and hepatic disease, and elevated cancer risk, principally of myeloid neoplasms.⁵⁷ Fanconi anaemia, the prototypical inherited bone-marrow-failure syndrome from a defect in DNA interstrand-crosslink repair, is usually diagnosed in childhood on marrow failure plus birth defects, but up to 10% of published cases are diagnosed only in adulthood, typically without the classic paediatric features and instead presenting with early head-and-neck or gynaecological squamous-cell carcinoma — a subgroup easy to miss in a clinic not looking for it.⁵⁸

None of these are rare curiosities to a transplant team. The shared clinical lesson is the same across all of them: a family history of "mild" cytopenia, unexplained early bone marrow failure, or a cluster of myeloid cancers in relatives should trigger germline testing before a related donor is chosen, because the donor may carry, and could later manifest, the same lesion.⁴⁷

Acute myeloid leukaemia

The driver landscape

The first systematic map of AML's genome came from The Cancer Genome Atlas, which sequenced 200 adult de novo cases. AML turned out to be genetically quiet compared with most solid cancers — an average of only 13 mutated genes per case, about five of them recurrently mutated across AML. The study nonetheless found at least one plausible driver in nearly every sample, across nine functional categories: transcription-factor fusions (18% of cases), NPM1 mutation (27%), tumour-suppressor mutation (16%), DNA-methylation genes (44%), signalling genes (59%), chromatin-modifying genes (30%), myeloid transcription-factor genes (22%), cohesin-complex genes (13%) and spliceosome genes (14%).⁵⁹ A larger analysis of 1540 patients on three prospective trials grouped these drivers into 11 genomic classes defined by co-mutation pattern rather than any single lesion, highlighting three classes beyond the already-recognised subtypes: chromatin/splicing mutations (18% of patients, poor outcome), TP53 mutation and/or aneuploidy (13%, poor outcome), and a small provisional group defined by IDH2(R172) mutation (1%).⁹ The Beat AML programme, pairing genomic and transcriptomic profiling of 672 specimens from 562 patients with ex vivo drug-sensitivity testing, extended this from classification to treatment selection, showing drug response tracked specific mutations and mutation combinations rather than any one lesion alone.⁶⁰ Saudi data are sparse. In 87 adults with de novo AML treated at one centre in Jeddah between 2011 and 2018, the most frequent cytogenetic abnormality was t(15;17) (17·2%), the most frequent molecular abnormalities were PML::RARA (16·1%) and NPM1 mutation (11·5%), and FLT3-ITD was independently associated with lower overall survival.⁶¹ A single-centre series of this size cannot give population frequencies; the high share of acute promyelocytic leukaemia may reflect referral pattern.

Grouping the individual genes by function is more useful clinically than any single list. Signalling genes — FLT3 (internal tandem duplication and tyrosine-kinase-domain point mutation), KIT, and the RAS genes — drive proliferation and carry the most mature targeted therapy. NPM1 mutation, present in 27% of the TCGA cohort,⁵⁹ is its own category, usually favourable in isolation, but that effect is modulated by co-mutations. DNMT3A, TET2, and IDH1/IDH2 alter DNA methylation; mutant IDH1 and IDH2 proteins each neomorphically produce the oncometabolite 2-hydroxyglutarate, which blocks differentiation and is directly druggable. The chromatin, spliceosome and cohesin genes overlap heavily with the myelodysplasia-related mutations described below and mark a poorer-risk group without a rescuing lesion. TP53 mutation defines its own aggressive category, largely independent of everything else on this list. RUNX1 and CEBPA matter differently depending on how they are mutated: WHO-HAEM5 recognises AML with CEBPA mutation whether the lesion is biallelic (any location) or a single mutation confined to the basic leucine zipper (bZIP) region, and it is specifically the bZIP-region cases whose favourable prognosis has been demonstrated across cohorts of children and adults up to age 70,¹ whereas somatic RUNX1 mutation on its own overlaps too many other categories to define a distinct entity.

Fusions and structural drivers

A minority of AML is defined by a chromosomal rearrangement producing a fusion oncoprotein, and several of these fusions define AML regardless of blast count. PML::RARA, from t(15;17), causes acute promyelocytic leukaemia and is the most treatment-responsive AML subtype known. RUNX1::RUNX1T1, from t(8;21), and CBFB::MYH11, from inv(16) or t(16;16), make up the core-binding-factor leukaemias, generally favourable. KMT2A rearrangements, with a large and still-growing list of fusion partners, carry an adverse prognosis and occur in up to 10% of acute leukaemias.⁶² DEK::NUP214, from t(6;9), and rearrangements of MECOM at 3q26 are both adverse, as are the growing list of NUP98 rearrangements, particularly in children and young adults. Both KMT2A-rearranged and NPM1-mutated leukaemia depend on the interaction between menin and the KMT2A protein. This is why the two, unrelated at the DNA level, share a drug target (see Genotype-directed therapy, below).⁶²

Co-mutation and clonal architecture

AML genomes are not random collections of independent hits; mutations accumulate in a preferred order and co-occur in preferred combinations. The archetype is NPM1-mutated AML with concurrent DNMT3A and FLT3-ITD mutation: age-related clonal haematopoiesis typically establishes a DNMT3A-mutant clone first, NPM1 mutation follows and is usually the disease-defining event, and FLT3-ITD is often a late, sometimes subclonal, addition driving the proliferative phase of overt leukaemia. The 1540-patient genomic-classification analysis found that co-mutation systematically modulates the prognosis of the disease-defining lesion: patterns of co-mutation split NPM1-mutated AML into favourable and adverse groups a single-gene view cannot see, and the prognostic effect of one driver mutation was often significantly altered by the presence or absence of others.⁹

Genetic classification: WHO-HAEM5 and the ICC

Two classifications of AML were published within months of each other in 2022, and clinicians will meet both. WHO-HAEM5 and the ICC agree on far more than they disagree on, but the disagreements are exactly the kind that change a diagnosis. WHO-HAEM5 removed the 20% blast requirement for AML with a defining genetic abnormality, keeping it only for AML with BCR::ABL1 fusion and AML with CEBPA mutation; every other genetically defined type follows the genetics once judged to be driving the disease, whatever the blast count.¹ The ICC instead lowered the threshold: AML with a recurrent genetic abnormality needs ≥10% blasts (≥20% for BCR::ABL1). AML with mutated TP53, with myelodysplasia-related changes, or not otherwise specified still needs ≥20%, and the same disease with 10–19% blasts is called MDS/AML, an explicit intermediate category.² A comparative review of both frameworks singles out this blast-threshold divergence, the MDS-AML interface, and the handling of TP53-mutated disease as the three areas where the two systems produce materially different diagnoses on the same bone marrow.⁶³

The two systems also disagree on which genes define the myelodysplasia-related category — AML arising from an antecedent MDS-type genetic background rather than de novo. WHO-HAEM5 uses eight genes: ASXL1, BCOR, EZH2, SF3B1, SRSF2, STAG2, U2AF1 and ZRSR2, plus a defined list of cytogenetic abnormalities (complex karyotype, deletions or losses of 5q, 7q, 11q, 12p, 13, 17p, and related rearrangements).¹ The ICC uses the same eight plus a ninth, RUNX1, folding the formerly separate provisional entity of AML with mutated RUNX1 into this category;² WHO-HAEM5 took the opposite view, judging that somatic RUNX1 mutation overlaps too broadly with other defining lesions to stand as its own entity.¹ The two systems handle TP53 differently. WHO-HAEM5 has no separate category for TP53-mutated AML; its TP53-defined type is MDS with biallelic TP53 inactivation.¹ The ICC has a group of myeloid neoplasms with mutated TP53: any somatic TP53 mutation with a variant allele fraction above 10% qualifies for AML or MDS/AML with mutated TP53, whereas MDS with mutated TP53 needs a multi-hit mutation, or a mutation above 10% with a complex karyotype.²

European LeukemiaNet 2022 genetic risk

Genetic risk stratification for AML is set out by the European LeukemiaNet (ELN), most recently updated in 2022 by an international expert panel.⁶⁴ An independent validation cohort of 1118 intensively treated patients set out exactly what changed from the 2017 version.⁶⁵ First, the FLT3-internal-tandem-duplication allelic ratio was dropped: a patient with FLT3-ITD, without a core-binding-factor rearrangement or an adverse-risk marker, is now intermediate-risk regardless of the mutant-to-wild-type ratio or NPM1 co-mutation that used to allow a favourable call. Second, favourable-risk CEBPA mutation was redefined around a single feature — an in-frame mutation in the basic leucine zipper (bZIP) domain, whether it occurs alone or with a second CEBPA mutation — rather than requiring two mutated alleles as the 2017 version had. Third, mutations in BCOR, EZH2, SF3B1, SRSF2, STAG2, U2AF1 or ZRSR2, in the absence of a favourable-risk marker, were added to the adverse-risk group, alongside the new fusions t(8;16)/KAT6A::CREBBP and t(3q26.2;v)/MECOM. In the validation cohort these changes reclassified 15% of patients — 3% to a more favourable group, 12% to a more adverse one, mostly on the strength of the newly adverse gene mutations.⁶⁵

Measurable residual disease by NPM1 quantitative PCR

Because NPM1 mutation is both common and leukaemia-specific, quantifying the mutant transcript by reverse-transcriptase quantitative PCR after treatment is one of the most sensitive tools available for detecting residual disease. In the UK NCRI AML17 trial, 2569 samples from 346 patients with NPM1-mutated AML were tracked through treatment. Persistence of NPM1-mutated transcripts after the second chemotherapy cycle, seen in 15% of patients, predicted a 3-year relapse rate of 82% against 30% in patients who had cleared the transcript (hazard ratio 4·80, 95% CI 2·95–7·80), and a survival rate of 24% against 75% (hazard ratio for death 4·38, 2·57–7·47). Sequencing split the same patients into more than 150 co-mutation subgroups, too many to predict outcome; residual disease was the only independent prognostic factor for death (hazard ratio 4·84, 2·57–9·15).³ The mutation identifies the disease; its clearance, or not, predicts relapse.

Genotype-directed therapy

Genotype increasingly dictates treatment as well as risk. Adding midostaurin to induction and consolidation chemotherapy in 717 newly diagnosed FLT3-mutated patients (screened from 3277) improved overall survival over chemotherapy alone in the RATIFY trial.⁶⁷ In relapsed or refractory FLT3-mutated AML, gilteritinib extended median overall survival to 9·3 months against 5·6 months with salvage chemotherapy (hazard ratio 0·64, 0·49–0·83) in ADMIRAL,⁶⁸ and in newly diagnosed FLT3-ITD-positive AML, quizartinib added to chemotherapy raised median overall survival to 31·9 months against 15·1 months with placebo (hazard ratio 0·78, 0·62–0·98) in QuANTUM-First.⁶⁹ IDH1 mutation, present in 6–10% of AML, is targeted by ivosidenib, which produced complete remission or complete remission with partial recovery in 30·4% of a relapsed/refractory population as monotherapy,⁷⁰ and, combined with azacitidine in patients unfit for intensive chemotherapy, nearly tripled median overall survival (24·0 against 7·9 months) in the AGILE trial.⁷¹ A second IDH1 inhibitor, olutasidenib, gave complete remission or complete remission with partial recovery in 35% of 153 patients with relapsed or refractory IDH1-mutated AML.⁷² IDH2 mutation, present in around 12% of AML, is targeted by enasidenib, with a 40·3% overall response rate in relapsed or refractory disease and a median survival of 19·7 months among those achieving complete remission.⁷³ Most recently, the shared dependence of NPM1-mutated and KMT2A-rearranged AML on the menin–KMT2A interaction has been targeted directly: the first-in-human trial of revumenib, which enrolled both myeloid and lymphoblastic leukaemias with KMT2A rearrangement or NPM1 mutation, reported 30% complete remission or complete remission with partial recovery in heavily pretreated disease, with asymptomatic QT prolongation the only dose-limiting toxicity.⁶² The trial was not powered to give a response rate for either lineage alone. Ziftomenib produced complete remission in 7 of 20 NPM1-mutated patients (35%) at the recommended dose in its phase 1 portion,⁷⁴ with a subsequent 92-patient registration cohort reaching a 22% complete remission rate (95% CI 14–32, p=0·0058), 61% of them MRD-negative.⁷⁵ Both drugs are now approved in the US for relapsed or refractory disease; the indications and dates are in the table under Targetable lesions and approved agents.

Myelodysplastic neoplasms

WHO-HAEM5 renamed the myelodysplastic syndromes "myelodysplastic neoplasms"; the ICC keeps the older name.^1,2 The genomic landscape of MDS was mapped in two large sequencing studies published within a year of each other. Sequencing 111 genes across 738 patients with MDS or a closely related neoplasm found at least one oncogenic mutation in 78% of patients and complex pairwise associations between genes, especially between the spliceosome and epigenetic-modifier categories, supporting the idea that the earliest driver — usually a splicing-factor mutation — sets a trajectory later mutations follow rather than redirect.⁷⁶ Sequencing 104 genes across 944 patients found at least one mutation in 89·5% (median 3 per patient), with TET2, SF3B1, ASXL1, SRSF2, DNMT3A and RUNX1 each mutated in more than 10% of cases; combining 14 of these genes with clinical factors separated patients into four risk groups with 3-year survival from 95·2% down to 5·3%.⁷⁷

SF3B1 is the single mutation in MDS with the clearest link between genotype and morphology. In the original description, SF3B1 was mutated in 72 of 354 patients with MDS (20%), and in 53 of 82 patients whose disease was specifically characterised by ring sideroblasts (65%).²⁹ This link is now a diagnostic entity in both classifications — MDS with low blasts and SF3B1 mutation in WHO-HAEM5, and MDS with mutated SF3B1 (variant allele fraction ≥10%) in the ICC.^1,2 A 2020 proposal from the International Working Group for the Prognosis of MDS set out diagnostic criteria for SF3B1-mutant MDS — cytopenia, a somatic SF3B1 mutation, morphological dysplasia with or without ring sideroblasts, and bone marrow blasts under 5% (peripheral blood blasts under 1%) — and noted that the entity carries a relatively favourable prognosis and often responds to luspatercept with abolition of the transfusion requirement.⁷⁸

MDS with isolated del(5q) is the oldest example of a cytogenetically defined MDS entity and remains, notwithstanding the newer genetic entities, one of the few MDS subtypes defined by a single lesion rather than a risk score. The deleted region's disease gene was identified by an RNA-interference screen as RPS14: partial loss of RPS14 function reproduces the disorder's defect in erythroid differentiation in normal progenitor cells, and restoring RPS14 expression rescues it in patient-derived marrow, linking the 5q- syndrome's biology to the ribosomopathy that also causes Diamond–Blackfan anaemia.⁷⁹ Clinically, the lesion also predicts treatment response: in 148 patients with transfusion-dependent MDS and a 5q31 deletion, lenalidomide produced transfusion independence in 67% (99 of 148), and of 85 evaluable patients, 62 had cytogenetic improvement, with 38 reaching complete cytogenetic remission.⁸⁰

TP53 mutation in MDS behaves entirely differently depending on whether both copies of the gene are functionally lost. Analysing 3324 patients with MDS for TP53 mutation and allelic imbalance found that one-third of TP53-mutated patients carried a single (monoallelic) mutation, while two-thirds had multiple hits consistent with biallelic inactivation ("multi-hit"). The established associations of mutant TP53 with complex karyotype, few co-occurring mutations, high-risk disease and poor survival held specifically for the multi-hit group; monoallelic patients did not differ from TP53 wild-type patients in outcome or treatment response, and multi-hit state predicted death and leukaemic transformation independently of the Revised International Prognostic Scoring System.³¹ Both classifications build this into MDS. WHO-HAEM5 defines MDS with biallelic TP53 inactivation as two or more TP53 mutations, or one mutation with TP53 copy-number loss or copy-neutral loss of heterozygosity;¹ the ICC's MDS with mutated TP53 is set out under Genetic classification, above.² Because the definitions differ, a patient can fall inside one and outside the other, and outcome data from studies using different systems are hard to compare.⁶³

Prognostic scoring in MDS has moved from clinical and cytogenetic parameters alone to include the mutation data above. The Revised International Prognostic Scoring System (IPSS-R), built from a combined database of 7012 patients (against 816 in the original IPSS), refined cytogenetic risk from three to five subgroups, split blast-percentage and cytopenia weighting more finely, and produced five rather than four prognostic categories.⁸¹ The Molecular IPSS (IPSS-M), a decade later, added somatic mutation data from 152 sequenced genes in 2957 patients, validated in 754 Japanese patients. TP53 multi-hit state, FLT3 mutation and partial tandem duplication of KMT2A (MLL-PTD) were the strongest independent predictors of adverse outcome, while SF3B1 mutation was independently favourable, modulated by co-mutation. The final model uses haematological parameters, cytogenetics and mutations in 31 genes to give six risk categories. It improved discrimination for every endpoint tested and restratified 46% of patients against the IPSS-R.⁸² Of the 31 genes, 16 carry an independent "main effect" and 15 are "residual" genes that add risk when mutated.⁸³

Two newer drugs treat anaemia in lower-risk MDS. Neither targets a mutation. Luspatercept, which reduces SMAD2/3 signalling downstream of the transforming growth factor β superfamily, produced transfusion independence for at least 8 weeks in 38% of transfusion-dependent, lower-risk MDS patients with ring sideroblasts against 13% on placebo in the MEDALIST trial.⁸⁴ MEDALIST selected by ring sideroblasts, not by genotype, but ring-sideroblast MDS is largely SF3B1-mutant (65% in the series above).²⁹ Imetelstat, a telomerase inhibitor, produced 8-week transfusion independence in 40% of heavily transfused lower-risk MDS patients ineligible for or refractory to erythropoiesis-stimulating agents, against 15% on placebo, in the IMerge trial, with responses lasting roughly a year.⁸⁵

Myeloproliferative neoplasms

Polycythaemia vera, essential thrombocythaemia and primary myelofibrosis are driven, in almost all cases, by a mutation that switches on JAK–STAT signalling without a cytokine to trigger it. Four groups reported the same mutation within weeks of each other in 2005: a single acquired substitution in JAK2, valine to phenylalanine at codon 617 (V617F), in the kinase's regulatory pseudokinase (JH2) domain.^11–14 In one of the founding cohorts, JAK2 V617F was present in 97% of 73 patients with polycythaemia vera, 57% of 51 with essential thrombocythaemia, and 50% of 16 with primary myelofibrosis.¹⁴ A minority of V617F-negative polycythaemia vera is explained by a different set of mutations in exon 12 of the same gene, first described in ten V617F-negative patients with polycythaemia vera or isolated erythrocytosis; these patients typically present with erythrocytosis alone and a low serum erythropoietin.⁸⁶

JAK2-negative essential thrombocythaemia and myelofibrosis were, until 2013, unexplained in most cases. Two groups, sequencing patients with primary myelofibrosis or essential thrombocythaemia who had neither a JAK2 nor an MPL mutation, independently found frameshift insertions or deletions in exon 9 of CALR, encoding calreticulin, that all generate the same novel, mutant C-terminal peptide.^15,16 CALR mutations were absent from polycythaemia vera and mutually exclusive with JAK2 and MPL mutations elsewhere. Among JAK2/MPL-unmutated patients, CALR mutations were found in 67% of essential thrombocythaemia and 88% of primary myelofibrosis in one series,¹⁵ and in 70–84% of JAK2-unmutated myeloproliferative neoplasm samples, and in 8% of myelodysplasia samples, in the other.¹⁶ The third driver, an activating substitution in the thrombopoietin receptor gene MPL (W515L), had been described the previous year in JAK2-negative myelofibrosis;⁸⁷ a 1182-patient screen across myeloid disorders then put its frequency at roughly 5% in myelofibrosis and 1% in essential thrombocythaemia, and found it absent from polycythaemia vera, with six patients carrying an MPL mutation alongside JAK2 V617F.⁸⁸

A minority of patients with essential thrombocythaemia or primary myelofibrosis carry none of the three drivers and are termed triple-negative. In a 254-patient Mayo Clinic cohort of primary myelofibrosis, 58% carried JAK2, 25% CALR, and 8·3% MPL, with 8·7% triple-negative and one patient carrying both JAK2 and CALR.⁸⁹ This is a large single-centre cohort rather than a population-representative sample, and the proportions are not identical to the discovery cohorts above.

Phenotype, thrombosis and prognosis by driver

CALR-mutated patients differ clinically from those with JAK2, and the two common types of CALR mutation (a 52 bp deletion, type 1; a 5 bp insertion, type 2) are not interchangeable in prognostic models, which treat "type 1/like" mutations as the favourable group.⁹⁰ Across myeloproliferative neoplasms, CALR-mutated patients present with higher platelet counts and lower haemoglobin than JAK2-mutated patients, and had a lower risk of thrombosis and longer overall survival in the discovery cohort.¹⁵ In the 254-patient primary myelofibrosis cohort, CALR mutation was associated with younger age, higher platelet count and a lower DIPSS-plus score, and carried a favourable effect on survival independent of both DIPSS-plus risk and ASXL1 status (hazard ratio 3·4 for triple-negative and 2·2 for JAK2-mutated patients, relative to CALR-mutated).⁸⁹ In a larger, two-centre cohort of 1581 patients with molecularly annotated essential thrombocythaemia, polycythaemia vera or myelofibrosis, mutational status did not affect the survival advantage of essential thrombocythaemia over polycythaemia vera, but in the 428 evaluable myelofibrosis patients outcome tracked the driver: median survival was 16 years with CALR mutation against 2·3 years for triple-negative disease (hazard ratio 5·1, 95% CI 3·2–8·0), and blast transformation occurred in 6·5% of CALR-mutated versus 25% of triple-negative patients (hazard ratio 7·6, 95% CI 2·8–20·2).⁹¹ Median overall survival across the whole cohort was 19·8 years in essential thrombocythaemia, 13·5 years in polycythaemia vera, and 5·9 years in primary myelofibrosis.⁹¹

Order of mutation acquisition and the clonal timeline

Most patients with a myeloproliferative neoplasm who also carry a TET2 mutation acquire the two mutations in a definite order, and the order matters. Genotyping haematopoietic colonies and next-generation sequencing in patients with both mutations showed that patients in whom JAK2 was acquired first ("JAK2-first") were more likely to present with polycythaemia vera than essential thrombocythaemia, had a higher risk of thrombosis, and had JAK2-mutant progenitors that were more sensitive to ruxolitinib in vitro than patients in whom TET2 was acquired first ("TET2-first"); prior TET2 mutation blunted the transcriptional, proliferative effect of JAK2 V617F on stem and progenitor cells.⁹² This is mechanistic and clonal-lineage work in patient-derived cells, not a clinical trial, but it establishes that mutation order, not only mutation presence, shapes phenotype.

Whole-genome sequencing of clonal haematopoietic colonies has also put a number on how early these driver mutations arise. In twelve patients with a myeloproliferative neoplasm, phylogenetic reconstruction from 580,133 somatic mutations across 1013 colonies estimated that JAK2 V617F was acquired by 33 weeks of gestation to 10·8 years of age in the five patients in whom it was the first event, and that the mean latency between JAK2 V617F acquisition and diagnosis was 30 years (range 11–54 years).⁹³ A separate lineage-tracing study of two patients reached a similar conclusion by a different route, dating the JAK2 V617F mutation to a single haematopoietic stem cell at age 9±2 years in a person diagnosed at 34, and at age 19±3 years in a person diagnosed at 63, decades before either presented.⁹⁴ Both are small studies — twelve and two patients respectively — but they converge on the same picture: the driver mutation is typically decades old by the time a myeloproliferative neoplasm is diagnosed, with childhood or even fetal acquisition in some patients.

High-molecular-risk mutations and genetics-based prognostic models

Beyond the three drivers, a further group of mutations — in ASXL1, EZH2, SRSF2, IDH1, IDH2, and U2AF1 at codon Q157 — carries independent adverse prognostic weight in primary myelofibrosis and is termed high-molecular-risk. Three prognostic models incorporate these mutations alongside clinical and, in two of them, cytogenetic data (Table 7). MIPSS70, built and validated in 805 transplantation-age patients with primary myelofibrosis, selects haemoglobin under 100 g/L, leukocytes over 25×10⁹/L, platelets under 100×10⁹/L, circulating blasts of 2% or more, marrow fibrosis grade 2 or higher, constitutional symptoms, absence of a type-1/like CALR mutation, and the presence of one or two-or-more high-molecular-risk mutations; three risk groups had 5-year overall survival of 95%, 70% and 29%. The companion model that adds cytogenetics, MIPSS70-plus, produced four groups at 91%, 66%, 42% and 7% 5-year survival (hazard ratios 3·2, 6·4 and 17·0 relative to low risk).⁹⁰ A later revision, MIPSS70+ version 2.0, adds more detailed karyotype and mutation information.⁹⁵ A purely genetic alternative, GIPSS, uses only karyotype and mutation status — no clinical variables — and in 641 cytogenetically annotated patients delineated four groups with 5-year survival of 94%, 73%, 40% and 14%, with predictive accuracy reported as comparable to MIPSS70-plus.⁹⁶

JAK inhibitors and interferon

None of the approved JAK inhibitors is clone-selective: they block the downstream kinase signal from the mutant allele but do not eliminate the mutant clone, so molecular response is partial at best and the drugs are used for symptom and spleen control rather than as disease-modifying therapy in the way a tyrosine kinase inhibitor is in chronic myeloid leukaemia (below). Ruxolitinib, a JAK1/2 inhibitor, reduced spleen volume by 35% or more at 24 weeks in 41·9% of patients versus 0·7% on placebo in the COMFORT-I trial, with an overall survival benefit at follow-up (hazard ratio 0·50, 95% CI 0·25–0·98).⁹⁷ The companion COMFORT-II trial, against best available therapy, found the same spleen-response endpoint met by 28% on ruxolitinib against 0% on best available therapy at 48 weeks.⁹⁸ Fedratinib, a JAK2-selective inhibitor, met the same 24-week spleen-response endpoint in 36–40% of patients against 1% on placebo in a 289-patient phase 3 trial, and Wernicke encephalopathy was reported in four women on the higher dose (supported by MRI in three, suspected clinically in the fourth).⁹⁹ Fedratinib received its first global regulatory approval in the USA in August 2019, for adults with intermediate-2 or high-risk primary or secondary myelofibrosis.¹⁰⁰ Pacritinib, which does not require dose reduction for thrombocytopenia, outperformed best available therapy (including ruxolitinib) for spleen response in patients with myelofibrosis and a platelet count of 100×10⁹/L or less in the PERSIST-2 trial (twice-daily pacritinib 22% versus 3%, p=0·001).¹⁰¹ Momelotinib, which also inhibits activin A receptor type 1 and so addresses the anaemia that the other three do not, achieved a 50% or greater reduction in symptom score in 25% of patients against 9% on danazol in the MOMENTUM trial of JAK-inhibitor-exposed patients with anaemia and myelofibrosis.¹⁰²

In polycythaemia vera, ropeginterferon alfa-2b, a monopegylated interferon, was compared with hydroxyurea in the PROUD-PV trial and its CONTINUATION-PV extension; the primary endpoint, complete haematological response with normal spleen size at 12 months, was met by 21% on ropeginterferon against 28% on hydroxyurea, and non-inferiority was not shown. Responses to ropeginterferon kept rising: at 36 months in CONTINUATION-PV, complete haematological response with improved disease burden was 53% against 38% (p=0·044).¹⁰³ The authors also report durable molecular responses. This is why interferon is favoured in younger patients, where reducing the mutant clone is a realistic goal; whether it changes long-term outcome is not established.

Chronic myeloid leukaemia

Chronic myeloid leukaemia was the first human cancer traced to a specific, recurrent chromosomal abnormality: the "Philadelphia chromosome", a shortened chromosome 22 first reported in 1960, and shown in 1973 by quinacrine fluorescence and Giemsa banding to be the product of a translocation, t(9;22).¹⁰⁴ The translocation fuses BCR on chromosome 22 with ABL1 on chromosome 9, producing the fusion oncoprotein BCR::ABL1, a constitutively active tyrosine kinase. The breakpoint within BCR determines which fusion transcript and protein result — p210 in most chronic-phase disease, p190 more often associated with Philadelphia-positive acute lymphoblastic leukaemia,¹⁰⁵ and the rare p230 associated with a more indolent, neutrophilic phenotype¹⁰⁶ — though the underlying mechanism, constitutive kinase activation driving proliferation independent of normal growth-factor control, is shared across all three.

The tyrosine kinase inhibitor paradigm

Imatinib, a small-molecule inhibitor of the BCR::ABL1 kinase, was compared with interferon alfa plus cytarabine in newly diagnosed chronic-phase chronic myeloid leukaemia in the IRIS trial.¹⁰⁷ Long-term follow-up, focused on the imatinib group because 65·6% of those assigned to interferon and cytarabine had crossed over, found a median follow-up of 10·9 years, an estimated 10-year overall survival of 83·3%, and a complete cytogenetic response in 82·8% of those originally assigned to imatinib, with serious drug-related adverse events uncommon and occurring mostly in the first year.¹⁰⁸ This established the template that the rest of the field has followed since: a targeted kinase inhibitor turning a once-fatal leukaemia into a chronic disease with, for most chronic-phase patients, a life expectancy that approaches that of the general population.¹⁰⁹

Resistance mutations and later-generation inhibitors

Resistance to imatinib is, in most cases examined, associated with reactivation of BCR::ABL1 signalling — most often a point mutation in the kinase domain, and in a minority of cases progressive gene amplification.¹¹⁰ In the first series, six of nine resistant patients carried the same substitution, threonine to isoleucine at position 315 (T315I), at a residue that forms a critical hydrogen bond with the drug.¹¹⁰ T315I also resists the second-generation inhibitors.¹¹¹ Ponatinib, designed to bind BCR::ABL1 despite the T315I substitution, produced a major cytogenetic response in 56% of 267 chronic-phase patients overall and in 70% of those specifically selected for the T315I mutation in a phase 2 trial of heavily pretreated patients.¹¹¹ Asciminib works by a different mechanism: rather than competing at the ATP-binding site, it is a STAMP inhibitor (Specifically Targeting the ABL Myristoyl Pocket), an allosteric site distinct from where the earlier inhibitors bind, and so remains active against most ATP-site resistance mutations. In the phase 3 ASCEMBL trial, asciminib produced a major molecular response at 24 weeks in 25·5% of patients previously treated with two or more tyrosine kinase inhibitors, against 13·2% on bosutinib, with fewer grade 3-or-higher adverse events and fewer discontinuations for toxicity.¹¹² Moved into the first-line setting in the ASC4FIRST trial, asciminib produced a major molecular response at 48 weeks in 67·7% of newly diagnosed patients against 49·0% on an investigator-selected tyrosine kinase inhibitor (69·3% versus 40·2% within the imatinib-comparator stratum specifically), again with fewer grade 3-or-higher adverse events; a head-to-head comparison against second-generation inhibitors specifically was not a primary objective of the trial.¹¹³

Additional mutations in disease progression

Transformation from chronic phase to blast phase is accompanied by an accumulation of mutations beyond BCR::ABL1 itself. A deep-sequencing study of patients in blast crisis detected additional mutations in 76·9% of cases.¹¹⁴ A subsequent genetic and transcriptomic profiling study of 59 patients found recurrent somatic mutations in blast phase in ABL1 (37%), ASXL1 (26%), RUNX1 (16%) and BCOR (16%), with mutation signatures in blast phase resembling those of de novo acute myeloid leukaemia, and found that patients responding poorly to treatment carried a higher mutation load than optimal responders at diagnosis and at follow-up.¹¹⁵ This is consistent with the broader picture in myeloid malignancy: progression is a matter of additional, cooperating mutations accumulating on the founding lesion, not a change in the founding lesion itself.

Molecular monitoring

Response in chronic myeloid leukaemia is tracked by quantitative PCR for BCR::ABL1 transcript, reported on the International Scale, a standardised scale that lets laboratories compare results against a common reference. A major molecular response is conventionally defined as BCR::ABL1 transcript levels of 0·1% or less on this scale — a three-log reduction from a standardised baseline — and is the milestone used as a trial endpoint in the studies above.¹¹³ The European LeukemiaNet's 2020 recommendations, the current guideline against which practice is measured, set a tyrosine kinase inhibitor as first-line treatment, recommend quantitative PCR monitoring wherever available, define greater than 10% BCR::ABL1 at three months as treatment failure once confirmed, and describe treatment discontinuation for treatment-free remission as an option in patients with a durable deep molecular response.¹¹⁶

Myelodysplastic/myeloproliferative overlap and related entities

A group of myeloid neoplasms sits between the myelodysplastic and myeloproliferative categories, combining dysplastic and proliferative features, and each has its own characteristic genetics.

Chronic myelomonocytic leukaemia

Chronic myelomonocytic leukaemia (CMML) carries at least one somatic mutation in the great majority of patients — 199 of 214 (93%) in one series — but no single mutation defines it the way JAK2 or BCR::ABL1 define the neoplasms above.¹¹⁷ Adding genetics to the existing clinical and cytogenetic prognostic score (CPSS) improved risk stratification: in a model built on cytogenetic abnormalities together with mutations in RUNX1, NRAS, SETBP1 and ASXL1, the resulting CPSS-molecular defined four risk groups with median overall survival ranging from more than 144 months to 18 months (hazard ratio 2·69) and 4-year cumulative incidence of leukaemic evolution ranging from 0% to 48% (hazard ratio 3·84), and this held up in an independent 260-patient validation cohort.¹¹⁷

Atypical chronic myeloid leukaemia and chronic neutrophilic leukaemia

Atypical chronic myeloid leukaemia is BCR::ABL1-negative despite the clinical resemblance to the eponymous disease above, and instead carries recurrent somatic mutations in SETBP1, most clustered between codons 858 and 871 and identical to the changes seen in the germline disorder Schinzel–Giedion syndrome; in a targeted resequencing series, 17 of 70 atypical chronic myeloid leukaemia cases (24·3%, 95% CI 16–35%) carried a SETBP1 mutation, and mutated patients had higher white-cell counts and worse prognosis.¹¹⁸ A second, smaller recurrent lesion in the same disease is ETNK1, found in 6 of 68 atypical chronic myeloid leukaemia cases (8·8%) and in 2 of 77 chronic myelomonocytic leukaemia cases (2·6%) in one series, clustering at two residues in the kinase domain and reducing the enzyme's catalytic activity in functional assays.¹¹⁹ Chronic neutrophilic leukaemia, a rarer and closely related entity, is defined in most cases by activating mutations in CSF3R, the granulocyte colony-stimulating factor receptor gene: 16 of 27 patients (59%) with chronic neutrophilic leukaemia or atypical chronic myeloid leukaemia carried a CSF3R mutation in the discovery series, with the mutations segregating into two regions that signal preferentially through either SRC-family/TNK2 or JAK kinases and predict differential sensitivity to kinase inhibitors; one patient with a JAK-activating mutation improved markedly on ruxolitinib.¹²⁰

Systemic mastocytosis

Systemic mastocytosis is driven, in almost all adult cases, by the activating point mutation KIT D816V, an amino-acid substitution in the gene's kinase domain. In a prospective study of 113 adults with systemic mastocytosis, D816V was found in virtually all patients (93%) with indolent or aggressive forms, but in only 29% of the well-differentiated variant, and patients with poor-prognosis disease more often carried the mutation across two or more bone-marrow myeloid lineages (81%) than did those with indolent disease (27%).¹²¹ Midostaurin, a multikinase inhibitor active against KIT D816V, produced an overall response rate of 60% (95% CI 49–70%) in an open-label trial of 116 patients with advanced systemic mastocytosis, with a median overall survival of 28·7 months across the efficacy population and 9·4 months in the subgroup with mast-cell leukaemia.¹²² Avapritinib, designed specifically against KIT D816V, has been tested in two linked trials: the phase 1 EXPLORER study and the phase 2 PATHFINDER study, both reporting safety and efficacy in advanced systemic mastocytosis and both published together in 2021.^123,124

Myeloid/lymphoid neoplasms with eosinophilia and tyrosine kinase gene fusions

A distinct, smaller group of eosinophilic neoplasms is defined not by a point mutation but by a fusion gene involving a tyrosine kinase — most often PDGFRA, and less often PDGFRB, FGFR1, or JAK2. The paradigm case is the cryptic fusion FIP1L1::PDGFRA, found in a substantial fraction of patients previously classified as having idiopathic hypereosinophilic syndrome; the fusion kinase is inhibited by imatinib at low concentration. In the founding series, 9 of 11 patients treated with imatinib had responses lasting more than three months, and the fusion was found in 9 of 16 patients with the syndrome; relapse in one patient came with an acquired T674I resistance mutation in PDGFRA.¹²⁵ This imatinib sensitivity is specific to the PDGFRA- and PDGFRB-rearranged subgroups; the FGFR1- and JAK2-rearranged fusions in this category are not imatinib-sensitive by the same mechanism and are managed differently, Both current classifications (WHO-HAEM5 and ICC, 2022) keep this as a named category because the fusion partner changes management.^1,2

B-lymphoblastic leukaemia

B-lymphoblastic leukaemia (B-ALL) is defined by a small number of founding genetic lesions — an abnormal chromosome number or a chromosomal rearrangement — each of which arrests B-lymphoid development at a different point and, until recently, left a large minority of cases with no identifiable driver at all. Large sequencing cohorts have shown that some of the founding lesions are themselves gateways to a shared downstream biology rather than single diseases, and transcriptome sequencing of thousands of cases has resolved most of the previously unclassified fraction into new subtypes defined by gene expression, point mutation or both.¹²⁶

Aneuploidy: the founding classification

The oldest and largest classification by chromosome number remains prognostically load-bearing. High hyperdiploidy (modal chromosome number ≥51) is the single most frequent lesion in childhood B-ALL and one of the two lesions independently associated with a lower risk of relapse in the UK Medical Research Council ALL97/99 trial of 1725 children (hazard ratio 0·60, 95% CI 0·47–0·78, against the other cytogenetic groups).¹²⁷ At the other end of chromosome number, hypodiploid ALL splits into two biologically distinct groups on integrated genomic analysis of 124 cases. Near-haploid ALL (24–31 chromosomes) is characterised by alterations of receptor tyrosine kinase and Ras signalling in 71% of cases and of the lymphoid transcription factor IKZF3 in 13%. Low-hypodiploid ALL (32–39 chromosomes) is characterised by TP53 alterations in 91·2% of cases, commonly present in the patient's non-tumour cells, that is, germline, together with IKZF2 (53%) and RB1 (41%) alterations.¹²⁸ The germline component of low-hypodiploid TP53 alteration means that a diagnosis of low-hypodiploid ALL in a child is itself an indication for Li-Fraumeni syndrome assessment, a point the genomic-landscape paper makes directly.¹²⁸

The founding fusions

Below hyperdiploidy in frequency but longer established clinically, the classical fusions retain their prognostic direction. ETV6::RUNX1 is the most common single fusion in children and, like hyperdiploidy, an independent marker of lower relapse risk in the same UK trial (hazard ratio 0·51, 95% CI 0·38–0·70).¹²⁷ BCR::ABL1 (the Philadelphia chromosome), KMT2A-rearrangements (commonly with AFF1, MLLT1, MLLT3 or MLLT10) and intrachromosomal amplification of chromosome 21 (iAMP21) sit at the opposite end. In the same analysis, iAMP21 carried the highest relapse hazard of any single lesion studied (hazard ratio 6·04, 95% CI 3·90–9·35), ahead of BCR::ABL1 (3·55, 2·21–5·72) and KMT2A-rearrangement (2·98, 1·71–5·20).¹²⁷ iAMP21 is itself genetically complex: genomic characterisation of 94 cases found secondary abnormalities of IKZF1 (22%), CDKN2A/B (17%), PAX5 (8%), ETV6 (19%) and RB1 (37%), all arising after and dependent on the chromosome 21 rearrangement, and a uniformly poor outcome on standard-risk therapy regardless of which secondary lesion was present.¹²⁹ An international study of 530 iAMP21 cases subsequently showed that treating the group as high risk from diagnosis reproduces the improved outcome seen in single-trial series, whatever the chemotherapy backbone, and argued for iAMP21 to be recognised as its own World Health Organization entity on that basis.¹³⁰ TCF3::PBX1 and, much more rarely, TCF3::HLF complete the classical fusion set. TCF3::HLF-positive ALL carries a dismal prognosis: in the series that sequenced it in detail, the affected children died of disease progression or treatment-related toxicity within two years on average, with only one patient in remission at last follow-up, in contrast with the treatment-responsive TCF3::PBX1 subtype.¹³¹

Subtypes defined by expression and sequencing

The largest remaining question in B-ALL classification — what drives the cases with no recognised cytogenetic lesion — has been answered largely by transcriptome sequencing. Genomic profiling of 1725 B-ALL cases and detailed study of 154 of them identified Philadelphia chromosome-like (Ph-like) ALL: a gene-expression profile resembling BCR::ABL1-positive ALL but without the fusion itself, present in 10% of standard-risk children and rising to 27% of young adults, and carrying a poor outcome.¹⁰ Kinase-activating lesions were found in 91% of Ph-like cases and fell into two therapeutically distinct groups: an ABL-class of rearrangements involving ABL1, ABL2, CSF1R or PDGFRB, sensitive in vitro to dasatinib, and a JAK-STAT/CRLF2 class involving EPOR, JAK2 rearrangement or CRLF2 overexpression together with JAK1/2/3, IL7R or SH2B3 sequence mutations, with the EPOR and JAK2 subset sensitive to ruxolitinib in the same models.¹⁰

Integrated genomic analysis of 1988 childhood and adult B-ALL cases, combining RNA sequencing, copy-number and mutation data, resolved the disease into 23 subtypes and gave each an age-stratified frequency (Table 8).¹³² Two of the new subtypes centre on PAX5: PAX5alt (7·4% of the cohort), comprising diverse rearrangements, intragenic amplifications and point mutations of PAX5, and a separate, more homogeneous subtype first described as its own entity by a French adult-ALL group and defined by the single missense mutation PAX5 p.Pro80Arg (P80R),¹³³ present in all 44 cases assigned to it in the 1988-case cohort and in only 4 of the remaining 1944. A third subtype is defined by a single missense mutation in a different transcription factor, IKZF1 p.Asn159Tyr (N159Y), in 8 cases; unlike PAX5 P80R, the non-mutated IKZF1 allele remained expressed. The same study reported that PAX5 P80R and PAX5alt cases had intermediate-to-poor outcomes in children and adults treated on Children's Oncology Group protocols, consistent with the more general finding that stem-cell-like B-ALL — KMT2A-rearranged and IKZF1-mutated Ph-positive or Ph-like disease — carries an inferior outcome.¹³² Minimal residual disease at the end of induction, in children on the AALL0232 protocol, was positive in 7·2% of PAX5 P80R cases against 29·4% of PAX5alt cases, underlining that the two PAX5-centred subtypes are not equivalent despite the shared gene.¹³²

DUX4-rearranged ALL, in which IGH::DUX4 or ERG::DUX4 fusions drive DUX4 overexpression and frequently co-occur with intragenic ERG deletion, was first delineated by RNA sequencing of 195 paediatric cases, representing 4% of that cohort; the same analysis identified a second new subtype with an ETV6::RUNX1-like expression profile but coexisting ETV6 and IKZF1 alterations rather than the fusion itself.¹³⁴ DUX4-rearranged ALL is associated with a favourable outcome, irrespective of the presence of concomitant lesions that predict poorly in other B-ALL subtypes, such as IKZF1 deletion — a paradox the discovery cohort attributes to DUX4/ERG deregulation itself rather than to the co-occurring lesions.¹³⁵ MEF2D-rearranged ALL (most often MEF2D::BCL9 or MEF2D::HNRNPUL1) made up 2·4% of a single Japanese trial cohort, presented with an aberrant pre-B immunophenotype and a high frequency of PHF6 mutation (50%), and carried a significantly worse outcome, with 53·3% of patients relapsing and subsequently dying; relapse was frequently associated with acquired CDKN2A/CDKN2B deletion.¹³⁶ ZNF384-rearranged ALL, most often as TCF3::ZNF384, made up 4·1% of a second cohort from the same trial group (2·4% for TCF3::ZNF384 specifically), shared a characteristic weak-CD10, aberrant CD13/CD33 immunophenotype across fusion partners, but differed in outcome by partner: TCF3::ZNF384 cases had a poorer steroid response and higher relapse rate than the alternative EP300::ZNF384.¹³⁷

IKZF1 deletion, whether or not it defines a named subtype, is independently prognostic in its own right: in a cohort of 221 children with high-risk B-ALL (excluding BCR::ABL1-positive, hypodiploid and infant ALL), copy-number alterations of regulators of B-cell development were found in 66·8% of patients, most often affecting PAX5 (31·7%) and IKZF1 (28·6%); a poor-outcome predictor built from these alterations was validated in an independent 258-patient cohort and was strongly associated with IKZF1 alteration specifically, its gene-expression signature resembling that of BCR::ABL1-positive ALL.¹³⁸ The composite marker "IKZF1plus" — IKZF1 deletion together with deletion of CDKN2A, CDKN2B or PAX5 — was associated with lower disease-free survival even in an otherwise favourable, chemotherapy-free treatment context for adult Ph-positive ALL, discussed below.¹³⁹

Germline predisposition and age distribution

Common, low-penetrance germline variants also shape risk. A genome-wide association study of 907 childhood ALL cases and 2398 controls identified risk loci at 7p12.2 (IKZF1, odds ratio 1·69), 10q21.2 (ARID5B, odds ratio 1·65) and 14q11.2 (CEBPE, odds ratio 1·34), with the ARID5B association selective for hyperdiploid B-ALL.¹⁴⁰ A parallel study of 317 cases and 17,958 controls confirmed two ARID5B variants that distinguished hyperdiploid ALL from other subtypes in an independent 124-child validation cohort.¹⁴¹ A separate genome-wide study specific to Ph-like ALL, in 511 cases and 6661 controls, found a susceptibility locus at GATA3 (rs3824662; odds ratio 3·85 for Ph-like ALL against non-ALL, 3·25 against non-Ph-like ALL); the risk allele was associated with the somatic lesions that define Ph-like disease (CRLF2 rearrangement, JAK mutation, IKZF1 deletion) and with early treatment response.¹⁴² The same group later identified a susceptibility locus at the ERG locus specific to Hispanic children, whose effect size rose with increasing Native American genetic ancestry, in a genome-wide study of 940 Hispanic ALL cases against 681 ancestry-matched controls — one contributor, among others not addressed here, to the higher incidence and inferior survival of ALL in Hispanic children.¹⁴³

Rare, high-penetrance germline variants define a smaller predisposed group. Germline ETV6 mutation causes a syndrome of thrombocytopenia and predisposition to B-ALL.¹⁴⁴ A recurrent germline PAX5 mutation (G183S) segregates with B-ALL susceptibility in affected families.¹⁴⁵ Germline coding variation in IKZF1 itself, distinct from the somatic deletions discussed above, was found in familial childhood ALL and in 0·9% of presumed sporadic B-ALL (28 unique variants across 45 children), most of which impaired IKZF1 function and drug responsiveness.¹⁴⁶ Germline TP53 variants underlie much of the low-hypodiploid subtype described above and warrant Li-Fraumeni assessment when that subtype is diagnosed.¹²⁸ Age is the single strongest driver of subtype distribution across the whole disease (Table 8): ETV6::RUNX1 and high hyperdiploidy dominate standard-risk childhood ALL and become rare by adulthood, while KMT2A-rearrangement, BCR::ABL1 and low-hypodiploidy are all several-fold more frequent in adults than in standard-risk children, and Ph-like ALL is more common in adolescents and young adults than in any other age group.^10,132

Genotype-linked therapy

Ph-positive ALL is the clearest case of genotype dictating treatment. In a phase 3 trial of 245 adults with newly diagnosed Ph-positive ALL randomised 2:1 to frontline ponatinib or imatinib, each with reduced-intensity chemotherapy, the primary endpoint — minimal residual disease-negative complete remission at the end of cycle 3 — was reached by 34·4% (53/154) on ponatinib against 16·7% (13/78) on imatinib (risk difference 0·18, 95% CI 0·06–0·29; p=0·002); event-free survival had not been reached with ponatinib at the data cutoff, against a median of 29 months with imatinib.¹⁴⁷ A separate, chemotherapy-free strategy — the Italian GIMEMA LAL2116 (D-ALBA) trial — combined dasatinib with glucocorticoids for induction, then blinatumomab, in 63 adults with newly diagnosed Ph-positive ALL (median age 54 years, range 24–82): complete remission was reached in 98%, a molecular response rose from 29% after dasatinib induction to 60% after two blinatumomab cycles, and at a median follow-up of 18 months overall survival was 95% and disease-free survival 88% — lower in patients with IKZF1plus.¹³⁹ These come from different evidence bases — a randomised phase 3 trial against a single-group phase 2 study — with different populations, age ranges and primary endpoints, and no randomised trial has directly compared the two strategies.

Menin inhibitors act on KMT2A-rearranged ALL as well as AML. The first revumenib trial enrolled both lineages, and the US approval of revumenib covers relapsed or refractory KMT2A-translocated acute leukaemia in adults and children aged one year and older; there is as yet no ALL-specific response rate (see Genotype-directed therapy under Acute myeloid leukaemia).^62,148

In Ph-like ALL, the kinase lesions point to drugs already in use: dasatinib for the ABL-class fusions (ABL1, ABL2, CSF1R, PDGFRB) and ruxolitinib for the EPOR and JAK2 rearrangements. The evidence is from cell lines and xenografts.¹⁰ In patients, the evidence is from case reports and small non-randomised series rather than a completed randomised trial, so a benefit from adding a kinase inhibitor in Ph-like ALL is not proven.¹⁴⁹

T-lymphoblastic leukaemia

T-lymphoblastic leukaemia (T-ALL) is organised less by chromosome number than by which T-cell transcription factor is aberrantly activated and at what stage of thymocyte development the leukaemic clone is arrested. An early gene-expression study of paediatric T-ALL identified five oncogenes — HOX11 (TLX1), TAL1, LYL1, LMO1 and LMO2 — recurrently activated in the absence of the chromosomal translocations that had originally implicated them, each marking arrest at a specific thymocyte developmental stage (a pro-T LYL1+ signature, an early cortical HOX11+ signature and a late cortical TAL1+ signature); a related subgroup activates HOX11L2 (TLX3). Clinically, HOX11/TLX1 activation was associated with a favourable prognosis, while TAL1, LYL1 and, unexpectedly, HOX11L2/TLX3 expression conferred a markedly worse response to treatment.¹⁵⁰ Integrated genomic analysis of 264 childhood and young-adult cases extended this framework, identifying 106 putative T-ALL driver genes — about half not previously described in childhood T-ALL — and showing that particular pathway mutations cluster with particular subtypes: NRAS/FLT3 mutation with immature T-ALL, JAK3/STAT5B mutation with HOXA-deregulated ALL, PTPN2 mutation with TLX1-deregulated T-ALL, and PIK3R1/PTEN mutation with TAL1-deregulated T-ALL.¹⁵¹

Underlying most of these subgroups is activation of NOTCH1 signalling. More than half of human T-ALL — across all major molecular subtypes — carries activating mutations of the extracellular heterodimerisation domain or the C-terminal PEST domain of NOTCH1, a finding that reframed T-ALL as, in large part, a disease of dysregulated NOTCH signalling and provided the rationale for NOTCH-directed therapy.¹⁵² FBXW7, which normally targets NOTCH1 for degradation, is mutated less often, with the same net effect of prolonging NOTCH1 activity. In a cohort of 98 children treated on the MRD risk-directed CCLG-ALL 2008 protocol, 51·02% of patients carried a NOTCH1 mutation and 8·75% an FBXW7 mutation, and NOTCH1-mutant patients had a more favourable 10-year overall survival (82·7% vs 62·4%, p=0·020), event-free survival (80·9% vs 48·4%, p=0·001) and disease-free survival (82·9% vs 52·9%, p=0·001) than NOTCH1-wild-type patients. NOTCH1 status was an independent prognostic factor alongside post-induction minimal residual disease; the authors describe the wider literature on NOTCH1/FBXW7 prognostic significance in T-ALL as contested.¹⁵³ A larger, adult cohort points the same way as this paediatric one: in 212 patients with T-ALL in the multicentre GRAALL-2003/2005 trials, a NOTCH1 and/or FBXW7 mutation was found in 67%, and an oncogenetic classifier defined low risk as NOTCH1/FBXW7 mutation without a RAS pathway or PTEN abnormality (51% of evaluable patients). On multivariable analysis it was the only significant prognostic factor: patients outside the low-risk group had a hazard ratio of 3·2 for events (95% CI 1·9–5·15) and 3·2 for death (1·9–5·6).¹⁵⁴ Loss of the CDKN2A/CDKN2B (INK4/ARF) tumour-suppressor locus is, together with NOTCH1 activation, described as a hallmark of T-ALL, and in the same 264-case genomic landscape cohort a CDKN2A/CDKN2B alteration was found in 78·4% of patients — the single most frequent lesion in the series.¹⁵¹

A biologically distinct minority does not fit the thymocyte-arrest framework at all. Early T-cell precursor ALL (ETP-ALL), defined immunophenotypically by an early, myeloid-adjacent stage of arrest, carries a mutation profile that resembles myeloid rather than lymphoid leukaemia. Whole-genome sequencing of 12 ETP-ALL cases, with mutation frequencies assessed in a further 94 T-ALL cases, found activating mutations of cytokine receptor and Ras signalling genes, lesions disrupting haematopoietic development (58%; GATA3, ETV6, RUNX1, IKZF1, EP300) and histone-modifying gene mutations (48%; EZH2, EED, SUZ12, SETD2, EP300). The global transcriptional profile resembled normal and leukaemic myeloid stem cells, which the authors read as an argument for trials of myeloid-directed therapy in this poor-prognosis subgroup.¹⁵⁵

Chronic lymphocytic leukaemia

Chronic lymphocytic leukaemia (CLL) is staged clinically but risk-stratified genetically, and its genetics divide into three layers that answer three different questions: the cell of origin (IGHV mutational status), the chromosomal lesion present at diagnosis (fluorescence in situ hybridisation, or FISH), and the mutation acquired under the selective pressure of a specific targeted drug (resistance genetics). All three are now part of routine management.

IGHV mutational status and stereotyped receptors

Two independent studies published together in 1999 established that CLL cells with unmutated immunoglobulin heavy-chain variable (IGHV) genes mark a more aggressive disease than those with somatically mutated IGHV genes — the first evidence that CLL's cell of origin, naive versus post-germinal-centre B cell, carries prognostic information independent of clinical stage. In a series of 84 patients, 45·2% had unmutated IGHV genes (defined as ≥98% sequence homology to the nearest germline gene), and median survival for stage A patients with unmutated genes was 95 months against 293 months for those with mutated genes (p=0·0008).¹⁵⁶ A parallel cohort linked unmutated IGHV status to high CD38 expression (≥30% of cells) and showed that both markers predicted poor response to chemotherapy, including fludarabine, and shorter survival.¹⁵⁷ The IGHV repertoire in CLL is also non-random: in a series of 916 patients, 21·9% expressed one of 48 subsets of near-identical ("stereotyped") heavy-chain complementarity-determining region 3 sequences, and particular stereotyped subsets carried their own clinical behaviour independent of overall IGHV mutation status — the IGHV3-21/IGLV3-21 subset following an aggressive course regardless of mutation status, and the IGHV4-34/IGKV2-30 subset a strikingly indolent one.¹⁵⁸

The FISH hierarchy and TP53

FISH for four abnormalities, applied to 325 patients, established the hierarchy that still anchors genetic risk stratification in CLL: del(17p) carried the shortest median survival (32 months) and the shortest treatment-free interval (9 months); del(11q) an intermediate survival (79 months); trisomy 12 (114 months) and a normal karyotype (111 months) similar to each other; and del(13q) as the sole abnormality the longest survival (133 months).¹⁵⁹ del(17p) often, but not always, comes with TP53 mutation. The 2018 update of the International Workshop on CLL (iwCLL) guideline revised its recommendations for diagnosis, treatment indications and response assessment in light of the genomic advances of the previous decade. On testing, the guideline states: "Examination of leukemia-cell cytogenetics (eg, metaphase karyotyping, FISH [in particular for del(17p)]) and analysis for inactivating mutations in TP53 should be performed before any line of therapy," with mutations reported at a threshold of about 10% variant allele frequency, and it recommends repeating del(17p)/TP53 testing before any subsequent line of treatment because additional genetic abnormalities may be acquired as the disease progresses.¹⁶⁰ The 2026 European Hematology Association guideline keeps the same rule. IGHV status is tested once, usually before first-line therapy, because it does not change over the disease course. del(17p) by FISH and TP53 mutation by sequencing are tested before every line of treatment; next-generation sequencing is preferred because it detects small subclones that a DNA-damaging therapy may later select.¹⁶¹

The recurrent mutation landscape and complex karyotype

Two large next-generation sequencing studies, published together in 2015, extended the genetic picture beyond FISH. Whole-exome sequencing of 538 CLL cases and matched germline DNA (278 from a prospective clinical trial) identified 44 recurrently mutated genes and 11 recurrent copy-number changes, including two previously unrecognised putative drivers, RPS15 and IKZF3, and reconstructed the clonal evolution between paired pre-treatment and relapse samples from 59 patients, demonstrating frequent clonal shifts under treatment pressure.¹⁶² A companion study, combining whole-genome and exome sequencing of 452 CLL cases and 54 cases of monoclonal B-lymphocytosis (the precursor state discussed below), extended the driver list further (ZNF292, ZMYM3, ARID1A, PTPN11) and identified recurrent non-coding mutations — in the 3' region of NOTCH1, causing aberrant splicing and increased NOTCH1 activity, and in a chromosome 9p13 enhancer, reducing expression of PAX5 — with the cumulative number of driver alterations (0 to ≥4 of the alterations tracked) discriminating between clinical courses.¹⁶³ Together these studies place NOTCH1, SF3B1, ATM, BIRC3, MYD88 and XPO1 among the recurrently mutated genes in CLL. Of the two novel drivers the whole-exome series itself identified, RPS15 was mutated in 23 of 538 cases (4·3%) and IKZF3 in 11 of 538 (2·0%), both clustering at highly conserved residues.¹⁶²

Structural complexity carries independent prognostic weight. In a retrospective study of 5290 patients with chromosome-banding analysis, patients with a complex karyotype (≥3 chromosomal aberrations) and specifically ≥5 aberrations ("high-CK") had uniformly poor outcomes independent of clinical stage, TP53 aberration and IGHV mutational status, whereas complex karyotypes of 3–4 aberrations behaved aggressively only in the presence of a TP53 aberration, and complex karyotypes accompanied by trisomy 12 and trisomy 19 together were, by contrast, exceptionally indolent — evidence that complex karyotype should not be treated as uniformly unfavourable in CLL.¹⁶⁴

CLL-IPI and Richter transformation

The CLL International Prognostic Index (CLL-IPI) combined the layers above into a single validated score. A meta-analysis of individual patient data from 3472 treatment-naive patients in eight phase 3 trials identified five independent prognostic factors on multivariate analysis — TP53 status (deletion and/or mutation versus neither), IGHV mutational status, serum β2-microglobulin (≤3·5 mg/L versus >3·5 mg/L), clinical stage (Binet A/Rai 0 versus higher), and age (≤65 versus >65 years) — and derived four risk groups with 5-year overall survival of 93·2% (low), 79·3% (intermediate), 63·3% (high) and 23·3% (very high), confirmed in internal and two external validation cohorts.¹⁶⁵ Richter transformation — progression of CLL to an aggressive, usually diffuse large B-cell, lymphoma — was characterised at the genomic level by whole-genome sequencing of paired circulating-CLL and transformed-tissue biopsies from 17 patients: the transformed tissue was enriched for mutations in poor-risk CLL drivers and in DNA-damage-response pathway genes, and carried genomic aberrations not previously implicated in Richter transformation (PTPRD, TRAF3) alongside mutational signatures pointing to DNA-damage response as the dominant mechanism of transformation.¹⁶⁶

Resistance mutations to targeted therapy

Acquired resistance to each of the three targeted drug classes used in CLL has its own genetics. Whole-exome sequencing at baseline and relapse in six patients with acquired resistance to the covalent BTK inhibitor ibrutinib identified a cysteine-to-serine mutation at the ibrutinib binding site of BTK (C481S) in five, converting irreversible to reversible inhibition, and gain-of-function mutations in the downstream signalling molecule PLCγ2 (R665W, L845F) in two; none of these mutations was found in patients with prolonged lymphocytosis but no disease progression on ibrutinib.¹⁶⁷ The noncovalent BTK inhibitor pirtobrutinib, designed to bypass C481S, has its own resistance genetics: among 55 treated patients, 9 developed relapsed or refractory disease with acquired resistance, and genomic analysis identified BTK kinase-domain mutations (V416L, A428D, M437R, T474I, L528W) and PLCγ2 mutations that conferred resistance to noncovalent inhibitors and, for a proportion of the mutations, to covalent inhibitors as well.¹⁶⁸ For the BCL2 inhibitor venetoclax, paired pre-treatment and progression samples from 15 patients on venetoclax trials identified a recurrent BCL2 Gly101Val mutation at progression in 7, absent at study entry, detectable 19–42 months before clinical progression, and reducing BCL2's binding affinity for venetoclax roughly 180-fold — a mutation that both explains a mechanism of resistance and offers a biomarker of impending relapse.¹⁶⁹

Monoclonal B-cell lymphocytosis

Monoclonal B-cell lymphocytosis (MBL) — a clonal CLL-phenotype B-cell population below the diagnostic threshold for CLL of 5 × 10⁹ clonal B cells per litre (5000 per mm³) — is CLL's precursor state. In a study of 1520 people aged 62–80 years with a normal blood count and 2228 with lymphocytosis, MBL was found by flow cytometry in 5·1% and 13·9% respectively, and CLL-phenotype MBL shared del(13q) and trisomy 12 frequencies, and a skewed IGHV repertoire, with overt CLL. Among 185 subjects with lymphocytosis followed for a median of 6·7 years, progressive lymphocytosis developed in 28%, progressive CLL in 15%, and 7% required chemotherapy; overall, CLL requiring treatment developed at a rate of 1·1% per year among subjects with CLL-phenotype MBL and lymphocytosis.¹⁷⁰ The same population-based framing underlies the CLL genomic-landscape sequencing study above, which included 54 MBL cases alongside 452 CLL cases specifically to study this precursor relationship at the mutational level.¹⁶³

Diffuse large B-cell lymphoma

Diffuse large B-cell lymphoma (DLBCL) is the most common lymphoma in adults, and it is not one disease. The first division came from gene-expression profiling, which split DLBCL into a germinal-centre B-cell-like (GCB) group and an activated B-cell-like (ABC) group, with a smaller unclassified remainder.¹⁷¹ The two groups arise from different stages of B-cell differentiation, carry different mutations, and respond differently to chemotherapy, with ABC tumours faring worse than GCB tumours on standard immunochemotherapy.¹⁷¹ Cell of origin describes a transcriptional state, not a specific lesion, and it does not by itself explain why some ABC tumours respond to a targeted drug and others do not.

The genetic subtypes: MCD, BN2, N1, EZB, and LymphGen

In 2018, two groups published overlapping but not identical genetic classifications of DLBCL from large sequencing cohorts. Schmitz and colleagues sequenced 574 biopsies and, by looking for genetic lesions that co-occur more often than chance, defined four subtypes: MCD (co-occurring MYD88^L265P and CD79B mutations), BN2 (BCL6 fusions with NOTCH2 mutations), N1 (NOTCH1 mutations), and EZB (EZH2 mutations with BCL2 translocations).¹⁷² BN2 and EZB had favourable survival on immunochemotherapy; MCD and N1 had inferior outcomes.¹⁷² The same year, Chapuy and colleagues analysed 304 primary DLBCLs by consensus clustering of a wider set of genetic drivers and identified five clusters plus a small group with no detectable drivers: a low-risk cluster of ABC tumours of extrafollicular or marginal-zone origin, two GCB clusters with different outcomes, an ABC cluster with MYD88 L265P and CD79B mutations, and a cluster defined by biallelic TP53 inactivation and CDKN2A loss with genomic instability, independent of cell of origin.¹⁷³ The two schemes overlap — both recognise an MYD88/CD79B-driven ABC subgroup and an EZH2/BCL2-driven GCB subgroup — but they were built independently and do not map onto each other one for one.

Wright and colleagues resolved the overlap with LymphGen, an algorithm trained on both cohorts that assigns each tumour a probability of belonging to one of seven genetic subtypes, keeping the original four (MCD, BN2, N1, EZB) and adding A53 (biallelic TP53 inactivation with aneuploidy) and ST2 (SGK1 and TET2 mutations). EZB is split in two by the double-hit gene-expression signature, into EZB-MYC+ and EZB-MYC−, which makes seven. Tumours that fit none are left unclassified.¹⁷⁴ Across the cohorts, MCD had inferior survival, BN2 was favourable among ABC tumours, and EZB fared worse than ST2 among GCB tumours.¹⁷⁴ LymphGen subtypes share genetic features with several indolent and extranodal lymphomas — BN2 with marginal-zone lymphoma, for instance — which the authors read as evidence of shared pathogenesis across lymphoma types rather than DLBCL being genetically self-contained.¹⁷⁴

MYD88, CD79B and chronic active B-cell receptor signalling

The MCD subtype's biology was worked out before its name was. Chronic, low-level (as opposed to antigen-triggered, transient) B-cell receptor signalling was shown to be required for survival of a subset of DLBCL cell lines, distinguishing them from GCB-derived lines.³⁴ The mutations that sustain this signal were identified as activating mutations of MYD88, most often the L265P substitution, found in around 29% of ABC-DLBCLs and rare in GCB-DLBCL in the discovery cohort; MYD88 L265P triggers IRAK-mediated NF-κB signalling and JAK–STAT3 signalling.³⁷ CD79B mutations, which impair a negative feedback loop on B-cell receptor signalling, frequently co-occur with MYD88 mutations in the same tumours — the basis of the MCD designation — and both lesions recur at extranodal sites, including the testis and the central nervous system, where MCD and its associated biology are disproportionately represented.^37,172

EZH2, KMT2D, CREBBP and the EZB/BCL2 axis

EZH2 encodes the catalytic subunit of PRC2, a histone methyltransferase; a recurrent Tyr641 substitution, first described in follicular lymphoma and GCB-DLBCL together, is a gain-of-function mutation that alters its substrate specificity.²⁵ Large sequencing surveys of non-Hodgkin lymphoma broadened the epigenetic picture, finding that KMT2D (also called MLL2) is among the most frequently mutated genes across follicular lymphoma and DLBCL.²⁷ CREBBP and its paralogue EP300 — both histone/transcription-factor acetyltransferases — are inactivated, predominantly by missense mutation of their acetyltransferase domain, in a recurrent subset of B-cell lymphomas, and their loss impairs two acetylation steps: the one that inactivates the BCL6 oncoprotein and the one that activates p53.¹⁷⁵ In the EZB/LymphGen subtype, these epigenetic lesions co-occur with BCL2 translocations — the t(14;18) also seen in follicular lymphoma — placing EZB DLBCL on a shared pathogenic axis with follicular lymphoma, from which a substantial proportion of EZB tumours are thought to arise by transformation or share a common progenitor clone.^172,173

A subset of GCB-DLBCL, largely overlapping with EZB, carries a "dark-zone" gene-expression signature associated with a high-grade phenotype. Ennishi and colleagues derived a 104-gene double-hit signature (DHITsig) from 157 GCB-DLBCLs; it assigned 27% of GCB-DLBCLs to a DHITsig-positive group, only half of which actually carried concurrent MYC and BCL2 rearrangements, and DHITsig-positive tumours had a lower 5-year time-to-progression rate after R-CHOP than DHITsig-negative tumours (57% vs 81%).¹⁷⁶ The signature therefore captures a biological state — a dark-zone-like, non-light-zone germinal-centre cell of origin with a distinct mutational landscape — that is broader than, and only partly explained by, the presence of a MYC/BCL2 double hit itself.¹⁷⁶

High-grade B-cell lymphoma with MYC and BCL2 rearrangements: WHO-HAEM5 and the ICC

The two current classifications diverge on how to name a large B-cell lymphoma that carries both a MYC rearrangement and a BCL2 and/or BCL6 rearrangement. WHO-HAEM5 (2022) names the entity "diffuse large B-cell lymphoma/high-grade B-cell lymphoma with MYC and BCL2 rearrangements", defined by MYC plus BCL2 whatever the cell size. A MYC/BCL6 double hit without BCL2 is excluded, because these tumours are more diverse, and is classed as DLBCL, not otherwise specified, or high-grade B-cell lymphoma, not otherwise specified, by morphology.¹⁷⁷ The ICC (2022) keeps high-grade B-cell lymphoma with MYC and BCL2 rearrangements and adds a separate provisional entity, high-grade B-cell lymphoma with MYC and BCL6 rearrangements.¹⁷⁸ So a MYC/BCL6 double hit is a named entity in one system and not in the other. In neither is a MYC rearrangement alone enough for these categories. Where a laboratory report and a treating team appear to disagree on nomenclature for the same biopsy, checking which classification and which edition each used usually resolves it.

Genotype and therapy: ibrutinib and polatuzumab vedotin

Genetic subtype has not yet changed standard treatment. The clearest signal comes from PHOENIX, a phase 3 trial that added the BTK inhibitor ibrutinib to R-CHOP in non-GCB DLBCL and showed a survival benefit only in younger patients. A genetic reanalysis of trial biopsies found MCD, BN2 and N1 tumours. In patients aged 60 or younger, 3-year event-free survival with ibrutinib plus R-CHOP was 100% in both MCD and N1, against 42·9% and 50% with R-CHOP alone.¹⁷⁹ This is a retrospective, biopsy-driven subgroup analysis of a trial that was not designed around genetic subtype, and it has not yet been prospectively confirmed; it is evidence that genetic subtype explains a drug effect the overall trial result had obscured, not yet evidence for genotype-directed prescribing outside a trial.¹⁷⁹

Polatuzumab vedotin, an antibody–drug conjugate against CD79b, is not genotype-directed. The POLARIX trial randomised 879 previously untreated patients with intermediate- or high-risk DLBCL to polatuzumab vedotin plus R-CHP (replacing vincristine) or standard R-CHOP, unselected by genetic subtype; 2-year progression-free survival was 76·7% with pola-R-CHP against 70·2% with R-CHOP (hazard ratio 0·73, 95% CI 0·57–0·95), with no significant difference in overall survival at 2 years.¹⁸⁰ CD79b is broadly expressed across DLBCL regardless of subtype, and the trial's benefit was not restricted to MCD or any other CD79B-mutant subgroup; the drug and the genetics of MCD share a target gene but not a rationale.¹⁸⁰

Follicular lymphoma

Follicular lymphoma's defining lesion, t(14;18)(q32;q21), places BCL2 under the control of the IGH enhancer, producing constitutive BCL2::IGH expression and a survival advantage rather than a proliferative one. The translocation alone is not sufficient for lymphoma: it is detectable by sensitive PCR in the blood of a substantial proportion of healthy adults with no lymphoma and no clinical consequence, at low copy number and without progressing in most who carry it.¹⁸¹ A second hit — most often an epigenetic one — is needed to drive overt disease.

Follicular lymphoma is among the most epigenetically deranged of the B-cell lymphomas. The three lesions described under diffuse large B-cell lymphoma recur here: KMT2D mutation, in 89% of cases in the founding cohort,²⁷ deletion or mutation of CREBBP or EP300, in 41%,¹⁷⁵ and EZH2 Tyr641 mutation.²⁵ These lesions are thought to act early and are often shared with the clone that later transforms into DLBCL, so follicular lymphoma is a disease of altered chromatin as much as of its founding translocation.

Pastore and colleagues combined seven of these mutations — EZH2, ARID1A, MEF2B, EP300, FOXO1, CREBBP and CARD11 — with the clinical FLIPI score and performance status into the m7-FLIPI model. In the training cohort (151 patients on R-CHOP), the resulting high-risk group (28% of patients) had a 5-year failure-free survival of 38·29% against 77·21% for the low-risk group (hazard ratio 4·14, 95% CI 2·47–6·93), and m7-FLIPI outperformed a model built from the mutations alone; the model was validated in an independent, population-based cohort treated with R-CVP.¹⁸² m7-FLIPI is a prognostic, not yet a treatment-selecting, tool: it identifies who is more likely to fail first-line immunochemotherapy, not which alternative regimen to give instead.

Because EZH2-mutant follicular lymphoma depends on the mutant enzyme's altered activity, it is directly targetable. Tazemetostat, an oral EZH2 inhibitor, was tested in a single-arm phase 2 trial of 99 patients with relapsed or refractory follicular lymphoma stratified by EZH2 mutation status; the objective response rate was 69% in the EZH2-mutant cohort against 35% in the wild-type cohort, with median progression-free survival of 13·8 months and 11·1 months respectively.¹⁸³ The US approval of June 2020 reflects both results: it covers EZH2-mutant follicular lymphoma after at least two previous treatments, and follicular lymphoma of any EZH2 status when there is no other option.¹⁸⁴ The genotype raises the response rate; it is not required for the drug to work.

Mantle cell lymphoma

Mantle cell lymphoma is founded on t(11;14)(q13;q32), which places CCND1 (cyclin D1) under the IGH enhancer and drives its constitutive overexpression — present in essentially all cases by definition, since it is part of how the entity is diagnosed. A small minority of morphologically and immunophenotypically identical cases lack CCND1 rearrangement and cyclin D1 expression altogether; these cyclin D1-negative cases instead carry CCND2 rearrangements in more than half of cases, most often with an immunoglobulin light-chain gene as partner, and are recognised by the transcription factor SOX11, which is expressed in both cyclin D1-positive and cyclin D1-negative mantle cell lymphoma and functions as the more reliable diagnostic marker of the two.¹⁸⁵

Beyond the founding translocation, whole-genome and whole-exome sequencing of mantle cell lymphoma has repeatedly converged on ATM, TP53 and the NOTCH pathway as the recurrently mutated genes that drive its clinical heterogeneity, alongside chromatin modifiers such as WHSC1 (NSD2) and MEF2B.¹⁸⁶ TP53 mutation is the single strongest adverse prognostic lesion identified. In 183 younger patients treated on the Nordic MCL2 and MCL3 trials — regimens that represent an intensive, transplant-based standard of care — TP53 mutations (11% of patients) and deletions (16%) were associated with markedly inferior outcomes: median overall survival of 1·8 years in TP53-mutated disease against 12·7 years in unmutated disease, with poor response to both induction chemotherapy and high-dose consolidation.¹⁸⁷ The authors concluded that patients with TP53-mutated mantle cell lymphoma derive little benefit from this intensive approach and should be directed to trials of novel agents rather than to more chemotherapy.¹⁸⁷

NOTCH1 mutations, clustering in the PEST domain that normally targets the NOTCH1 protein for degradation, were found in 12% of clinical mantle cell lymphoma samples and are associated with aggressive disease and poor outcome; NOTCH2 mutations occur as a largely alternative, mutually exclusive event in a further subset of aggressive cases.^186,188 At the other clinical extreme, a leukaemic, non-nodal presentation — circulating disease with little or no lymphadenopathy, hypermutated immunoglobulin genes, a low proliferation signature, and a simpler karyotype — behaves indolently and can often be observed rather than treated at diagnosis; this indolent form shares mantle cell lymphoma's core gene-expression profile but under-expresses a signature dominated by SOX11, distinguishing it at the molecular level from the conventional, nodal, SOX11-high disease with which it would otherwise be indistinguishable by cyclin D1 status alone.¹⁸⁹

BTK inhibitors are used in relapsed or refractory mantle cell lymphoma without genetic selection; no validated genetic marker yet identifies who will respond.

Burkitt lymphoma

Burkitt lymphoma is defined by a MYC translocation — most often t(8;14)(q24;q32) placing MYC::IGH under the immunoglobulin heavy-chain enhancer, with the variant t(2;8) (IGK::MYC) and t(8;22) (MYC::IGL) accounting for the remaining minority of cases — but MYC dysregulation alone is not sufficient to cause the disease; the same translocation, at low frequency, can be found in reactive tissue. Two independent whole-exome sequencing studies published in 2012 identified the cooperating lesions. Schmitz and colleagues, using RNA sequencing and RNA-interference screening across sporadic, endemic and HIV-associated Burkitt lymphoma, found that 70% of sporadic cases carried mutations in TCF3 (E2A) or its negative regulator ID3 that create a dependency on TCF3, which in turn drives tonic B-cell receptor signalling and activates the PI3K pathway; a further 38% of sporadic cases carried oncogenic CCND3 mutations producing a stabilised cyclin D3 protein that drives cell-cycle progression.¹⁹⁰ Love and colleagues, sequencing 59 tumours and comparing them against 94 DLBCL exomes, independently identified ID3 mutations in 34% of Burkitt lymphomas and in no DLBCLs, alongside recurrent mutations in GNA13, RET, PIK3R1 and the SWI/SNF genes ARID1A and SMARCA4.¹⁹¹ ID3/TCF3 and CCND3 mutations are now understood as the second and third genetic hits, alongside MYC, that define the disease's cooperating oncogenic pathways.

Endemic (largely sub-Saharan African, EBV-positive) and sporadic (EBV-positive in up to 30% of cases outside endemic regions¹⁹²) Burkitt lymphoma are not genetically identical. A genome-wide comparison of paediatric endemic and sporadic cases found a striking, genome-wide increase in aberrant somatic hypermutation in EBV-positive tumours, consistent with EBV activating the AICDA (AID) mutator enzyme, and EBV-positive tumours carried significantly fewer driver mutations overall — particularly in apoptosis genes — than EBV-negative tumours, alongside novel candidate genes including SIN3A, USP7 and CHD8.¹⁹² The authors read this as evidence that EBV substitutes, in part, for some of the genetic hits that an EBV-negative tumour must otherwise acquire on its own, and it is consistent with a distinct pathogenic route to the same morphological and clinical entity depending on EBV status and geography, rather than EBV positivity being an incidental passenger.¹⁹²

Marginal zone and lymphoplasmacytic lymphoma

Waldenström macroglobulinaemia, the lymphoplasmacytic lymphoma associated with an IgM paraprotein, is driven in the great majority of cases by the same MYD88^L265P mutation described above in DLBCL. Whole-genome sequencing of bone-marrow lymphoplasmacytic cells found the mutation in all ten paired tumour/normal samples and in 49 of 54 unpaired Waldenström samples (91% of all patients with lymphoplasmacytic lymphoma tested), while it was essentially absent from healthy donors and rare in multiple myeloma, marginal-zone lymphoma or IgM monoclonal gammopathy of undetermined significance — making it a useful differentiator of Waldenström macroglobulinaemia from those look-alike disorders rather than a marker specific to it alone.³⁸ In the same study, two of the three patients with wild-type MYD88 had MLL2 (KMT2D) variants.³⁸

A second recurrent lesion, CXCR4 mutation (typically nonsense or frameshift, producing a WHIM-syndrome-like truncated receptor), co-occurs with MYD88^L265P in a subset of Waldenström macroglobulinaemia and was found in a large single-centre cohort to be a determinant of clinical presentation — associated with higher disease burden and a distinct symptom profile — and of overall survival, independent of MYD88 status.¹⁹³ CXCR4 status also shapes the response to BTK inhibitors, which otherwise depends on MYD88 L265P-driven signalling for their activity. In a prospective trial of ibrutinib in 63 previously treated patients with Waldenström macroglobulinaemia, the overall and major response rates were highest in MYD88^L265P/CXCR4^WT patients (100% and 91·2%), lower in MYD88^L265P/CXCR4^WHIM patients (85·7% and 61·9%), and lowest in patients wild-type for both genes (71·4% and 28·6%).¹⁹⁴

Splenic marginal zone lymphoma has a different genetic signature from Waldenström macroglobulinaemia. Whole-exome sequencing identified NOTCH2 mutations — predicted, in every case, to truncate the protein's C-terminal PEST domain and so impair its proteasomal degradation — in around 20% of cases, restricted among indolent B-cell lymphoproliferative disorders to splenic marginal zone lymphoma and therefore of some diagnostic use, with other NOTCH-pathway genes (NOTCH1, SPEN, DTX1) recurrently altered in a further share of cases.⁴⁰ A second, largely non-overlapping lesion is recurrent mutation of KLF2, a transcription factor required for normal marginal-zone B-cell development, identified across an independent multicentre series.¹⁹⁵

Extranodal marginal zone (MALT) lymphoma is genetically distinguished by recurrent chromosomal translocations that activate NF-κB, of which t(11;18)(q21;q21), producing the BIRC3::MALT1 fusion (also written API2::MALT1), is the best characterised; the resulting fusion protein has constitutive protease activity against NF-κB pathway substrates.¹⁹⁶ The translocation predicts that antibiotics will fail. Of H pylori-associated gastric MALT lymphomas, 20–30% do not regress after antibiotic therapy alone. In a series of ten antibiotic-responsive and twelve non-responsive gastric MALT lymphomas, the fusion transcript was detected in 9 of 12 (75%) non-responders and in none of the responders — identifying the translocation as a marker for patients who need lymphoma-directed treatment rather than antibiotics from the outset.¹⁹⁷

Hodgkin lymphoma

Classical Hodgkin lymphoma is unusual among the lymphomas discussed here in that the malignant Hodgkin and Reed–Sternberg cells are rare within the tumour mass, most of which is a reactive inflammatory infiltrate. This scarcity long delayed genomic study of the malignant clone, and is why several of the key findings came from laser-capture microdissection or, more recently, from sequencing circulating tumour DNA rather than the tumour tissue directly.

9p24.1 amplification and the case for PD-1 blockade

The genetic lesion with the clearest therapeutic consequence in Hodgkin lymphoma is amplification of chromosome 9p24.1, the locus containing PD-L1, PD-L2 and JAK2. Integrative analysis of copy number and transcription in Hodgkin and primary mediastinal large B-cell lymphoma cell lines and primary tumours found that 9p24.1 amplification was restricted to nodular-sclerosing Hodgkin lymphoma, the subtype most closely related to mediastinal B-cell lymphoma, and that the co-amplified JAK2 further induces PD-1 ligand transcription and sensitises the amplified cells to JAK2 inhibition.¹⁹⁸ In a validation cohort of 108 newly diagnosed classical Hodgkin lymphomas treated with the Stanford V regimen, 97% had a concordant genetic alteration of the PD-L1/PD-L2 locus — polysomy in 5%, copy gain in 56%, amplification in 36% — with amplification associated with advanced-stage disease and shorter progression-free survival.¹⁹⁹ Loss of β2-microglobulin/MHC class I expression, which impairs antigen presentation and CD8 T-cell recognition, was independently associated with inferior outcome, additive to and independent of 9p24.1 status.²⁰⁰ Together these findings explain, at the genetic level, why Hodgkin lymphoma is one of the most PD-1-blockade-responsive cancers known: essentially every case carries a genetic route to PD-1 ligand overexpression, whether or not it also escapes immune recognition by other means.

NF-κB, JAK-STAT and antigen presentation

Constitutive NF-κB activation is close to universal in Hodgkin and Reed–Sternberg cells, and one mechanism was identified as biallelic inactivating mutation of TNFAIP3 (A20), a negative regulator of NF-κB, acting as a tumour suppressor gene in both classical Hodgkin lymphoma and primary mediastinal B-cell lymphoma.²⁰¹ Microdissection of roughly 50 000 singly isolated Hodgkin and Reed–Sternberg cells from 34 biopsies, followed by coding-genome sequencing, found recurrent mutations in STAT6 (32% of cases), GNA13 (24%), XPO1 (18%) and ITPKB (16%); mutant STAT6 sustained tumour-cell viability and cooperated genetically and functionally with disruption of SOCS1, a JAK-STAT pathway inhibitor. Overall, 87% of cases showed JAK-STAT pathway dysregulation across this and several further genes (including STAT3, STAT5B, JAK1, JAK2 and PTPN1), reinforcing the JAK2-driven signal already amplified by 9p24.1.²⁰² Flow-sorted exome sequencing of primary Hodgkin and Reed-Sternberg cells has independently catalogued the coding mutations that make up this oncogenome, complementing the microdissection-based studies above.²⁰³

Circulating tumour DNA

Because the malignant clone is so sparse within a Hodgkin lymphoma biopsy, several of the genotyping studies above drew instead on circulating tumour DNA. Targeted genotyping of plasma from patients with classical Hodgkin lymphoma reliably recovered the tumour's genetic lesions, tracked clonal evolution over treatment, and detected residual or relapsed disease — establishing circulating tumour DNA as a practical route to genotyping a tumour that is otherwise difficult to sample directly, and as a tool for monitoring response that does not depend on repeat biopsy.²⁰⁴

Nodular lymphocyte-predominant Hodgkin lymphoma

Nodular lymphocyte-predominant Hodgkin lymphoma is genetically and clinically distinct from classical Hodgkin lymphoma: its malignant lymphocyte-predominant cells retain a B-cell immunophenotype (CD20-positive, typically CD30- and CD15-negative) that classical Hodgkin and Reed-Sternberg cells lack, and it behaves as a more indolent disease, often managed with less intensive therapy. The ICC renamed it nodular lymphocyte predominant B-cell lymphoma, stressing its differences from classical Hodgkin lymphoma.¹⁷⁸ WHO-HAEM5 keeps the Hodgkin name, so as not to disturb ongoing trials, but accepts the B-cell name ahead of a future change.¹⁷⁷

Mature T-cell and NK-cell lymphomas

The mature T- and NK-cell lymphomas remained the least genetically characterised major group of lymphoid cancers for longest, in part because peripheral T-cell lymphoma, not otherwise specified (PTCL-NOS), was for years a residual category defined by exclusion. Gene-expression profiling of 372 cases across the major entities has since resolved much of that residual category into molecularly coherent groups and identified two biological subgroups within what remains: a GATA3-high subgroup, associated with poor overall survival, and a TBX21-high subgroup, in which a high-cytotoxic-gene-expression signature also predicts poor outcome; together these accounted for roughly a third and a half, respectively, of PTCL-NOS cases still unclassifiable by other means.²⁰⁵ The same study reclassified 37% of morphologically diagnosed PTCL-NOS cases into other specific entities on the strength of their molecular signature, underlining how much of PTCL-NOS as traditionally diagnosed is really under-recognised angioimmunoblastic T-cell lymphoma (AITL), anaplastic large cell lymphoma, or extranodal NK/T-cell lymphoma.²⁰⁵

AITL and the other nodal lymphomas of T follicular helper (TFH) origin share a consistent mutational hierarchy. TET2 mutations, an early and non-specific lesion shared with several myeloid and lymphoid cancers, are present in a majority of cases and are also found in the surrounding non-tumour haematopoietic cells, implying they arise in a pre-malignant clone before the lymphoma proper develops.²⁰⁶ Against that background, a highly specific somatic mutation encoding RHOA p.Gly17Val was independently identified by two groups in 2014: present in 68% of AITL samples in one series (with every Gly17Val case also carrying a TET2 mutation) and in 67% of AITL and 18% of PTCL-NOS samples in the other; the mutant protein fails to bind GTP and dominantly interferes with wild-type RHOA function.^206,207 The same 2014 exome-sequencing study also identified less frequent but recurrent mutations in FYN, ATM, B2M and CD58 in PTCL, implicating impaired DNA damage response and immune escape alongside the RHOA/epigenetic axis.²⁰⁷ IDH2 mutations at the same arginine residues altered in glioma and acute myeloid leukaemia were found in around 20% of AITL in the discovery series and around 45% in an independent validation series — restricted, among peripheral T-cell lymphomas, to AITL, and the second common genetic lesion identified in the disease after TET2.²⁰⁸ DNMT3A mutations also recur, found in 33% (28 of 85) of AITL in the targeted mutational-landscape series that described them, every one of them alongside a TET2 mutation.²⁰⁹ Together with RHOA G17V, these epigenetic lesions give nodal TFH-derived lymphoma its genetic profile.

Anaplastic large cell lymphoma (ALCL) splits cleanly by ALK status. ALK-positive ALCL is defined by NPM1::ALK, from t(2;5), or one of several rarer ALK fusion partners, and carries a comparatively favourable prognosis. The fusion is itself directly targetable: in a Children's Oncology Group trial of the ALK inhibitor crizotinib in 26 patients with relapsed or refractory ALK-positive ALCL, the overall response rate was 83–90% depending on dose; on this evidence, the US Food and Drug Administration approved crizotinib in January 2021 for relapsed or refractory systemic ALCL in patients aged one year and older and young adults, the first approval for paediatric systemic ALCL.^214,215 ALK-negative ALCL lacks any ALK rearrangement and was, until recently, treated as a single prognostic entity; genetic subtyping across 73 ALK-negative and 32 ALK-positive ALCLs showed instead that ALK-negative disease splits into three groups with widely disparate outcomes. DUSP22 rearrangement, present in 30% of ALK-negative cases, defined a group with a 5-year overall survival of 90% — statistically indistinguishable from ALK-positive disease's 85% — while TP63 rearrangement, present in 8%, defined a group with a 5-year overall survival of only 17%; cases with neither genetic marker fell in between, at 42%.²¹⁰ DUSP22 and TP63 rearrangements were mutually exclusive and absent from ALK-positive tumours, and the outcome differences held after adjusting for the International Prognostic Index and persisted in patients treated with anthracycline-based chemotherapy without transplantation — making these two rearrangements the clearest example, outside DLBCL, of a genetic lesion that reclassifies risk within a morphologically uniform entity.²¹⁰ Brentuximab vedotin, an antibody–drug conjugate against CD30, is active across ALCL regardless of ALK or DUSP22/TP63 status: in a phase 2 trial of 58 patients with relapsed or refractory systemic ALCL, the objective response rate was 86% (50 of 58 patients), with 57% achieving complete remission and a further 29% a partial remission.²¹⁶ At a median observation of around six years, median overall survival had not been reached, no progressions occurred beyond 40 months, 66% of patients (38 of 58) had achieved a complete response by investigator assessment, and 8 patients (14% of the whole enrolled cohort) remained in sustained remission on single-agent brentuximab vedotin alone, without stem-cell transplantation or any further anticancer therapy.²¹⁷ CD30 expression, not genotype, is what selects patients for brentuximab.

Extranodal NK/T-cell lymphoma, nasal type, is almost always EBV-driven and carries a distinct mutational profile from the nodal TFH lymphomas above, with recurrent activating mutations in STAT3 and STAT5B — also found in the related entity of γδ-T-cell-derived lymphomas — providing a rationale for JAK-STAT-directed therapy, and recurrent mutation of the RNA helicase DDX3X, identified by whole-exome sequencing as a novel driver in the disease.^211,212 Mycosis fungoides and Sézary syndrome, the cutaneous T-cell lymphomas, differ again: whole-genome and exome sequencing found that somatic copy-number variants, not single-nucleotide mutations, dominate their genomic landscape, accounting for 92% of driver events against a mean of only one point mutation per case, and implicating genes in T-cell activation, apoptosis, NF-κB signalling, chromatin remodelling and the DNA damage response.²¹⁸ Adult T-cell leukaemia/lymphoma (ATLL) is caused by HTLV-1 infection, but the virus alone does not explain the malignancy. Integrated sequencing of 426 cases found that the acquired mutations converge on the HTLV-1 Tax protein's own interactome and on T-cell receptor–NF-κB signalling, with recurrent activating mutations in PLCG1, PRKCB, CARD11, VAV1, IRF4, CCR4 and CCR7, and recurrent gene fusions CTLA4::CD28 and ICOS::CD28 that are thought to help the malignant clone escape immune surveillance. The virus starts the disease; acquired mutations in the host cell complete it.²¹³

Multiple myeloma

Myeloma is preceded, in essentially every case, by an asymptomatic clonal plasma-cell state. In a prospective US screening cohort, serum samples collected from 71 people who went on to develop myeloma, taken 2 to nearly 10 years before diagnosis, showed monoclonal gammopathy of undetermined significance (MGUS) in 100% of samples taken 2 years before diagnosis and in 82·4% of samples taken 8 or more years before.²¹⁹ Between MGUS and overt myeloma sits an intermediate, still asymptomatic stage — smouldering myeloma — defined by a higher tumour burden than MGUS but the continued absence of the organ damage that defines active disease; both stages carry a genetic profile that already overlaps substantially with the myeloma that eventually emerges.

Two broadly separate genetic routes lead into myeloma, and a review of the field's molecular pathogenesis frames the disease around this split: chromosomal translocation into the immunoglobulin heavy-chain (IGH) locus on one hand, and hyperdiploidy on the other, with the two routes converging on a shared endpoint — dysregulation of the cyclin D/retinoblastoma pathway that drives plasma-cell proliferation regardless of which upstream lesion produced it.²²⁰ The recurrent IGH translocations juxtapose a partner oncogene with the strong IGH enhancer: t(4;14) dysregulates NSD2 (also called WHSC1/MMSET) and, because the translocation breakpoint sits close to it, often FGFR3 as well; t(11;14) dysregulates CCND1; t(14;16) dysregulates MAF; t(14;20) dysregulates MAFB; and t(6;14) dysregulates CCND3. Hyperdiploid myeloma instead carries trisomies of the odd-numbered chromosomes (typically 3, 5, 7, 9, 11, 15, 19 and 21) and tends not to carry a primary IGH translocation. Large-scale sequencing of newly diagnosed patients has since mapped how these primary events set up specific dependencies for what follows: t(4;14) associates with mutation of FGFR3, DIS3 and PRKD2; t(11;14) with mutation of CCND1 and IRF4; t(14;16) with mutation of MAF, BRAF, DIS3 and ATM; and hyperdiploidy with gain of chromosome 11q, mutation of FAM46C, and MYC rearrangement — a pattern the authors describe as a non-random accumulation of genetic hits built on the primary event, rather than a random walk through the genome.²²¹

Secondary genetic events accumulate on top of either primary route and are where most of the prognostically important, and potentially targetable, biology sits. Gain of chromosome 1q, and amplification of 1q with three or more copies, is one of the most consistently adverse secondary lesions and was significant enough to be added explicitly to a later risk-staging revision (below).²²² Deletion of 1p, deletion of 17p (removing one TP53 allele) and MYC rearrangement are further recurrent secondary events. Whole-exome sequencing of the UK Myeloma XI trial cohort (463 patients) found the mutational spectrum of newly diagnosed myeloma dominated by the RAS pathway, mutated in 43% of patients (chiefly KRAS, NRAS and BRAF), and by NF-κB pathway genes, mutated in 17%; mutations in CCND1 and in DNA-repair genes (TP53, ATM, ATR) carried a negative effect on survival, while mutations in IRF4 and EGR1 were associated with a more favourable course.²²³ A larger integrated genomic analysis of 1273 newly diagnosed patients added IDH1, IDH2, HUWE1, KLHL6 and PTPN11 to the list of significantly mutated driver genes, and found oncogene mutations to be more clonal — present in a larger fraction of the tumour — than tumour-suppressor mutations, consistent with the oncogenes exerting stronger positive selection.²²¹ Whole-exome sequencing of serial samples through the disease course has also identified kataegis — localised clustering of mutations attributed to APOBEC cytidine-deaminase activity — alongside hypermutation consistent with AID activity expected in a plasma-cell malignancy, the two processes contributing in varying amounts as the disease evolves, together with linear, branching and differential subclonal patterns of clonal evolution described in the section that follows.²²⁴

Risk stratification has moved from a purely clinical staging system to one that incorporates these genetic lesions directly. The Revised International Staging System (R-ISS), built from a pooled cohort of 4445 newly diagnosed patients across 11 international trials, combines the original ISS (serum β2-microglobulin and albumin) with high-risk cytogenetic abnormalities on fluorescence in situ hybridisation — del(17p), t(4;14) or t(14;16) — and serum lactate dehydrogenase. Five-year overall survival was 82% for R-ISS I, 62% for R-ISS II and 40% for R-ISS III.²²⁵ A second revision (R2-ISS), built from 10,843 patients across 16 trials in the European Myeloma Network's HARMONY project, added chromosome 1q gain/amplification as an explicit risk feature alongside ISS stage, del(17p) and LDH, and produced four risk groups by an additive points score: R2-ISS I (low, 19·2% of patients), II (low-intermediate, 30·8%), III (intermediate-high, 41·2%) and IV (high, the remainder).²²² Most recently, a 2025 consensus from the International Myeloma Society and the International Myeloma Working Group moved beyond staging to a specific genomic definition of high-risk disease, the Consensus Genomic Staging (CGS): patients are classed as genomically high-risk if they carry del(17p) at a clonal fraction above 20% and/or a TP53 mutation, an IGH translocation — t(4;14), t(14;16) or t(14;20) — together with 1q gain and/or del(1p32), monoallelic del(1p32) together with 1q gain, biallelic del(1p32), or a β2-microglobulin of 5·5 mg/L or more with a normal creatinine.²²⁶

The clearest example so far of matching a myeloma drug to a specific genetic lesion is venetoclax in t(11;14) disease, where the translocation drives a BCL2-dependent biology the drug is designed against. The randomised BELLINI trial added venetoclax or placebo to bortezomib and dexamethasone in unselected relapsed or refractory myeloma. Progression-free survival favoured venetoclax (22·4 vs 11·5 months; hazard ratio 0·63, 95% CI 0·44–0·90; p=0·010), but overall mortality was higher in the venetoclax group, driven mainly by fatal infection, and the trial's own conclusion was stated plainly: increased mortality was seen with venetoclax, underscoring the importance of selecting patients appropriately for this treatment.²²⁷ Because of that signal, venetoclax in myeloma is used, if at all, in t(11;14) disease, and off-label — no regulator has approved it for myeloma.²²⁸ A review of the trial's biomarker subgroups states that the increased mortality seen in BELLINI was primarily driven by patients without t(11;14),²²⁸ consistent with the earlier phase 1 monotherapy study, in which 12 of the 14 objective responses (86%) occurred in the 46% of patients who carried t(11;14) and the response rate in that subgroup (40%) was roughly double the rate overall (21%).²²⁹

Clonal evolution and co-mutation

A haematological cancer is rarely a single clone. Deep sequencing of paired diagnosis and relapse samples has shown that both linear evolution — where the relapse clone descends directly from the dominant diagnostic clone, having acquired further mutations — and branching evolution — where relapse arises from a minor subclone already present, sometimes barely detectable, at diagnosis — occur, sometimes in the same disease. Whole-genome sequencing of paired diagnosis/relapse AML samples showed two dominant patterns: either the founding clone gained mutations and evolved into the relapse clone, or a subclone of the founding clone survived chemotherapy, gained additional mutations, and expanded to cause relapse — direct evidence that chemotherapy can select for a pre-existing minor subclone rather than only causing new mutations.²³⁰ In CLL, sequencing of paired samples over the course of treatment showed subclones carrying specific driver mutations, including SF3B1 and TP53, were preferentially selected by chemotherapy and expanded at relapse, with the degree of subclonal complexity at diagnosis predicting how quickly the disease would progress.²³¹

Cytotoxic therapy itself can be a selective force that produces a second cancer. In four patients with therapy-related AML or MDS, the exact TP53 mutation found at diagnosis was already present at very low frequency (0·003–0·7%) in blood or marrow 3–6 years earlier, in two cases before any chemotherapy. In mice carrying both normal and Tp53-heterozygous stem cells, the Tp53-mutant cells expanded preferentially after chemotherapy. Chemotherapy did not create the TP53 mutation; it selected a clone that was already there.²³² The same logic applies to PPM1D (see Clonal haematopoiesis).

Single-cell sequencing — genotyping individual leukaemic cells rather than the bulk population — has clarified clonal architecture that bulk sequencing can only infer indirectly. Single-cell mutation analysis across several myeloid malignancies showed that clones distinguished only by their combination of mutations, invisible to bulk sequencing because the same mutations exist elsewhere in different combinations, can have markedly different fates during treatment and relapse.²³³ High-throughput single-cell genomics applied specifically to AML resolved the order in which mutations were acquired within individual patients and showed that the clonal hierarchy present at diagnosis directly shapes which clone survives chemotherapy and drives relapse.²³⁴

Targetable lesions and approved agents

Most genes named in this review have no drug built against them. The table lists the lesions that do, with the pivotal evidence, the regulator and the approval date.

Several widely used drugs are not in the table because they are not selected by genotype. JAK inhibitors (ruxolitinib and others) act on JAK–STAT signalling in myelofibrosis and polycythaemia vera whether the driver is JAK2, CALR or MPL: the drug blocks the kinase, not the mutation (see Myeloproliferative neoplasms). Bruton tyrosine kinase (BTK) inhibitors act on B-cell-receptor signalling in CLL and B-cell lymphomas and are prescribed without genetic selection. Genotype still changes how well they work — in Waldenström macroglobulinaemia, response depends on MYD88 and CXCR4 status (see Marginal zone and lymphoplasmacytic lymphoma) — and in CLL, BTK and PLCG2 mutations explain resistance (see Chronic lymphocytic leukaemia). Many of the genes named earlier have no directed agent at all: TP53, DNMT3A, TET2, ASXL1 and the spliceosome genes other than the ring-sideroblast/luspatercept association above are prognostic and classificatory, not yet targets a drug is built against. "Targetable" describes a minority of the genetic landscape, not most of it.

Testing in practice

The tests needed at diagnosis are set disease by disease. In AML, the ELN 2022 risk groups need karyotype and a mutation panel that covers at least the genes in its risk table (see European LeukemiaNet 2022 genetic risk);⁶⁴ FLT3, IDH1, IDH2 and NPM1 results also decide drug choice. In MDS, the IPSS-M needs sequencing of its 31 genes (see Myelodysplastic neoplasms). In CLL, IGHV status is tested once and del(17p)/TP53 before every line of treatment (see Chronic lymphocytic leukaemia). In myeloma, the 2025 genomic high-risk definition needs FISH for the IGH translocations, 1q gain, del(1p32) and del(17p), and TP53 sequencing (see Multiple myeloma).²²⁶ In CML, quantitative PCR on the International Scale tracks response (see Chronic myeloid leukaemia).¹¹⁶

Where a variant found in blood or marrow might be germline rather than acquired — relevant whenever a myeloid neoplasm could reflect an inherited predisposition syndrome, and increasingly relevant given how many transplant donors are screened for the same variant in the family — confirmation requires sequencing a non-haematopoietic tissue, because a variant present in both blood and marrow could still be somatic and clonal rather than truly germline. A pathology review of the diagnostic approach to germline predisposition in myeloid neoplasms states that the ideal tissue source for this confirmatory germline testing is cultured skin fibroblasts, obtained by skin biopsy and grown in culture to remove any contaminating haematopoietic cells before sequencing.²⁶¹

Uncertainties and future research

• Two classifications. WHO-HAEM5 and the ICC differ on blast thresholds, the boundary between MDS and AML, and TP53-mutated disease, and cohort studies show that many patients are classed differently by the two.⁶³ Until they are merged, reports and trials need to say which system they used.
• New adverse genes in AML. ELN 2022 moved patients with myelodysplasia-related gene mutations into the adverse group. In an independent cohort of 1118 patients, these patients did better than other adverse-risk patients, and the validators did not support the change.⁶⁵
• Clonal haematopoiesis. A risk score now picks out the small group of carriers at high risk of a myeloid neoplasm.⁴² What should be done for them, beyond follow-up, is not settled.
• Genetic subtypes of lymphoma. The DLBCL subtypes predict outcome, but the one link to a drug effect, ibrutinib in younger patients with MCD and N1 disease, comes from a retrospective analysis and needs a prospective trial.¹⁷⁹
• Ph-like ALL. Kinase inhibitors are chosen for Ph-like ALL on laboratory evidence; no completed randomised trial has tested them.¹⁰
• Menin inhibitors. The evidence so far is from single-arm trials in relapsed disease.^62,75 Their place earlier in treatment, and with chemotherapy, is untested.
• Venetoclax in myeloma. Benefit appears confined to t(11;14) disease, but the one randomised trial enrolled an unselected population and showed higher mortality overall.²²⁷ A trial restricted to t(11;14) is needed before routine use.
• Regional data. The published genetic data from Saudi Arabia that this review found are a single-centre AML series of 87 patients.⁶¹ The large sequencing cohorts come from Europe, North America and East Asia, and the frequency of germline predisposition variants such as DDX41 in Arab populations is not known.

Conclusion

The genetics of haematological cancer now does three jobs: it names the disease, it sets the risk, and for a minority of lesions it chooses the drug. The same few pathways recur across diseases, and a clinician who knows them can read a new mutation report without memorising it gene by gene. Most prognostic genes, TP53 above all, still have no drug directed at them, and several widely used drugs, the JAK and BTK inhibitors among them, are not chosen by genotype. The main open problems are practical: two classifications that disagree, risk scores whose newest changes are not yet validated, and too few data from the populations these patients come from.

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: 1703–19. doi:10.1038/s41375-022-01613-1 PMID 35732831
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: 1200–28. doi:10.1182/blood.2022015850 PMID 35767897
3. Ivey A, Hills RK, Simpson MA, Jovanovic JV, Gilkes A, Grech A, et al. Assessment of minimal residual disease in standard-risk AML. N Engl J Med 2016; 374: 422–33. doi:10.1056/NEJMoa1507471 PMID 26789727
4. Vogelstein B, Papadopoulos N, Velculescu VE, Zhou S, Diaz LA Jr, Kinzler KW. Cancer genome landscapes. Science 2013; 339: 1546-58. doi:10.1126/science.1235122 PMID 23539594
5. Lawrence MS, Stojanov P, Polak P, Kryukov GV, Cibulskis K, Sivachenko A, et al. Mutational heterogeneity in cancer and the search for new cancer-associated genes. Nature 2013; 499: 214-8. doi:10.1038/nature12213 PMID 23770567
6. Alexandrov LB, Nik-Zainal S, Wedge DC, Aparicio SA, Behjati S, Biankin AV, et al. Signatures of mutational processes in human cancer. Nature 2013; 500: 415-21. doi:10.1038/nature12477 PMID 23945592
7. Küppers R, Dalla-Favera R. Mechanisms of chromosomal translocations in B cell lymphomas. Oncogene 2001; 20: 5580-94. doi:10.1038/sj.onc.1204640 PMID 11607811
8. Stephens PJ, Greenman CD, Fu B, Yang F, Bignell GR, Mudie LJ, et al. Massive genomic rearrangement acquired in a single catastrophic event during cancer development. Cell 2011; 144: 27-40. doi:10.1016/j.cell.2010.11.055 PMID 21215367
9. Papaemmanuil E, Gerstung M, Bullinger L, Gaidzik VI, Paschka P, Roberts ND, et al. Genomic classification and prognosis in acute myeloid leukemia. N Engl J Med 2016; 374: 2209–21. doi:10.1056/NEJMoa1516192 PMID 27276561
10. Roberts KG, Li Y, Payne-Turner D, Harvey RC, Yang YL, Pei D, et al. Targetable kinase-activating lesions in Ph-like acute lymphoblastic leukemia. N Engl J Med 2014; 371: 1005–15. doi:10.1056/NEJMoa1403088 PMID 25207766
11. Levine RL, Wadleigh M, Cools J, Ebert BL, Wernig G, Huntly BJ, et al. Activating mutation in the tyrosine kinase JAK2 in polycythemia vera, essential thrombocythemia, and myeloid metaplasia with myelofibrosis. Cancer Cell 2005; 7: 387–97. doi:10.1016/j.ccr.2005.03.023 PMID 15837627
12. James C, Ugo V, Le Couédic JP, Staerk J, Delhommeau F, Lacout C, et al. A unique clonal JAK2 mutation leading to constitutive signalling causes polycythaemia vera. Nature 2005; 434: 1144–8. doi:10.1038/nature03546 PMID 15793561
13. Kralovics R, Passamonti F, Buser AS, Teo SS, Tiedt R, Passweg JR, et al. A gain-of-function mutation of JAK2 in myeloproliferative disorders. N Engl J Med 2005; 352: 1779–90. doi:10.1056/NEJMoa051113 PMID 15858187
14. Baxter EJ, Scott LM, Campbell PJ, East C, Fourouclas N, Swanton S, et al. Acquired mutation of the tyrosine kinase JAK2 in human myeloproliferative disorders. Lancet 2005; 365: 1054–61. doi:10.1016/S0140-6736(05)71142-9 PMID 15781101
15. Klampfl T, Gisslinger H, Harutyunyan AS, Nivarthi H, Rumi E, Milosevic JD, et al. Somatic mutations of calreticulin in myeloproliferative neoplasms. N Engl J Med 2013; 369: 2379–90. doi:10.1056/NEJMoa1311347 PMID 24325356
16. Nangalia J, Massie CE, Baxter EJ, Nice FL, Gundem G, Wedge DC, et al. Somatic CALR mutations in myeloproliferative neoplasms with nonmutated JAK2. N Engl J Med 2013; 369: 2391–2405. doi:10.1056/NEJMoa1312542 PMID 24325359
17. Koskela HL, Eldfors S, Ellonen P, van Adrichem AJ, Kuusanmäki H, Andersson EI, et al. Somatic STAT3 mutations in large granular lymphocytic leukemia. N Engl J Med 2012; 366: 1905-13. doi:10.1056/NEJMoa1114885 PMID 22591296
18. Ley TJ, Ding L, Walter MJ, McLellan MD, Lamprecht T, Larson DE, et al. DNMT3A mutations in acute myeloid leukemia. N Engl J Med 2010; 363: 2424-33. doi:10.1056/NEJMoa1005143 PMID 21067377
19. Delhommeau F, Dupont S, Della Valle V, James C, Trannoy S, Massé A, et al. Mutation in TET2 in myeloid cancers. N Engl J Med 2009; 360: 2289-301. doi:10.1056/NEJMoa0810069 PMID 19474426
20. Dang L, White DW, Gross S, Bennett BD, Bittinger MA, Driggers EM, et al. Cancer-associated IDH1 mutations produce 2-hydroxyglutarate. Nature 2009; 462: 739-44. doi:10.1038/nature08617 PMID 19935646
21. Ward PS, Patel J, Wise DR, Abdel-Wahab O, Bennett BD, Coller HA, et al. The common feature of leukemia-associated IDH1 and IDH2 mutations is a neomorphic enzyme activity converting alpha-ketoglutarate to 2-hydroxyglutarate. Cancer Cell 2010; 17: 225-34. doi:10.1016/j.ccr.2010.01.020 PMID 20171147
22. Figueroa ME, Abdel-Wahab O, Lu C, Ward PS, Patel J, Shih A, et al. Leukemic IDH1 and IDH2 mutations result in a hypermethylation phenotype, disrupt TET2 function, and impair hematopoietic differentiation. Cancer Cell 2010; 18: 553-67. doi:10.1016/j.ccr.2010.11.015 PMID 21130701
23. Gelsi-Boyer V, Trouplin V, Adélaïde J, Bonansea J, Cervera N, Carbuccia N, et al. Mutations of polycomb-associated gene ASXL1 in myelodysplastic syndromes and chronic myelomonocytic leukaemia. Br J Haematol 2009; 145: 788-800. doi:10.1111/j.1365-2141.2009.07697.x PMID 19388938
24. Ernst T, Chase AJ, Score J, Hidalgo-Curtis CE, Bryant C, Jones AV, et al. Inactivating mutations of the histone methyltransferase gene EZH2 in myeloid disorders. Nat Genet 2010; 42: 722-6. doi:10.1038/ng.621 PMID 20601953
25. Morin RD, Johnson NA, Severson TM, et al. Somatic mutations altering EZH2 (Tyr641) in follicular and diffuse large B-cell lymphomas of germinal-centre origin. Nat Genet 2010; 42: 181–5. doi:10.1038/ng.518 PMID 20081860
26. Yap DB, Chu J, Berg T, Schapira M, Cheng SW, Moradian A, et al. Somatic mutations at EZH2 Y641 act dominantly through a mechanism of selectively altered PRC2 catalytic activity, to increase H3K27 trimethylation. Blood 2011; 117: 2451-9. doi:10.1182/blood-2010-11-321208 PMID 21190999
27. Morin RD, Mendez-Lago M, Mungall AJ, et al. Frequent mutation of histone-modifying genes in non-Hodgkin lymphoma. Nature 2011; 476: 298–303. doi:10.1038/nature10351 PMID 21796119
28. Yoshida K, Sanada M, Shiraishi Y, Nowak D, Nagata Y, Yamamoto R, et al. Frequent pathway mutations of splicing machinery in myelodysplasia. Nature 2011; 478: 64-9. doi:10.1038/nature10496 PMID 21909114
29. Papaemmanuil E, Cazzola M, Boultwood J, Malcovati L, Vyas P, Bowen D, et al. Somatic SF3B1 mutation in myelodysplasia with ring sideroblasts. N Engl J Med 2011; 365: 1384–95. doi:10.1056/NEJMoa1103283 PMID 21995386
30. Kon A, Shih LY, Minamino M, Sanada M, Shiraishi Y, Nagata Y, et al. Recurrent mutations in multiple components of the cohesin complex in myeloid neoplasms. Nat Genet 2013; 45: 1232-7. doi:10.1038/ng.2731 PMID 23955599
31. Bernard E, Nannya Y, Hasserjian RP, Devlin SM, Tuechler H, Medina-Martinez JS, et al. Implications of TP53 allelic state for genome stability, clinical presentation and outcomes in myelodysplastic syndromes. Nat Med 2020; 26: 1549–56. doi:10.1038/s41591-020-1008-z PMID 32747829
32. 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: 1012–7. doi:10.1038/ng.913 PMID 21892162
33. Mullighan CG, Miller CB, Radtke I, Phillips LA, Dalton J, Ma J, et al. BCR-ABL1 lymphoblastic leukaemia is characterized by the deletion of Ikaros. Nature 2008; 453: 110-4. doi:10.1038/nature06866 PMID 18408710
34. Davis RE, Ngo VN, Lenz G, et al. Chronic active B-cell-receptor signalling in diffuse large B-cell lymphoma. Nature 2010; 463: 88–92. doi:10.1038/nature08638 PMID 20054396
35. Lenz G, Davis RE, Ngo VN, Lam L, George TC, Wright GW, et al. Oncogenic CARD11 mutations in human diffuse large B cell lymphoma. Science 2008; 319: 1676-9. doi:10.1126/science.1153629 PMID 18323416
36. Compagno M, Lim WK, Grunn A, Nandula SV, Brahmachary M, Shen Q, et al. Mutations of multiple genes cause deregulation of NF-kappaB in diffuse large B-cell lymphoma. Nature 2009; 459: 717-21. doi:10.1038/nature07968 PMID 19412164
37. Ngo VN, Young RM, Schmitz R, et al. Oncogenically active MYD88 mutations in human lymphoma. Nature 2011; 470: 115–9. doi:10.1038/nature09671 PMID 21179087
38. Treon SP, Xu L, Yang G, et al. MYD88 L265P somatic mutation in Waldenström's macroglobulinaemia. N Engl J Med 2012; 367: 826–33. doi:10.1056/NEJMoa1200710 PMID 22931316
39. Puente XS, Pinyol M, Quesada V, Conde L, Ordóñez GR, Villamor N, et al. Whole-genome sequencing identifies recurrent mutations in chronic lymphocytic leukaemia. Nature 2011; 475: 101-5. doi:10.1038/nature10113 PMID 21642962
40. Rossi D, Trifonov V, Fangazio M, et al. The coding genome of splenic marginal zone lymphoma: activation of NOTCH2 and other pathways regulating marginal zone development. J Exp Med 2012; 209: 1537–51. doi:10.1084/jem.20120904 PMID 22891273
41. Souers AJ, Leverson JD, Boghaert ER, Ackler SL, Catron ND, Chen J, et al. ABT-199, a potent and selective BCL-2 inhibitor, achieves antitumor activity while sparing platelets. Nat Med 2013; 19: 202-8. doi:10.1038/nm.3048 PMID 23291630
42. Weeks LD, Niroula A, Neuberg D, Wong W, Lindsley RC, Luskin M, et al. Prediction of risk for myeloid malignancy in clonal hematopoiesis. NEJM Evid 2023; 2: EVIDoa2200310. doi:10.1056/evidoa2200310 PMID 37483562
43. Jaiswal S, Fontanillas P, Flannick J, Manning A, Grauman PV, Mar BG, et al. Age-related clonal hematopoiesis associated with adverse outcomes. N Engl J Med 2014; 371: 2488–98. doi:10.1056/NEJMoa1408617 PMID 25426837
44. Genovese G, Kähler AK, Handsaker RE, Lindberg J, Rose SA, Bakhoum SF, et al. Clonal hematopoiesis and blood-cancer risk inferred from blood DNA sequence. N Engl J Med 2014; 371: 2477–87. doi:10.1056/NEJMoa1409405 PMID 25426838
45. Jaiswal S, Natarajan P, Silver AJ, Gibson CJ, Bick AG, Shvartz E, et al. Clonal hematopoiesis and risk of atherosclerotic cardiovascular disease. N Engl J Med 2017; 377: 111–21. doi:10.1056/NEJMoa1701719 PMID 28636844
46. Bolton KL, Ptashkin RN, Gao T, Braunstein L, Devlin SM, Kelly D, et al. Cancer therapy shapes the fitness landscape of clonal hematopoiesis. Nat Genet 2020; 52: 1219–26. doi:10.1038/s41588-020-00710-0 PMID 33106634
47. Godley LA, Shimamura A. Genetic predisposition to hematologic malignancies: management and surveillance. Blood 2017; 130: 424–32. doi:10.1182/blood-2017-02-735290 PMID 28600339
48. 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: 534–49. doi:10.1182/blood.2022018221 PMID 36322930
49. 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: 658–70. doi:10.1016/j.ccell.2015.03.017 PMID 25920683
50. Kim B, Choi DH, Jang JH, Jung CW, Kim HJ, Kim HY. Clinical and molecular spectrum of DDX41 variants in Korean patients with hematologic malignancies. J Clin Med 2025; 14: 7999. doi:10.3390/jcm14227999 PMID 41303035
51. 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: 166–75. doi:10.1038/13793 PMID 10508512
52. Smith ML, Cavenagh JD, Lister TA, Fitzgibbon J. Mutation of CEBPA in familial acute myeloid leukemia. N Engl J Med 2004; 351: 2403–7. doi:10.1056/NEJMoa041331 PMID 15575056
53. 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: 180–5. doi:10.1038/ng.3177 PMID 25581430
54. 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: 115–20. doi:10.1016/j.ajhg.2010.12.006 PMID 21211618
55. 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: 792–7. doi:10.1038/ng.3569 PMID 27182967
56. 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: 1146–58. doi:10.1016/j.ajhg.2016.04.009 PMID 27259050
57. Savage SA. Telomeres and human disease. Cold Spring Harb Perspect Biol 2025; 17: a041684. doi:10.1101/cshperspect.a041684 PMID 40097157
58. Alter BP, Giri N, McReynolds LJ, Altintas B. Fanconi anaemia: a syndrome with distinct subgroups. Br J Haematol 2022; 197: 467–74. doi:10.1111/bjh.18091 PMID 35191533
59. Cancer Genome Atlas Research Network, Ley TJ, Miller C, Ding L, Raphael BJ, Mungall AJ, et al. Genomic and epigenomic landscapes of adult de novo acute myeloid leukemia. N Engl J Med 2013; 368: 2059–74. doi:10.1056/NEJMoa1301689 PMID 23634996
60. Tyner JW, Tognon CE, Bottomly D, Wilmot B, Kurtz SE, Savage SL, et al. Functional genomic landscape of acute myeloid leukaemia. Nature 2018; 562: 526–31. doi:10.1038/s41586-018-0623-z PMID 30333627
61. Alsobhi E, Farahat F, Daghistani M, Awad K, Al-Zahran O, Al-Saiari A, et al. Overall survival of adult acute myeloid leukemia based on cytogenetic and molecular abnormalities during 5 years in a single center study. Saudi Med J 2019; 40: 1171–6. doi:10.15537/smj.2019.11.24584 PMID 31707416
62. Issa GC, Aldoss I, DiPersio J, Cuglievan B, Stone R, Arellano M, et al. The menin inhibitor revumenib in KMT2A-rearranged or NPM1-mutant leukaemia. Nature 2023; 615: 920–4. doi:10.1038/s41586-023-05812-3 PMID 36922593
63. Koorse Germans S, Weinberg OK. Navigating diagnostic shifts: a comparative review of the WHO 5th edition and ICC 2022 classifications of myeloid neoplasms. Hum Pathol 2026; 169: 106077. doi:10.1016/j.humpath.2026.106077 PMID 41720165
64. 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: 1345–77. doi:10.1182/blood.2022016867 PMID 35797463
65. Rausch C, Rothenberg-Thurley M, Dufour A, Schneider S, Gittinger H, Sauerland C, et al. Validation and refinement of the 2022 European LeukemiaNet genetic risk stratification of acute myeloid leukemia. Leukemia 2023; 37: 1234–44. doi:10.1038/s41375-023-01884-2 PMID 37041198
66. Perrone S, Tresoldi C, Rigamonti S, Molica M, Zhdanovskaya N, Cicconi L. How to read a next-generation sequencing report for AML and MDS? What hematologists need to know. J Clin Med 2025; 14: 8681. doi:10.3390/jcm14248681 PMID 41464583
67. Stone RM, Mandrekar SJ, Sanford BL, Laumann K, Geyer S, Bloomfield CD, et al. Midostaurin plus chemotherapy for acute myeloid leukemia with a FLT3 mutation. N Engl J Med 2017; 377: 454–64. doi:10.1056/NEJMoa1614359 PMID 28644114
68. Perl AE, Martinelli G, Cortes JE, Neubauer A, Berman E, Paolini S, et al. Gilteritinib or chemotherapy for relapsed or refractory FLT3-mutated AML. N Engl J Med 2019; 381: 1728–40. doi:10.1056/NEJMoa1902688 PMID 31665578
69. Erba HP, Montesinos P, Kim HJ, Patkowska E, Vrhovac R, Žák P, et al. Quizartinib plus chemotherapy in newly diagnosed patients with FLT3-internal-tandem-duplication-positive acute myeloid leukaemia (QuANTUM-First): a randomised, double-blind, placebo-controlled, phase 3 trial. Lancet 2023; 401: 1571–83. doi:10.1016/S0140-6736(23)00464-6 PMID 37116523
70. DiNardo CD, Stein EM, de Botton S, Roboz GJ, Altman JK, Mims AS, et al. Durable remissions with ivosidenib in IDH1-mutated relapsed or refractory AML. N Engl J Med 2018; 378: 2386–98. doi:10.1056/NEJMoa1716984 PMID 29860938
71. Montesinos P, Recher C, Vives S, Zarzycka E, Wang J, Bertani G, et al. Ivosidenib and azacitidine in IDH1-mutated acute myeloid leukemia. N Engl J Med 2022; 386: 1519–31. doi:10.1056/NEJMoa2117344 PMID 35443108
72. de Botton S, Fenaux P, Yee K, Récher C, Wei AH, Montesinos P, et al. Olutasidenib (FT-2102) induces durable complete remissions in patients with relapsed or refractory IDH1-mutated AML. Blood Adv 2023; 7: 3117–27. doi:10.1182/bloodadvances.2022009411 PMID 36724515
73. Stein EM, DiNardo CD, Pollyea DA, Fathi AT, Roboz GJ, Altman JK, et al. Enasidenib in mutant IDH2 relapsed or refractory acute myeloid leukemia. Blood 2017; 130: 722–31. doi:10.1182/blood-2017-04-779405 PMID 28588020
74. Wang ES, Issa GC, Erba HP, Altman JK, Montesinos P, DeBotton S, et al. Ziftomenib in relapsed or refractory acute myeloid leukaemia (KOMET-001): a multicentre, open-label, multi-cohort, phase 1 trial. Lancet Oncol 2024; 25: 1310–24. doi:10.1016/S1470-2045(24)00386-3 PMID 39362248
75. Wang ES, Montesinos P, Foran J, Erba H, Rodríguez-Arbolí E, Fedorov K, et al. Ziftomenib in relapsed or refractory NPM1-mutated AML. J Clin Oncol 2025; 43: 3381–90. doi:10.1200/JCO-25-01694 PMID 40997296
76. Papaemmanuil E, Gerstung M, Malcovati L, Tauro S, Gundem G, Van Loo P, et al. Clinical and biological implications of driver mutations in myelodysplastic syndromes. Blood 2013; 122: 3616–27. doi:10.1182/blood-2013-08-518886 PMID 24030381
77. Haferlach T, Nagata Y, Grossmann V, Okuno Y, Bacher U, Nagae G, et al. Landscape of genetic lesions in 944 patients with myelodysplastic syndromes. Leukemia 2014; 28: 241–7. doi:10.1038/leu.2013.336 PMID 24220272
78. Malcovati L, Stevenson K, Papaemmanuil E, Neuberg D, Bejar R, Boultwood J, et al. SF3B1-mutant MDS as a distinct disease subtype: a proposal from the International Working Group for the Prognosis of MDS. Blood 2020; 136: 157–70. doi:10.1182/blood.2020004850 PMID 32347921
79. Ebert BL, Pretz J, Bosco J, Chang CY, Tamayo P, Galili N, et al. Identification of RPS14 as a 5q- syndrome gene by RNA interference screen. Nature 2008; 451: 335–9. doi:10.1038/nature06494 PMID 18202658
80. List A, Dewald G, Bennett J, Giagounidis A, Raza A, Feldman E, et al. Lenalidomide in the myelodysplastic syndrome with chromosome 5q deletion. N Engl J Med 2006; 355: 1456–65. doi:10.1056/NEJMoa061292 PMID 17021321
81. Greenberg PL, Tuechler H, Schanz J, Sanz G, Garcia-Manero G, Solé F, et al. Revised international prognostic scoring system for myelodysplastic syndromes. Blood 2012; 120: 2454–65. doi:10.1182/blood-2012-03-420489 PMID 22740453
82. Bernard E, Tuechler H, Greenberg PL, Hasserjian RP, Arango Ossa JE, Nannya Y, et al. Molecular International Prognostic Scoring System for myelodysplastic syndromes. NEJM Evid 2022; 1: EVIDoa2200008. doi:10.1056/EVIDoa2200008 PMID 38319256
83. Pazzi C, DeZern AE. SOHO state of the art updates and next questions: progress in higher-risk myelodysplastic syndromes. Clin Lymphoma Myeloma Leuk 2026; 26: e987–95. doi:10.1016/j.clml.2026.06.010 PMID 42436094
84. Fenaux P, Platzbecker U, Mufti GJ, Garcia-Manero G, Buckstein R, Santini V, et al. Luspatercept in patients with lower-risk myelodysplastic syndromes. N Engl J Med 2020; 382: 140–51. doi:10.1056/NEJMoa1908892 PMID 31914241
85. Platzbecker U, Santini V, Fenaux P, Sekeres MA, Savona MR, Madanat YF, et al. Imetelstat in patients with lower-risk myelodysplastic syndromes who have relapsed or are refractory to erythropoiesis-stimulating agents (IMerge): a multinational, randomised, double-blind, placebo-controlled, phase 3 trial. Lancet 2024; 403: 249–60. doi:10.1016/S0140-6736(23)01724-5 PMID 38048786
86. Scott LM, Tong W, Levine RL, Scott MA, Beer PA, Stratton MR, et al. JAK2 exon 12 mutations in polycythemia vera and idiopathic erythrocytosis. N Engl J Med 2007; 356: 459–68. doi:10.1056/NEJMoa065202 PMID 17267906
87. Pikman Y, Lee BH, Mercher T, McDowell E, Ebert BL, Gozo M, et al. MPLW515L is a novel somatic activating mutation in myelofibrosis with myeloid metaplasia. PLoS Med 2006; 3: e270. doi:10.1371/journal.pmed.0030270 PMID 16834459
88. Pardanani AD, Levine RL, Lasho T, Pikman Y, Mesa RA, Wadleigh M, et al. MPL515 mutations in myeloproliferative and other myeloid disorders: a study of 1182 patients. Blood 2006; 108: 3472–6. doi:10.1182/blood-2006-04-018879 PMID 16868251
89. Tefferi A, Lasho TL, Finke CM, Knudson RA, Ketterling R, Hanson CH, et al. CALR vs JAK2 vs MPL-mutated or triple-negative myelofibrosis: clinical, cytogenetic and molecular comparisons. Leukemia 2014; 28: 1472–7. doi:10.1038/leu.2014.3 PMID 24402162
90. Guglielmelli P, Lasho TL, Rotunno G, Mudireddy M, Mannarelli C, Nicolosi M, et al. MIPSS70: Mutation-Enhanced International Prognostic Score System for Transplantation-Age Patients With Primary Myelofibrosis. J Clin Oncol 2018; 36: 310–8. doi:10.1200/JCO.2017.76.4886 PMID 29226763
91. Tefferi A, Guglielmelli P, Larson DR, Finke C, Wassie EA, Pieri L, et al. Long-term survival and blast transformation in molecularly annotated essential thrombocythemia, polycythemia vera, and myelofibrosis. Blood 2014; 124: 2507–13. doi:10.1182/blood-2014-05-579136 PMID 25037629
92. Ortmann CA, Kent DG, Nangalia J, Silber Y, Wedge DC, Grinfeld J, et al. Effect of mutation order on myeloproliferative neoplasms. N Engl J Med 2015; 372: 601–12. doi:10.1056/NEJMoa1412098 PMID 25671252
93. Williams N, Lee J, Mitchell E, Moore L, Baxter EJ, Hewinson J, et al. Life histories of myeloproliferative neoplasms inferred from phylogenies. Nature 2022; 602: 162–8. doi:10.1038/s41586-021-04312-6 PMID 35058638
94. Van Egeren D, Escabi J, Nguyen M, Liu S, Reilly CR, Patel S, et al. Reconstructing the lineage histories and differentiation trajectories of individual cancer cells in myeloproliferative neoplasms. Cell Stem Cell 2021; 28: 514–23.e9. doi:10.1016/j.stem.2021.02.001 PMID 33621486
95. Tefferi A, Guglielmelli P, Lasho TL, Gangat N, Ketterling RP, Pardanani A, et al. MIPSS70+ Version 2.0: Mutation and Karyotype-Enhanced International Prognostic Scoring System for Primary Myelofibrosis. J Clin Oncol 2018; 36: 1769–70. doi:10.1200/JCO.2018.78.9867 PMID 29708808
96. Tefferi A, Guglielmelli P, Nicolosi M, Mannelli F, Mudireddy M, Bartalucci N, et al. GIPSS: genetically inspired prognostic scoring system for primary myelofibrosis. Leukemia 2018; 32: 1631–42. doi:10.1038/s41375-018-0107-z PMID 29654267
97. Verstovsek S, Mesa RA, Gotlib J, Levy RS, Gupta V, DiPersio JF, et al. A double-blind, placebo-controlled trial of ruxolitinib for myelofibrosis. N Engl J Med 2012; 366: 799–807. doi:10.1056/NEJMoa1110557 PMID 22375971
98. Harrison C, Kiladjian JJ, Al-Ali HK, Gisslinger H, Waltzman R, Stalbovskaya V, et al. JAK inhibition with ruxolitinib versus best available therapy for myelofibrosis. N Engl J Med 2012; 366: 787–98. doi:10.1056/NEJMoa1110556 PMID 22375970
99. Pardanani A, Harrison C, Cortes JE, Cervantes F, Mesa RA, Milligan D, et al. Safety and Efficacy of Fedratinib in Patients With Primary or Secondary Myelofibrosis: A Randomized Clinical Trial. JAMA Oncol 2015; 1: 643–51. doi:10.1001/jamaoncol.2015.1590 PMID 26181658
100. Blair HA. Fedratinib: first approval. Drugs 2019; 79: 1719–25. doi:10.1007/s40265-019-01205-x PMID 31571162
101. Mascarenhas J, Hoffman R, Talpaz M, Gerds AT, Stein B, Gupta V, et al. Pacritinib vs Best Available Therapy, Including Ruxolitinib, in Patients With Myelofibrosis: A Randomized Clinical Trial. JAMA Oncol 2018; 4: 652–9. doi:10.1001/jamaoncol.2017.5818 PMID 29522138
102. Verstovsek S, Gerds AT, Vannucchi AM, Al-Ali HK, Lavie D, Kuykendall AT, et al. Momelotinib versus danazol in symptomatic patients with anaemia and myelofibrosis (MOMENTUM): results from an international, double-blind, randomised, controlled, phase 3 study. Lancet 2023; 401: 269–80. doi:10.1016/S0140-6736(22)02036-0 PMID 36709073
103. Gisslinger H, Klade C, Georgiev P, Krochmalczyk D, Gercheva-Kyuchukova L, Egyed M, et al. Ropeginterferon alfa-2b versus standard therapy for polycythaemia vera (PROUD-PV and CONTINUATION-PV): a randomised, non-inferiority, phase 3 trial and its extension study. Lancet Haematol 2020; 7: e196–208. doi:10.1016/S2352-3026(19)30236-4 PMID 32014125
104. Rowley JD. A new consistent chromosomal abnormality in chronic myelogenous leukaemia identified by quinacrine fluorescence and Giemsa staining. Nature 1973; 243: 290–3. doi:10.1038/243290a0 PMID 4126434
105. Lichty BD, Keating A, Callum J, Yee K, Croxford R, Corpus G, et al. Expression of p210 and p190 BCR-ABL due to alternative splicing in chronic myelogenous leukaemia. Br J Haematol 1998; 103: 711–5. doi:10.1046/j.1365-2141.1998.01033.x PMID 9858221
106. Verstovsek S, Lin H, Kantarjian H, Saglio G, De Micheli D, Pane F, et al. Neutrophilic-chronic myeloid leukemia: low levels of p230 BCR/ABL mRNA and undetectable BCR/ABL protein may predict an indolent course. Cancer 2002; 94: 2416–25. doi:10.1002/cncr.10490 PMID 12015767
107. O'Brien SG, Guilhot F, Larson RA, Gathmann I, Baccarani M, Cervantes F, et al. Imatinib compared with interferon and low-dose cytarabine for newly diagnosed chronic-phase chronic myeloid leukemia. N Engl J Med 2003; 348: 994–1004. doi:10.1056/NEJMoa022457 PMID 12637609
108. Hochhaus A, Larson RA, Guilhot F, Radich JP, Branford S, Hughes TP, et al. Long-Term Outcomes of Imatinib Treatment for Chronic Myeloid Leukemia. N Engl J Med 2017; 376: 917–27. doi:10.1056/NEJMoa1609324 PMID 28273028
109. Bower H, Björkholm M, Dickman PW, Höglund M, Lambert PC, Andersson TM. Life expectancy of patients with chronic myeloid leukemia approaches the life expectancy of the general population. J Clin Oncol 2016; 34: 2851–7. doi:10.1200/JCO.2015.66.2866 PMID 27325849
110. Gorre ME, Mohammed M, Ellwood K, Hsu N, Paquette R, Rao PN, et al. Clinical resistance to STI-571 cancer therapy caused by BCR-ABL gene mutation or amplification. Science 2001; 293: 876–80. doi:10.1126/science.1062538 PMID 11423618
111. Cortes JE, Kim DW, Pinilla-Ibarz J, le Coutre P, Paquette R, Chuah C, et al. A phase 2 trial of ponatinib in Philadelphia chromosome-positive leukemias. N Engl J Med 2013; 369: 1783–96. doi:10.1056/NEJMoa1306494 PMID 24180494
112. Réa D, Mauro MJ, Boquimpani C, Minami Y, Lomaia E, Voloshin S, et al. A phase 3, open-label, randomized study of asciminib, a STAMP inhibitor, vs bosutinib in CML after 2 or more prior TKIs. Blood 2021; 138: 2031–41. doi:10.1182/blood.2020009984 PMID 34407542
113. Hochhaus A, Wang J, Kim DW, Kim DDH, Mayer J, Goh YT, et al. Asciminib in Newly Diagnosed Chronic Myeloid Leukemia. N Engl J Med 2024; 391: 885–98. doi:10.1056/NEJMoa2400858 PMID 38820078
114. Grossmann V, Kohlmann A, Zenger M, Schindela S, Eder C, Weissmann S, et al. A deep-sequencing study of chronic myeloid leukemia patients in blast crisis (BC-CML) detects mutations in 76.9% of cases. Leukemia 2011; 25: 557–60. doi:10.1038/leu.2010.298 PMID 21274004
115. Adnan Awad S, Kankainen M, Ojala T, Koskenvesa P, Eldfors S, Ghimire B, et al. Mutation accumulation in cancer genes relates to nonoptimal outcome in chronic myeloid leukemia. Blood Adv 2020; 4: 546–59. doi:10.1182/bloodadvances.2019000943 PMID 32045476
116. Hochhaus A, Baccarani M, Silver RT, Schiffer C, Apperley JF, Cervantes F, et al. European LeukemiaNet 2020 recommendations for treating chronic myeloid leukemia. Leukemia 2020; 34: 966–84. doi:10.1038/s41375-020-0776-2 PMID 32127639
117. Elena C, Gallì A, Such E, Meggendorfer M, Germing U, Rizzo E, et al. Integrating clinical features and genetic lesions in the risk assessment of patients with chronic myelomonocytic leukemia. Blood 2016; 128: 1408–17. doi:10.1182/blood-2016-05-714030 PMID 27385790
118. Piazza R, Valletta S, Winkelmann N, Redaelli S, Spinelli R, Pirola A, et al. Recurrent SETBP1 mutations in atypical chronic myeloid leukemia. Nat Genet 2013; 45: 18–24. doi:10.1038/ng.2495 PMID 23222956
119. Gambacorti-Passerini CB, Donadoni C, Parmiani A, Pirola A, Redaelli S, Signore G, et al. Recurrent ETNK1 mutations in atypical chronic myeloid leukemia. Blood 2015; 125: 499–503. doi:10.1182/blood-2014-06-579466 PMID 25343957
120. Maxson JE, Gotlib J, Pollyea DA, Fleischman AG, Agarwal A, Eide CA, et al. Oncogenic CSF3R mutations in chronic neutrophilic leukemia and atypical CML. N Engl J Med 2013; 368: 1781–90. doi:10.1056/NEJMoa1214514 PMID 23656643
121. Garcia-Montero AC, Jara-Acevedo M, Teodosio C, Sanchez ML, Nunez R, Prados A, et al. KIT mutation in mast cells and other bone marrow hematopoietic cell lineages in systemic mast cell disorders: a prospective study of the Spanish Network on Mastocytosis (REMA) in a series of 113 patients. Blood 2006; 108: 2366–72. doi:10.1182/blood-2006-04-015545 PMID 16741248
122. Gotlib J, Kluin-Nelemans HC, George TI, Akin C, Sotlar K, Hermine O, et al. Efficacy and Safety of Midostaurin in Advanced Systemic Mastocytosis. N Engl J Med 2016; 374: 2530–41. doi:10.1056/NEJMoa1513098 PMID 27355533
123. DeAngelo DJ, Radia DH, George TI, Robinson WA, Quiery AT, Drummond MW, et al. Safety and efficacy of avapritinib in advanced systemic mastocytosis: the phase 1 EXPLORER trial. Nat Med 2021; 27: 2183–91. doi:10.1038/s41591-021-01538-9 PMID 34873347
124. Gotlib J, Reiter A, Radia DH, Deininger MW, George TI, Panse J, et al. Efficacy and safety of avapritinib in advanced systemic mastocytosis: interim analysis of the phase 2 PATHFINDER trial. Nat Med 2021; 27: 2192–9. doi:10.1038/s41591-021-01539-8 PMID 34873345
125. Cools J, DeAngelo DJ, Gotlib J, Stover EH, Legare RD, Cortes J, et al. A tyrosine kinase created by fusion of the PDGFRA and FIP1L1 genes as a therapeutic target of imatinib in idiopathic hypereosinophilic syndrome. N Engl J Med 2003; 348: 1201–14. doi:10.1056/NEJMoa025217 PMID 12660384
126. Iacobucci I, Mullighan CG. Genetic basis of acute lymphoblastic leukemia. J Clin Oncol 2017; 35: 975–83. doi:10.1200/JCO.2016.70.7836 PMID 28297628
127. Moorman AV, Ensor HM, Richards SM, Chilton L, Schwab C, Kinsey SE, et al. Prognostic effect of chromosomal abnormalities in childhood B-cell precursor acute lymphoblastic leukaemia: results from the UK Medical Research Council ALL97/99 randomised trial. Lancet Oncol 2010; 11: 429–38. doi:10.1016/S1470-2045(10)70066-8 PMID 20409752
128. Holmfeldt L, Wei L, Diaz-Flores E, Walsh M, Zhang J, Ding L, et al. The genomic landscape of hypodiploid acute lymphoblastic leukemia. Nat Genet 2013; 45: 242–52. doi:10.1038/ng.2532 PMID 23334668
129. Rand V, Parker H, Russell LJ, Schwab C, Ensor H, Irving J, et al. Genomic characterization implicates iAMP21 as a likely primary genetic event in childhood B-cell precursor acute lymphoblastic leukemia. Blood 2011; 117: 6848–55. doi:10.1182/blood-2011-01-329961 PMID 21527530
130. Harrison CJ, Moorman AV, Schwab C, Carroll AJ, Raetz EA, Devidas M, et al. An international study of intrachromosomal amplification of chromosome 21 (iAMP21): cytogenetic characterization and outcome. Leukemia 2014; 28: 1015–21. doi:10.1038/leu.2013.317 PMID 24166298
131. Fischer U, Forster M, Rinaldi A, Risch T, Sungalee S, Warnatz HJ, et al. Genomics and drug profiling of fatal TCF3-HLF-positive acute lymphoblastic leukemia identifies recurrent mutation patterns and therapeutic options. Nat Genet 2015; 47: 1020–29. doi:10.1038/ng.3362 PMID 26214592
132. Gu Z, Churchman ML, Roberts KG, Moore I, Zhou X, Nakitandwe J, et al. PAX5-driven subtypes of B-progenitor acute lymphoblastic leukemia. Nat Genet 2019; 51: 296–307. doi:10.1038/s41588-018-0315-5 PMID 30643249
133. Passet M, Boissel N, Sigaux F, Saillard C, Bargetzi M, Ba I, et al. PAX5 P80R mutation identifies a novel subtype of B-cell precursor acute lymphoblastic leukemia with favorable outcome. Blood 2019; 133: 280–4. doi:10.1182/blood-2018-10-882142 PMID 30510083
134. Lilljebörn H, Henningsson R, Hyrenius-Wittsten A, Olsson L, Orsmark-Pietras C, von Palffy S, et al. Identification of ETV6-RUNX1-like and DUX4-rearranged subtypes in paediatric B-cell precursor acute lymphoblastic leukaemia. Nat Commun 2016; 7: 11790. doi:10.1038/ncomms11790 PMID 27265895
135. Zhang J, McCastlain K, Yoshihara H, Xu B, Chang Y, Churchman ML, et al. Deregulation of DUX4 and ERG in acute lymphoblastic leukemia. Nat Genet 2016; 48: 1481–9. doi:10.1038/ng.3691 PMID 27776115
136. Ohki K, Kiyokawa N, Saito Y, Hirabayashi S, Nakabayashi K, Ichikawa H, et al. Clinical and molecular characteristics of MEF2D fusion-positive B-cell precursor acute lymphoblastic leukemia in childhood. Haematologica 2019; 104: 128–37. doi:10.3324/haematol.2017.186320 PMID 30171027
137. Hirabayashi S, Ohki K, Nakabayashi K, Ichikawa H, Momozawa Y, Okamura K, et al. ZNF384-related fusion genes define a subgroup of childhood B-cell precursor acute lymphoblastic leukemia with a characteristic immunotype. Haematologica 2017; 102: 118–29. doi:10.3324/haematol.2016.151035 PMID 27634205
138. Mullighan CG, Su X, Zhang J, Radtke I, Phillips LA, Miller CB, et al. Deletion of IKZF1 and prognosis in acute lymphoblastic leukemia. N Engl J Med 2009; 360: 470–80. doi:10.1056/NEJMoa0808253 PMID 19129520
139. Foà R, Bassan R, Vitale A, Elia L, Piciocchi A, Puzzolo MC, et al. Dasatinib-blinatumomab for Ph-positive acute lymphoblastic leukemia in adults. N Engl J Med 2020; 383: 1613–23. doi:10.1056/NEJMoa2016272 PMID 33085860
140. Papaemmanuil E, Hosking FJ, Vijayakrishnan J, Price A, Olver B, Sheridan E, et al. Loci on 7p12.2, 10q21.2 and 14q11.2 are associated with risk of childhood acute lymphoblastic leukemia. Nat Genet 2009; 41: 1006–10. doi:10.1038/ng.430 PMID 19684604
141. Treviño LR, Yang W, French D, Hunger SP, Carroll WL, Devidas M, et al. Germline genomic variants associated with childhood acute lymphoblastic leukemia. Nat Genet 2009; 41: 1001–5. doi:10.1038/ng.432 PMID 19684603
142. Perez-Andreu V, Roberts KG, Harvey RC, Yang W, Cheng C, Pei D, et al. Inherited GATA3 variants are associated with Ph-like childhood acute lymphoblastic leukemia and risk of relapse. Nat Genet 2013; 45: 1494–8. doi:10.1038/ng.2803 PMID 24141364
143. Qian M, Xu H, Perez-Andreu V, Roberts KG, Zhang H, Yang W, et al. Novel susceptibility variants at the ERG locus for childhood acute lymphoblastic leukemia in Hispanics. Blood 2019; 133: 724–9. doi:10.1182/blood-2018-07-862946 PMID 30510082
144. 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: e1005262. doi:10.1371/journal.pgen.1005262 PMID 26102509
145. Shah S, Schrader KA, Waanders E, Timms AE, Vijai J, Miething C, et al. A recurrent germline PAX5 mutation confers susceptibility to pre-B cell acute lymphoblastic leukemia. Nat Genet 2013; 45: 1226–31. doi:10.1038/ng.2754 PMID 24013638
146. Churchman ML, Qian M, Te Kronnie G, Zhang R, Yang W, Zhang H, et al. Germline genetic IKZF1 variation and predisposition to childhood acute lymphoblastic leukemia. Cancer Cell 2018; 33: 937–948.e8. doi:10.1016/j.ccell.2018.03.021 PMID 29681510
147. Jabbour E, Kantarjian HM, Aldoss I, Montesinos P, Leonard JT, Gómez-Almaguer D, et al. Ponatinib vs imatinib in frontline Philadelphia chromosome-positive acute lymphoblastic leukemia: a randomized clinical trial. JAMA 2024; 331: 1814–23. doi:10.1001/jama.2024.4783 PMID 38722621
148. Syed YY. Revumenib: First Approval. Drugs 2025; 85: 577-83. doi:10.1007/s40265-025-02161-5 PMID 40072775
149. Tan KW, Zhu YY, Qiu QC, Wang M, Shen HJ, Huang SM, et al. Rapid molecular response to dasatinib in Ph-like acute lymphoblastic leukemia patients with ABL1 rearrangements: case series and literature review. Ann Hematol 2023; 102: 2397–402. doi:10.1007/s00277-023-05236-z PMID 37103615
150. Ferrando AA, Neuberg DS, Staunton J, Loh ML, Huard C, Raimondi SC, et al. Gene expression signatures define novel oncogenic pathways in T cell acute lymphoblastic leukemia. Cancer Cell 2002; 1: 75–87. doi:10.1016/s1535-6108(02)00018-1 PMID 12086890
151. Liu Y, Easton J, Shao Y, Maciaszek J, Wang Z, Wilkinson MR, et al. The genomic landscape of pediatric and young adult T-lineage acute lymphoblastic leukemia. Nat Genet 2017; 49: 1211–8. doi:10.1038/ng.3909 PMID 28671688
152. Weng AP, Ferrando AA, Lee W, Morris JP 4th, Silverman LB, Sanchez-Irizarry C, et al. Activating mutations of NOTCH1 in human T cell acute lymphoblastic leukemia. Science 2004; 306: 269–71. doi:10.1126/science.1102160 PMID 15472075
153. Yuan Y, Li J, Xue TL, Hu HR, Lin W, Liu SG, et al. Prognostic significance of NOTCH1/FBXW7 mutations in pediatric T cell acute lymphoblastic leukemia: a study of minimal residual disease risk-directed CCLG-ALL 2008 treatment protocol. Leuk Lymphoma 2022; 63: 1624–33. doi:10.1080/10428194.2022.2032033 PMID 35129045
154. Trinquand A, Tanguy-Schmidt A, Ben Abdelali R, Lambert J, Beldjord K, Lengliné E, et al. Toward a NOTCH1/FBXW7/RAS/PTEN-based oncogenetic risk classification of adult T-cell acute lymphoblastic leukemia: a Group for Research in Adult Acute Lymphoblastic Leukemia study. J Clin Oncol 2013; 31: 4333–42. doi:10.1200/JCO.2012.48.5292 PMID 24166518
155. Zhang J, Ding L, Holmfeldt L, Wu G, Heatley SL, Payne-Turner D, et al. The genetic basis of early T-cell precursor acute lymphoblastic leukaemia. Nature 2012; 481: 157–63. doi:10.1038/nature10725 PMID 22237106
156. Hamblin TJ, Davis Z, Gardiner A, Oscier DG, Stevenson FK. Unmutated Ig V(H) genes are associated with a more aggressive form of chronic lymphocytic leukemia. Blood 1999; 94: 1848–54. PMID 10477713
157. Damle RN, Wasil T, Fais F, Ghiotto F, Valetto A, Allen SL, et al. Ig V gene mutation status and CD38 expression as novel prognostic indicators in chronic lymphocytic leukemia. Blood 1999; 94: 1840–7. PMID 10477712
158. Stamatopoulos K, Belessi C, Moreno C, Boudjograh M, Guida G, Smilevska T, et al. Over 20% of patients with chronic lymphocytic leukemia carry stereotyped receptors: pathogenetic implications and clinical correlations. Blood 2007; 109: 259–70. doi:10.1182/blood-2006-03-012948 PMID 16985177
159. Döhner H, Stilgenbauer S, Benner A, Leupolt E, Kröber A, Bullinger L, et al. Genomic aberrations and survival in chronic lymphocytic leukemia. N Engl J Med 2000; 343: 1910–6. doi:10.1056/NEJM200012283432602 PMID 11136261
160. Hallek M, Cheson BD, Catovsky D, Caligaris-Cappio F, Dighiero G, Döhner H, et al. iwCLL guidelines for diagnosis, indications for treatment, response assessment, and supportive management of CLL. Blood 2018; 131: 2745–60. doi:10.1182/blood-2017-09-806398 PMID 29540348
161. Eichhorst B, Ghia P, Bosch F, Clifford R, Gregor M, Guieze R, et al. EHA Guidelines on management of chronic lymphocytic leukemia and Richter transformation. Hemasphere 2026; 6: e70403. doi:10.1002/hem3.70403 PMID 42293463
162. Landau DA, Tausch E, Taylor-Weiner AN, Stewart C, Reiter JG, Bahlo J, et al. Mutations driving CLL and their evolution in progression and relapse. Nature 2015; 526: 525–30. doi:10.1038/nature15395 PMID 26466571
163. Puente XS, Beà S, Valdés-Mas R, Villamor N, Gutiérrez-Abril J, Martín-Subero JI, et al. Non-coding recurrent mutations in chronic lymphocytic leukaemia. Nature 2015; 526: 519–24. doi:10.1038/nature14666 PMID 26200345
164. Baliakas P, Jeromin S, Iskas M, Puiggros A, Plevova K, Nguyen-Khac F, et al. Cytogenetic complexity in chronic lymphocytic leukemia: definitions, associations, and clinical impact. Blood 2019; 133: 1205–16. doi:10.1182/blood-2018-09-873083 PMID 30602617
165. International CLL-IPI working group. An international prognostic index for patients with chronic lymphocytic leukaemia (CLL-IPI): a meta-analysis of individual patient data. Lancet Oncol 2016; 17: 779–90. doi:10.1016/S1470-2045(16)30029-8 PMID 27185642
166. Klintman J, Appleby N, Stamatopoulos B, Ridout K, Eyre TA, Robbe P, et al. Genomic and transcriptomic correlates of Richter transformation in chronic lymphocytic leukemia. Blood 2021; 137: 2800–16. doi:10.1182/blood.2020005650 PMID 33206936
167. Woyach JA, Furman RR, Liu TM, Ozer HG, Zapatka M, Ruppert AS, et al. Resistance mechanisms for the Bruton's tyrosine kinase inhibitor ibrutinib. N Engl J Med 2014; 370: 2286–94. doi:10.1056/NEJMoa1400029 PMID 24869598
168. Wang E, Mi X, Thompson MC, Montoya S, Notti RQ, Afaghani J, et al. Mechanisms of resistance to noncovalent Bruton's tyrosine kinase inhibitors. N Engl J Med 2022; 386: 735–43. doi:10.1056/NEJMoa2114110 PMID 35196427
169. Blombery P, Anderson MA, Gong JN, Thijssen R, Birkinshaw RW, Thompson ER, et al. Acquisition of the recurrent Gly101Val mutation in BCL2 confers resistance to venetoclax in patients with progressive chronic lymphocytic leukemia. Cancer Discov 2019; 9: 342–53. doi:10.1158/2159-8290.CD-18-1119 PMID 30514704
170. Rawstron AC, Bennett FL, O'Connor SJ, Kwok M, Fenton JA, Plummer M, et al. Monoclonal B-cell lymphocytosis and chronic lymphocytic leukemia. N Engl J Med 2008; 359: 575–83. doi:10.1056/NEJMoa075290 PMID 18687638
171. Alizadeh AA, Eisen MB, Davis RE, et al. Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling. Nature 2000; 403: 503–11. doi:10.1038/35000501 PMID 10676951
172. Schmitz R, Wright GW, Huang DW, et al. Genetics and pathogenesis of diffuse large B-cell lymphoma. N Engl J Med 2018; 378: 1396–1407. doi:10.1056/NEJMoa1801445 PMID 29641966
173. Chapuy B, Stewart C, Dunford AJ, et al. Molecular subtypes of diffuse large B cell lymphoma are associated with distinct pathogenic mechanisms and outcomes. Nat Med 2018; 24: 679–90. doi:10.1038/s41591-018-0016-8 PMID 29713087
174. Wright GW, Huang DW, Phelan JD, et al. A probabilistic classification tool for genetic subtypes of diffuse large B cell lymphoma with therapeutic implications. Cancer Cell 2020; 37: 551–68. doi:10.1016/j.ccell.2020.03.015 PMID 32289277
175. Pasqualucci L, Dominguez-Sola D, Chiarenza A, et al. Inactivating mutations of acetyltransferase genes in B-cell lymphoma. Nature 2011; 471: 189–95. doi:10.1038/nature09730 PMID 21390126
176. Ennishi D, Jiang A, Boyle M, et al. Double-hit gene expression signature defines a distinct subgroup of germinal centre B-cell-like diffuse large B-cell lymphoma. J Clin Oncol 2019; 37: 190–201. doi:10.1200/JCO.18.01583 PMID 30523716
177. Alaggio R, Amador C, Anagnostopoulos I, et al. The 5th edition of the World Health Organization classification of haematolymphoid tumours: lymphoid neoplasms. Leukemia 2022; 36: 1720–48. doi:10.1038/s41375-022-01620-2 PMID 35732829
178. Campo E, Jaffe ES, Cook JR, et al. The International Consensus Classification of Mature Lymphoid Neoplasms: a report from the Clinical Advisory Committee. Blood 2022; 140: 1229–53. doi:10.1182/blood.2022015851 PMID 35653592
179. Wilson WH, Wright GW, Huang DW, et al. Effect of ibrutinib with R-CHOP chemotherapy in genetic subtypes of DLBCL. Cancer Cell 2021; 39: 1643–53. doi:10.1016/j.ccell.2021.10.006 PMID 34739844
180. Tilly H, Morschhauser F, Sehn LH, et al. Polatuzumab vedotin in previously untreated diffuse large B-cell lymphoma. N Engl J Med 2022; 386: 351–63. doi:10.1056/NEJMoa2115304 PMID 34904799
181. Roulland S, Kelly RS, Morgado E, Sungalee S, Solal-Celigny P, Colombat P, et al. t(14;18) translocation: a predictive blood biomarker for follicular lymphoma. J Clin Oncol 2014; 32: 1347–55. doi:10.1200/JCO.2013.52.8190 PMID 24687831
182. Pastore A, Jurinovic V, Kridel R, et al. Integration of gene mutations in risk prognostication for patients receiving first-line immunochemotherapy for follicular lymphoma: a retrospective analysis of a prospective clinical trial and validation in a population-based registry. Lancet Oncol 2015; 16: 1111–22. doi:10.1016/S1470-2045(15)00169-2 PMID 26256760
183. Morschhauser F, Tilly H, Chaidos A, et al. Tazemetostat for patients with relapsed or refractory follicular lymphoma: an open-label, single-arm, multicentre, phase 2 trial. Lancet Oncol 2020; 21: 1433–42. doi:10.1016/S1470-2045(20)30441-1 PMID 33035457
184. Julia E, Salles G. EZH2 inhibition by tazemetostat: mechanisms of action, safety and efficacy in relapsed/refractory follicular lymphoma. Future Oncol 2021; 17: 2127-40. doi:10.2217/fon-2020-1244 PMID 33709777
185. Salaverria I, Royo C, Carvajal-Cuenca A, et al. CCND2 rearrangements are the most frequent genetic events in cyclin D1(-) mantle cell lymphoma. Blood 2013; 121: 1394–1402. doi:10.1182/blood-2012-08-452284 PMID 23255553
186. Beà S, Valdés-Mas R, Navarro A, et al. Landscape of somatic mutations and clonal evolution in mantle cell lymphoma. Proc Natl Acad Sci U S A 2013; 110: 18250–5. doi:10.1073/pnas.1314608110 PMID 24145436
187. Eskelund CW, Dahl C, Hansen JW, et al. TP53 mutations identify younger mantle cell lymphoma patients who do not benefit from intensive chemoimmunotherapy. Blood 2017; 130: 1903–10. doi:10.1182/blood-2017-04-779736 PMID 28819011
188. Kridel R, Meissner B, Rogic S, et al. Whole transcriptome sequencing reveals recurrent NOTCH1 mutations in mantle cell lymphoma. Blood 2012; 119: 1963–71. doi:10.1182/blood-2011-11-391474 PMID 22210878
189. Fernàndez V, Salamero O, Espinet B, et al. Genomic and gene expression profiling defines indolent forms of mantle cell lymphoma. Cancer Res 2010; 70: 1408–18. doi:10.1158/0008-5472.CAN-09-3419 PMID 20124476
190. Schmitz R, Young RM, Ceribelli M, et al. Burkitt lymphoma pathogenesis and therapeutic targets from structural and functional genomics. Nature 2012; 490: 116–20. doi:10.1038/nature11378 PMID 22885699
191. Love C, Sun Z, Jima D, et al. The genetic landscape of mutations in Burkitt lymphoma. Nat Genet 2012; 44: 1321–5. doi:10.1038/ng.2468 PMID 23143597
192. Grande BM, Gerhard DS, Jiang A, et al. Genome-wide discovery of somatic coding and noncoding mutations in pediatric endemic and sporadic Burkitt lymphoma. Blood 2019; 133: 1313–24. doi:10.1182/blood-2018-09-871418 PMID 30617194
193. Treon SP, Cao Y, Xu L, Yang G, Liu X, Hunter ZR. Somatic mutations in MYD88 and CXCR4 are determinants of clinical presentation and overall survival in Waldenstrom macroglobulinemia. Blood 2014; 123: 2791–6. doi:10.1182/blood-2014-01-550905 PMID 24553177
194. Treon SP, Tripsas CK, Meid K, Warren D, Varma G, Green R, et al. Ibrutinib in previously treated Waldenström's macroglobulinemia. N Engl J Med 2015; 372: 1430–40. doi:10.1056/NEJMoa1501548 PMID 25853747
195. Piva R, Deaglio S, Famà R, et al. The Krüppel-like factor 2 transcription factor gene is recurrently mutated in splenic marginal zone lymphoma. Leukemia 2015; 29: 503–7. doi:10.1038/leu.2014.294 PMID 25283840
196. Rosebeck S, Lucas PC, McAllister-Lucas LM. Protease activity of the API2-MALT1 fusion oncoprotein in MALT lymphoma development and treatment. Future Oncol 2011; 7: 613–7. doi:10.2217/fon.11.35 PMID 21568677
197. Liu H, Ruskoné-Fourmestraux A, Lavergne-Slove A, et al. Resistance of t(11;18) positive gastric mucosa-associated lymphoid tissue lymphoma to Helicobacter pylori eradication therapy. Lancet 2001; 357: 39–40. doi:10.1016/S0140-6736(00)03571-6 PMID 11197361
198. Green MR, Monti S, Rodig SJ, et al. Integrative analysis reveals selective 9p24.1 amplification, increased PD-1 ligand expression, and further induction via JAK2 in nodular sclerosing Hodgkin lymphoma and primary mediastinal large B-cell lymphoma. Blood 2010; 116: 3268–77. doi:10.1182/blood-2010-05-282780 PMID 20628145
199. Roemer MG, Advani RH, Ligon AH, et al. PD-L1 and PD-L2 genetic alterations define classical Hodgkin lymphoma and predict outcome. J Clin Oncol 2016; 34: 2690–7. doi:10.1200/JCO.2016.66.4482 PMID 27069084
200. Roemer MG, Advani RH, Redd RA, et al. Classical Hodgkin lymphoma with reduced β2M/MHC class I expression is associated with inferior outcome independent of 9p24.1 status. Cancer Immunol Res 2016; 4: 910–6. doi:10.1158/2326-6066.CIR-16-0201 PMID 27737878
201. Schmitz R, Hansmann ML, Bohle V, et al. TNFAIP3 (A20) is a tumor suppressor gene in Hodgkin lymphoma and primary mediastinal B cell lymphoma. J Exp Med 2009; 206: 981–9. doi:10.1084/jem.20090528 PMID 19380639
202. Tiacci E, Ladewig E, Schiavoni G, et al. Pervasive mutations of JAK-STAT pathway genes in classical Hodgkin lymphoma. Blood 2018; 131: 2454–65. doi:10.1182/blood-2017-11-814913 PMID 29650799
203. Reichel J, Chadburn A, Rubinstein PG, et al. Flow sorting and exome sequencing reveal the oncogenome of primary Hodgkin and Reed-Sternberg cells. Blood 2015; 125: 1061–72. doi:10.1182/blood-2014-11-610436 PMID 25488972
204. Spina V, Bruscaggin A, Cuccaro A, et al. Circulating tumor DNA reveals genetics, clonal evolution, and residual disease in classical Hodgkin lymphoma. Blood 2018; 131: 2413–25. doi:10.1182/blood-2017-11-812073 PMID 29449275
205. Iqbal J, Wright G, Wang C, et al. Gene expression signatures delineate biological and prognostic subgroups in peripheral T-cell lymphoma. Blood 2014; 123: 2915–23. doi:10.1182/blood-2013-11-536359 PMID 24632715
206. Sakata-Yanagimoto M, Enami T, Yoshida K, et al. Somatic RHOA mutation in angioimmunoblastic T cell lymphoma. Nat Genet 2014; 46: 171–5. doi:10.1038/ng.2872 PMID 24413737
207. Palomero T, Couronné L, Khiabanian H, et al. Recurrent mutations in epigenetic regulators, RHOA and FYN kinase in peripheral T cell lymphomas. Nat Genet 2014; 46: 166–70. doi:10.1038/ng.2873 PMID 24413734
208. Cairns RA, Iqbal J, Lemonnier F, et al. IDH2 mutations are frequent in angioimmunoblastic T-cell lymphoma. Blood 2012; 119: 1901–3. doi:10.1182/blood-2011-11-391748 PMID 22215888
209. Odejide O, Weigert O, Lane AA, Toscano D, Lunning MA, Kopp N, et al. A targeted mutational landscape of angioimmunoblastic T-cell lymphoma. Blood 2014; 123: 1293–6. doi:10.1182/blood-2013-10-531509 PMID 24345752
210. Parrilla Castellar ER, Jaffe ES, Said JW, et al. ALK-negative anaplastic large cell lymphoma is a genetically heterogeneous disease with widely disparate clinical outcomes. Blood 2014; 124: 1473–80. doi:10.1182/blood-2014-04-571091 PMID 24894770
211. Küçük C, Jiang B, Hu X, et al. Activating mutations of STAT5B and STAT3 in lymphomas derived from γδ-T or NK cells. Nat Commun 2015; 6: 6025. doi:10.1038/ncomms7025 PMID 25586472
212. Jiang L, Gu ZH, Yan ZX, et al. Exome sequencing identifies somatic mutations of DDX3X in natural killer/T-cell lymphoma. Nat Genet 2015; 47: 1061–6. doi:10.1038/ng.3358 PMID 26192917
213. Kataoka K, Nagata Y, Kitanaka A, et al. Integrated molecular analysis of adult T cell leukemia/lymphoma. Nat Genet 2015; 47: 1304–15. doi:10.1038/ng.3415 PMID 26437031
214. Mossé YP, Voss SD, Lim MS, Rolland D, Minard CG, Fox E, et al. Targeting ALK with crizotinib in pediatric anaplastic large cell lymphoma and inflammatory myofibroblastic tumor: a Children's Oncology Group study. J Clin Oncol 2017; 35: 3215–21. doi:10.1200/JCO.2017.73.4830 PMID 28787259
215. Merino M, Kasamon Y, Li H, Ma L, Leong R, Zhou J, et al. FDA approval summary: crizotinib for pediatric and young adult patients with relapsed or refractory systemic anaplastic large cell lymphoma. Pediatr Blood Cancer 2022; 69: e29602. doi:10.1002/pbc.29602 PMID 35561013
216. Pro B, Advani R, Brice P, et al. Brentuximab vedotin (SGN-35) in patients with relapsed or refractory systemic anaplastic large-cell lymphoma: results of a phase II study. J Clin Oncol 2012; 30: 2190–6. doi:10.1200/JCO.2011.38.0402 PMID 22614995
217. Pro B, Advani R, Brice P, et al. Five-year results of brentuximab vedotin in patients with relapsed or refractory systemic anaplastic large cell lymphoma. Blood 2017; 130: 2709–17. doi:10.1182/blood-2017-05-780049 PMID 28974506
218. Choi J, Goh G, Walradt T, et al. Genomic landscape of cutaneous T cell lymphoma. Nat Genet 2015; 47: 1011–9. doi:10.1038/ng.3356 PMID 26192916
219. Landgren O, Kyle RA, Pfeiffer RM, Katzmann JA, Caporaso NE, Hayes RB, et al. Monoclonal gammopathy of undetermined significance (MGUS) consistently precedes multiple myeloma: a prospective study. Blood 2009; 113: 5412-7. doi:10.1182/blood-2008-12-194241 PMID 19179464
220. Kuehl WM, Bergsagel PL. Molecular pathogenesis of multiple myeloma and its premalignant precursor. J Clin Invest 2012; 122: 3456-63. doi:10.1172/JCI61188 PMID 23023717
221. Walker BA, Mavrommatis K, Wardell CP, Ashby TC, Bauer M, Davies FE, et al. Identification of novel mutational drivers reveals oncogene dependencies in multiple myeloma. Blood 2018; 132: 587-97. doi:10.1182/blood-2018-03-840132 PMID 29884741
222. D'Agostino M, Cairns DA, Lahuerta JJ, Wester R, Bertsch U, Waage A, et al. Second Revision of the International Staging System (R2-ISS) for Overall Survival in Multiple Myeloma: A European Myeloma Network (EMN) Report Within the HARMONY Project. J Clin Oncol 2022; 40: 3406-18. doi:10.1200/JCO.21.02614 PMID 35605179
223. Walker BA, Boyle EM, Wardell CP, Murison A, Begum DB, Dahir NM, et al. Mutational Spectrum, Copy Number Changes, and Outcome: Results of a Sequencing Study of Patients With Newly Diagnosed Myeloma. J Clin Oncol 2015; 33: 3911-20. doi:10.1200/JCO.2014.59.1503 PMID 26282654
224. Bolli N, Avet-Loiseau H, Wedge DC, Van Loo P, Alexandrov LB, Martincorena I, et al. Heterogeneity of genomic evolution and mutational profiles in multiple myeloma. Nat Commun 2014; 5: 2997. doi:10.1038/ncomms3997 PMID 24429703
225. Palumbo A, Avet-Loiseau H, Oliva S, Lokhorst HM, Goldschmidt H, Rosinol L, et al. Revised International Staging System for Multiple Myeloma: A Report From International Myeloma Working Group. J Clin Oncol 2015; 33: 2863-9. doi:10.1200/JCO.2015.61.2267 PMID 26240224
226. Avet-Loiseau H, Davies FE, Samur MK, Corre J, D'Agostino M, Kaiser MF, et al. International Myeloma Society/International Myeloma Working Group Consensus Recommendations on the Definition of High-Risk Multiple Myeloma. J Clin Oncol 2025; 43: 2739-51. doi:10.1200/JCO-24-01893 PMID 40489728
227. Kumar SK, Harrison SJ, Cavo M, de la Rubia J, Popat R, Gasparetto C, et al. Venetoclax or placebo in combination with bortezomib and dexamethasone in patients with relapsed or refractory multiple myeloma (BELLINI): a randomised, double-blind, multicentre, phase 3 trial. Lancet Oncol 2020; 21: 1630-42. doi:10.1016/S1470-2045(20)30525-8 PMID 33129376
228. Chakraborty R, Bhutani D, Lentzsch S. How do we manage t(11;14) plasma cell disorders with venetoclax? Br J Haematol 2022; 199: 31-9. doi:10.1111/bjh.18243 PMID 35594184
229. Kumar S, Kaufman JL, Gasparetto C, Mikhael J, Vij R, Pegourie B, et al. Efficacy of venetoclax as targeted therapy for relapsed/refractory t(11;14) multiple myeloma. Blood 2017; 130: 2401-9. doi:10.1182/blood-2017-06-788786 PMID 29018077
230. Ding L, Ley TJ, Larson DE, Miller CA, Koboldt DC, Welch JS, et al. Clonal evolution in relapsed acute myeloid leukaemia revealed by whole-genome sequencing. Nature 2012; 481: 506-10. doi:10.1038/nature10738 PMID 22237025
231. Landau DA, Carter SL, Stojanov P, McKenna A, Stevenson K, Lawrence MS, et al. Evolution and impact of subclonal mutations in chronic lymphocytic leukemia. Cell 2013; 152: 714-26. doi:10.1016/j.cell.2013.01.019 PMID 23415222
232. Wong TN, Ramsingh G, Young AL, Miller CA, Touma W, Welch JS, et al. Role of TP53 mutations in the origin and evolution of therapy-related acute myeloid leukaemia. Nature 2015; 518: 552-5. doi:10.1038/nature13968 PMID 25487151
233. Miles LA, Bowman RL, Merlinsky TR, Csete IS, Ooi AT, Durruthy-Durruthy R, et al. Single-cell mutation analysis of clonal evolution in myeloid malignancies. Nature 2020; 587: 477-82. doi:10.1038/s41586-020-2864-x PMID 33116311
234. Morita K, Wang F, Jahn K, Hu T, Tanaka T, Sasaki Y, et al. Clonal evolution of acute myeloid leukemia revealed by high-throughput single-cell genomics. Nat Commun 2020; 11: 5327. doi:10.1038/s41467-020-19119-8 PMID 33087716
235. Johnson JR, Bross P, Cohen M, Rothmann M, Chen G, Zajicek A, et al. Approval summary: imatinib mesylate capsules for treatment of adult patients with newly diagnosed Philadelphia chromosome-positive chronic myelogenous leukemia in chronic phase. Clin Cancer Res 2003; 9: 1972-9. PMID 12796358
236. Kantarjian H, Shah NP, Hochhaus A, Cortes J, Shah S, Ayala M, et al. Dasatinib versus imatinib in newly diagnosed chronic-phase chronic myeloid leukemia. N Engl J Med 2010; 362: 2260-70. doi:10.1056/NEJMoa1002315 PMID 20525995
237. Okimoto RA, Van Etten RA. Navigating the road toward optimal initial therapy for chronic myeloid leukemia. Curr Opin Hematol 2011; 18: 89-97. doi:10.1097/MOH.0b013e32834399a5 PMID 21252655
238. Saglio G, Kim DW, Issaragrisil S, le Coutre P, Etienne G, Lobo C, et al. Nilotinib versus imatinib for newly diagnosed chronic myeloid leukemia. N Engl J Med 2010; 362: 2251-9. doi:10.1056/NEJMoa0912614 PMID 20525993
239. Cortes JE, Gambacorti-Passerini C, Deininger MW, Mauro MJ, Chuah C, Kim DW, et al. Bosutinib Versus Imatinib for Newly Diagnosed Chronic Myeloid Leukemia: Results From the Randomized BFORE Trial. J Clin Oncol 2018; 36: 231-7. doi:10.1200/JCO.2017.74.7162 PMID 29091516
240. Wehrle J, Pahl HL, von Bubnoff N. Ponatinib: a third-generation inhibitor for the treatment of CML. Recent Results Cancer Res 2014; 201: 99-107. doi:10.1007/978-3-642-54490-3_5 PMID 24756787
241. Pamuk GE, Chow ECY, Ionan AC, Chen H, Lee SL, Hsu V, et al. FDA Approval Summary: Asciminib for Ph+ CML in Chronic Phase Treated with Two or More Tyrosine Kinase Inhibitors and for the T315I Mutation. Clin Cancer Res 2024; 30: 4266-71. doi:10.1158/1078-0432.CCR-24-1086 PMID 39088257
242. Lo-Coco F, Avvisati G, Vignetti M, Thiede C, Orlando SM, Iacobelli S, et al. Retinoic acid and arsenic trioxide for acute promyelocytic leukemia. N Engl J Med 2013; 369: 111-21. doi:10.1056/NEJMoa1300874 PMID 23841729
243. Cohen MH, Hirschfeld S, Flamm Honig S, Ibrahim A, Johnson JR, O'Leary JJ, et al. Drug approval summaries: arsenic trioxide, tamoxifen citrate, anastrazole, paclitaxel, bexarotene. Oncologist 2001; 6: 4-11. doi:10.1634/theoncologist.6-1-4 PMID 11161223
244. Kim ES. Midostaurin: First Global Approval. Drugs 2017; 77: 1251-9. doi:10.1007/s40265-017-0779-0 PMID 28612232
245. Pulte ED, Norsworthy KJ, Wang Y, Xu Q, Qosa H, Gudi R, et al. FDA Approval Summary: Gilteritinib for Relapsed or Refractory Acute Myeloid Leukemia with a FLT3 Mutation. Clin Cancer Res 2021; 27: 3515-21. doi:10.1158/1078-0432.CCR-20-4271 PMID 33632926
246. Norsworthy KJ, Luo L, Hsu V, Gudi R, Dorff SE, Przepiorka D, et al. FDA Approval Summary: Ivosidenib for Relapsed or Refractory Acute Myeloid Leukemia with an Isocitrate Dehydrogenase-1 Mutation. Clin Cancer Res 2019; 25: 3205-9. doi:10.1158/1078-0432.CCR-18-3749 PMID 30692099
247. Woods A, Norsworthy KJ, Wang X, Vallejo J, Chiu Yuen Chow E, Li RJ, et al. FDA Approval Summary: Ivosidenib in Combination with Azacitidine for Treatment of Patients with Newly Diagnosed Acute Myeloid Leukemia with an IDH1 Mutation. Clin Cancer Res 2024; 30: 1226-31. doi:10.1158/1078-0432.CCR-23-2234 PMID 38010220
248. Woods AC, Norsworthy KJ, Choe M, Gehrke BJ, Chen H, Vallejo J, et al. FDA Approval Summary: Olutasidenib for Adult Patients with Relapsed or Refractory Acute Myeloid Leukemia with an Isocitrate Dehydrogenase 1 Mutation. Clin Cancer Res 2025; 31: 12-7. doi:10.1158/1078-0432.CCR-24-2196 PMID 39475462
249. Liu X, Gong Y. Isocitrate dehydrogenase inhibitors in acute myeloid leukemia. Biomark Res 2019; 7: 22. doi:10.1186/s40364-019-0173-z PMID 31660152
250. Issa GC, Aldoss I, Thirman MJ, DiPersio J, Arellano M, Blachly JS, et al. Menin Inhibition With Revumenib for KMT2A-Rearranged Relapsed or Refractory Acute Leukemia (AUGMENT-101). J Clin Oncol 2025; 43: 75-84. doi:10.1200/JCO.24.00826 PMID 39121437
251. US Food and Drug Administration. FDA approves revumenib for relapsed or refractory acute leukemia with a KMT2A translocation. 15 November 2024. https://www.fda.gov/drugs/resources-information-approved-drugs/fda-approves-revumenib-relapsed-or-refractory-acute-leukemia-kmt2a-translocation
252. US Food and Drug Administration. FDA approves revumenib for relapsed or refractory acute myeloid leukemia with a susceptible NPM1 mutation. 24 October 2025. https://www.fda.gov/drugs/resources-information-approved-drugs/fda-approves-revumenib-relapsed-or-refractory-acute-myeloid-leukemia-susceptible-npm1-mutation
253. US Food and Drug Administration. Drug Trials Snapshots: KOMZIFTI (ziftomenib). Approved 13 November 2025 for adult patients with relapsed or refractory acute myeloid leukemia with a susceptible NPM1 mutation and no satisfactory alternative treatment. https://www.fda.gov/drugs/drug-trials-snapshots/drug-trials-snapshots-komzifti
254. Gotlib J, Reiter A, DeAngelo DJ. Avapritinib for advanced systemic mastocytosis. Blood 2022; 140: 1667-73. doi:10.1182/blood.2021014612 PMID 35877999
255. Seymour JF, Kipps TJ, Eichhorst B, Hillmen P, D'Rozario J, Assouline S, et al. Venetoclax-Rituximab in Relapsed or Refractory Chronic Lymphocytic Leukemia. N Engl J Med 2018; 378: 1107-20. doi:10.1056/NEJMoa1713976 PMID 29562156
256. Held L, Siu C, Shadman M. Venetoclax as a therapeutic option for the treatment of chronic lymphocytic leukemia: the evidence so far. Expert Opin Pharmacother 2021; 22: 655-65. doi:10.1080/14656566.2021.1882421 PMID 33645381
257. DiNardo CD, Jonas BA, Pullarkat V, Thirman MJ, Garcia JS, Wei AH, et al. Azacitidine and Venetoclax in Previously Untreated Acute Myeloid Leukemia. N Engl J Med 2020; 383: 617-29. doi:10.1056/NEJMoa2012971 PMID 32786187
258. Pollyea DA. Venetoclax in AML: Where We Are and Where We Are Headed. Clin Lymphoma Myeloma Leuk 2020; 20 (suppl 1): S25-6. doi:10.1016/S2152-2650(20)30450-X PMID 32862856
259. Platzbecker U, Della Porta MG, Santini V, Zeidan AM, Komrokji RS, Shortt J, et al. Efficacy and safety of luspatercept versus epoetin alfa in erythropoiesis-stimulating agent-naive, transfusion-dependent, lower-risk myelodysplastic syndromes (COMMANDS): interim analysis of a phase 3, open-label, randomised controlled trial. Lancet 2023; 402: 373-85. doi:10.1016/S0140-6736(23)00874-7 PMID 37311468
260. Chang C, Suzuki T, Liang Y, Tong H, Usuki K, Liu Q, et al. Safety and efficacy of luspatercept in treating anemia associated with myelodysplastic syndrome with ring sideroblasts in Asian patients who require red blood cell transfusions: a phase II bridging study. Ther Adv Hematol 2025; 16: 20406207251321715. doi:10.1177/20406207251321715 PMID 39991012
261. Patel N, Calvo KR. How I diagnose myeloid neoplasms with germline predisposition. Am J Clin Pathol 2023; 160: 352-64. doi:10.1093/ajcp/aqad075 PMID 37458302