Myelodysplastic syndromes (MDS)

Overview

• Definition: clonal hematopoietic stem cell disorder characterized by ineffective hematopoiesis, peripheral cytopenias, and risk of progression to AML.
• Epidemiology: incidence 4 to 5 per 100,000/yr overall; 30 or more per 100,000 in age >70. Median age at dx ~70 yrs. Therapy-related MDS younger.
• Pathogenesis: stepwise acquisition of mutations in hematopoietic stem/progenitor cells (generally 3 to 4 driver mutations) leads to ineffective hematopoiesis (cytopenias despite hypercellular marrow) and proliferative advantage of the clone.

Diagnosis

• WHO minimal diagnostic criteria: meaningful cytopenia(s) with reactive causes ruled out, PLUS at least one of: increased blasts (5 to 19%), extensive dysplasia (≥10% in a lineage), or karyotypic evidence of clonality (excluding non-specific +8, -Y, del(20q), possibly +15).
• If cytopenia is present but none of the qualifying features: the case may be termed Idiopathic Cytopenias of Undetermined Significance (ICUS), i.e. cytopenias with a negative workup.

• Other causes of cytopenia and morphologic change that must be excluded: vitamin B12/folate deficiency, HIV or other viral infection, copper deficiency; alcohol abuse; medications (esp. methotrexate, azathioprine, recent chemotherapy); autoimmune conditions (ITP, Felty syndrome, SLE); congenital syndromes (Fanconi anemia); other hematologic disorders (aplastic anemia, LGL disorders, MPN).

Precursor states (differential diagnosis)

• ICUS (Idiopathic Cytopenias of Undetermined Significance): meaningful cytopenias, does NOT meet minimal MDS criteria (no clonality, <10% dysplasia, <5% blasts, no MDS-specific karyotype), other causes ruled out. Very low chance of progression to a myeloid neoplasm.
• CHIP (Clonal Hematopoiesis of Indeterminate Potential): qualifying clonal mutation (VAF at least 2%) without cytopenia and not meeting criteria for a hematologic neoplasm; physiologic marrow blasts and subdiagnostic dysplasia are compatible. A clonal mutation associated with hematologic neoplasia without meeting diagnostic criteria for any neoplasm. Associated with increased all-cause mortality via (1) progression to myeloid malignancy at ~0.5 to 1% per year and (2) cardiovascular disease (proinflammatory state increasing atherosclerosis).
• CCUS (Clonal Cytopenias of Undetermined Significance): meets ICUS criteria AND has clonality (a mutation in an MDS/leukemia-associated gene, e.g. DNMT3A, TET2, ASXL1, SRSF2 or other splicing gene). Spliceosome mutations plus co-mutation of TET2/DNMT3A/ASXL1 carry the highest predictive value for a myeloid neoplasm. Key point: Most isolated myeloid mutations without morphologic dysplasia (even with cytopenia) are NOT diagnostic of MDS (= CCUS). Exceptions per ICC 2022: MDS-SF3B1 (SF3B1 VAF ≥10%) and MDS with multi-hit TP53 are genetically defined and do not require diagnostic dysplasia (cytopenia still required); other isolated mutations without dysplasia are monitored as CHIP or CCUS.

Etiologies

• Primary (idiopathic) MDS: age-related; CHIP is a precursor.
• Therapy-related MDS (t-MDS):
  • Alkylating agent-related (5 to 7 yr latency): complex karyotype, -5/del(5q), -7/del(7q); poor prognosis. Agents: melphalan, cyclophosphamide, nitrogen mustard, chlorambucil, busulfan, carboplatin, cisplatin, dacarbazine, bendamustine.
  • Topoisomerase II inhibitor-related (etoposide) (1 to 3 yr latency): 11q23 (KMT2A/MLL); typically AML rather than MDS. Agents: etoposide, doxorubicin, mitoxantrone.
  • PARP inhibitor-related (emerging): increased rates of t-MDS/AML especially with prolonged exposure (olaparib for ovarian/breast/prostate).
• Inherited predisposition: IBMFS (Fanconi, dyskeratosis congenita, GATA2, Shwachman-Diamond, RUNX1, ANKRD26, DDX41 with adult onset).
• Environmental: benzene, ionizing radiation.
• Other: prior chemotherapy, autologous SCT.

Clinical features

• Cytopenias: anemia (most common; macrocytic), neutropenia, thrombocytopenia. Single or multilineage.
• Symptoms: fatigue, dyspnea (anemia); infections (neutropenia); bleeding/petechiae (thrombocytopenia).
• Splenomegaly uncommon (suggests CMML or MDS/MPN overlap).
• Smear: dysplastic cells: Pelger-Huet cells (bilobed neutrophils), hypogranular neutrophils, pseudo-Pelger-Huet, oval macrocytes, basophilic stippling, blasts (variable).
• Bone marrow: typically hypercellular (with peripheral cytopenias = ineffective hematopoiesis), dysplasia ≥10% in ≥1 lineage, ring sideroblasts in some, blasts (% defines subtype).

Mutations associated with MDS

• Mnemonic (EBIgEneTic = epigenetic regulators): EZH2, BCOR, IDH1/2, DNMT3A.
• Functional classes: RNA splicing (SF3B1, SRSF2, U2AF1, ZRSR2); cohesin (STAG2, RAD21, SMC3); transcription factors (RUNX1, ETV6); DNA damage (TP53); tyrosine-kinase signaling (JAK2, NRAS, KRAS).

Classification (WHO 2022 / ICC 2022)

• Modern frameworks emphasize molecular features. Both classifications updated 2022. The WHO scheme is useful diagnostically but has limited prognostic value; risk-scoring tools are used for prognosis.

• MDS with defining genetic abnormality: (1) MDS w/ low blasts and isolated 5q deletion (MDS-5q); (2) MDS w/ low blasts and SF3B1 mutation (MDS-SF3B1), favorable; (3) MDS w/ biallelic TP53 inactivation (MDS-bi-TP53), very poor prognosis (replaces the older "TP53-mutated" concept).
• MDS, morphologically defined: (1) MDS w/ low blasts (MDS-LB); (2) MDS, hypoplastic (MDS-h), NEW WHO 2022 category, ≤25% cellularity age-adjusted; (3) MDS w/ increased blasts: IB1 (5 to 9% BM / 2 to 4% PB), IB2 (10 to 19% BM / 5 to 19% PB or Auer rods); (4) MDS w/ fibrosis (MDS-f), 5 to 19% blasts plus significant fibrosis (worse).
• Key WHO 2022 / ICC 2022 changes:
  • AML vs MDS-EB blast threshold: WHO 2022 retains 20% blasts; ICC 2022 lowers to ≥10% with defining genetic features and creates a new MDS/AML category for 10 to 19% blasts. In WHO 2022 these genetically defined AML entities need no minimum blast count. ICC 2022 requires ≥10% blasts (10 to 19% blasts with an AML-defining lesion is still AML; MDS/AML is for 10 to 19% without one), except APL with PML::RARA, which is classified regardless of blast count.
  • Biallelic TP53 is its own MDS category (extremely poor prognosis).
  • SF3B1 alone now defines a better-prognosis category.
  • CMML is classified separately as MDS/MPN overlap (not MDS); defined by ≥0.5 × 10⁹/L blood monocytes and ≥10% of WBC (lowered from 1 × 10⁹/L in WHO 2022/ICC 2022). MDS/MPN-RS-T requires platelets ≥450 × 10⁹/L.

Risk stratification

• IPSS (1997): cytogenetics + blast % + cytopenias.
• IPSS-R (2012): 5 risk groups with refined cytogenetics (see table).
• IPSS-M (2022): adds 31 mutations to refine risk further. Now standard in clinical practice.

• Cytogenetic risk (per IPSS-R):
  • Very good: -Y, del(11q).
  • Good (most common): normal karyotype, del(5q), del(20q), del(12p) alone.
  • Intermediate: +8, del(7q) alone, i(17q), other single or double abnormalities.
  • Poor: -7 alone, complex (3 abnormalities).
  • Very poor: complex karyotype (>3 abnormalities).
• High-yield molecular markers:
  • TP53 (especially biallelic): very poor prognosis; allo-HSCT urgent (though benefit limited).
  • SF3B1: better prognosis, esp. without TP53; luspatercept very effective in MDS-RS.
  • SRSF2, U2AF1, ASXL1, EZH2, RUNX1: adverse.
  • IDH1, IDH2: may be targeted with IDH inhibitors.
  • NPM1: WHO 2022 classifies as AML at any blast count; ICC 2022 requires ≥10% blasts (FLT3 is not AML-defining).

Treatment: lower-risk MDS

• Asymptomatic: observation.
• Erythropoiesis-stimulating agents (ESAs): epoetin alfa 40,000 to 60,000 U/week or darbepoetin 500 mcg q3 weeks. ORR ~15 to 40% in low-risk MDS; no prospective survival benefit; off-label. Best response with low endogenous EPO (<500 mIU/mL) and low transfusion need (<2 U/8 weeks). CMS requires Hb <10 g/dL and iron-replete state for initiation. G-CSF may augment ESA response, particularly in MDS-RS.
• Lenalidomide:
  • del(5q) is the strongest genetic predictor of response in MDS; ORR ~60% with lenalidomide 10 mg (response duration ~2.2 yrs). ORR ~25% in low-risk MDS without del(5q) (median duration of transfusion independence ~30.9 weeks, MDS-005). FDA 2005.
  • Mechanism: binds/modulates the CRL4-CRBN (cereblon) E3 ubiquitin ligase. In del(5q) MDS it induces ubiquitination and degradation of CSNK1A1 (CK1α); haploinsufficiency for CSNK1A1 in del(5q) cells sensitizes them to the drug. (In myeloma it degrades IKZF1/Ikaros and IKZF3/Aiolos.)
  • Caveat: del(5q) has a high rate of TP53 mutation (esp. complex karyotype); TP53 mutation predicts shorter OS and relapse even in del(5q).
  • Trials: MDS-003 (List AF, NEJM 2006), MDS-004 (Fenaux P, Blood 2011).
• Luspatercept (Reblozyl): TGF-beta superfamily ligand trap (binds TGF-beta ligands and inhibits SMAD2/3 signaling in late-stage erythropoiesis), reduces ineffective erythropoiesis. MEDALIST trial (List AF, NEJM 2020): 1.0 mg/kg q21d improved RBC transfusion independence (≥8 wk) 38% vs 13% placebo in MDS-RS. at least 12-wk RBC transfusion independence (wk 1 to 24) ~28% vs 8% placebo; median longest transfusion-independent response ~30.6 weeks. FDA Apr 2020 for MDS-RS after ESA failure; expanded Aug 2023 to first-line lower-risk MDS (COMMANDS). Now a first-line SOC option for lower-risk anemia.
• Imetelstat (Rytelo): first-in-class telomerase inhibitor (oligonucleotide). FDA Jun 2024 for lower-risk MDS with transfusion-dependent anemia (≥4 RBC units/8 wk) refractory or ineligible for ESAs (IMerge phase 3, Lancet 2024). 8-wk transfusion independence ~40% vs 15% placebo. AEs: thrombocytopenia, neutropenia.
• Thrombopoiesis-stimulating agents: romiplostim (1 to 10 mcg/kg/week; best studied) and eltrombopag (50 to 150 mg/d) reduce platelet transfusions and bleeding but may increase AML progression; off-label. Do NOT give romiplostim to patients with excess blasts.
• Myeloid growth factors (G-CSF): never shown to improve survival (nor to worsen AML progression); may reduce infections/hospitalizations in severe neutropenia with recurrent infection. Pegfilgrastim risks: leukemoid reaction, splenic rupture.
• Immunosuppressive therapy (ATG +/- cyclosporine): ORR ~30% (Komrokji JCO 2011) to ~50% (Stahl, Blood Adv 2018), DOR ~1 yr. ATG given 4 days inpatient, then steroids to prevent serum sickness, with long-term low-dose cyclosporine. Best in selected hypoplastic MDS, younger patients, HLA-DR15+.
• Iron chelation for transfusion dependence: deferasirox (oral, preferred for ease) or deferoxamine (overnight infusion). Consider if ferritin >1000 ng/mL or other evidence of iron overload (liver damage, CHF), and think about it after 20 to 30 units of RBCs, especially in lower-risk MDS or transplant candidates. For ferritin >2500, aim to reduce to <1000. Do NOT give deferasirox if CrCl <40 mL/min (deferoxamine: avoid in severe renal disease or anuria).

Treatment: higher-risk MDS

• Hypomethylating agents (HMA):
  • Azacitidine (Vidaza): 75 mg/m² SC or IV daily × 7 days q4 weeks. AZA-001 (Fenaux, Lancet Oncol 2009): CR+PR 29% vs 12% (CR 17%), HI 49% vs 29% in higher-risk, 2-yr OS 50.8% vs 26.2%, median OS 24.5 vs 15.0 mo vs conventional care.
  • Decitabine: 20 mg/m² IV daily × 5 days q4 weeks. IV formulation did not improve median OS in Lubbert JCO 2011; remains a reasonable option despite weaker evidence.
  • Decitabine + cedazuridine (Inqovi): oral fixed combination (35 mg/100 mg PO daily × 5 d/cycle); FDA 2020; bioequivalent to IV decitabine (Garcia-Manero, Blood 2020, overall clinical response 60% (CR 21%); no oral-vs-IV OS comparison (PK crossover design)). Useful for outpatient HMA delivery.
  • Response: median time to response ~2 to 3 cycles; CR/PR rates ~25 to 50%; OS extension limited without transplant.
• Allo-HSCT: the only curative therapy; upfront for fit higher-risk (IPSS-R/M intermediate-high to very high) patients. usually deferred in lower-risk (delayed transplant maximizes survival for most), but consider earlier in selected lower-risk patients with adverse molecular features, severe cytopenias, high transfusion burden, or failure of non-transplant therapy, but can offer at progression. HMA can bridge to transplant during donor search.
• Reduced-intensity conditioning (RIC) has expanded eligibility; 3-yr OS ~30 to 50% depending on age and risk.
• Venetoclax + HMA: VERONA phase 3 (2025) showed no OS benefit over azacitidine in higher-risk MDS; not standard (trial only).
• IDH1-mut (ivosidenib): FDA expanded for MDS Oct 2023; IDH2-mut (enasidenib) off-label.
• TP53-mutated MDS: early allo-HSCT; eprenetapopt (APR-246) phase 3 negative. Sabatolimab (anti-TIM-3): phase 3 STIMULUS-MDS2 negative, development discontinued.

Special situations

• Hypoplastic MDS: overlap with aplastic anemia; HLA-DR2/DR15 association. IST (horse ATG + cyclosporine +/- eltrombopag) effective in selected patients. Predictors of IST response: age <60, ≤5% marrow blasts, hypocellular marrow, HLA-DR15+, PNH clone+, or STAT3-mutant cytotoxic T-cell clones.
• Therapy-related MDS: worse prognosis; often complex / -7 / -5q; allo-HSCT preferred if fit.
• Germline predisposition (DDX41, RUNX1, GATA2): family screening; avoid a germline-affected sibling as donor.
• Myeloid/lymphoid neoplasm with eosinophilia and PDGFRB rearrangement (5q32; not MDS): t(5;12) ETV6::PDGFRB activates the PDGFRB receptor tyrosine kinase; presents with leukocytosis and eosinophilia; imatinib effective (case series normalized counts by 4 weeks and cleared t(5;12) by 9 months).
• CMML: now classified separately as MDS/MPN overlap (monocytosis ≥10% of WBC and ≥0.5 × 10⁹/L monocytes per WHO 2022/ICC 2022; BCR-ABL1 negative). Managed elsewhere; treatment includes HMA and hydroxyurea for proliferative disease. Frequent mutations: TET2, SRSF2, ASXL1, RUNX1, NRAS, CBL.

Recent FDA approvals: MDS (2023 to 2026)

• Luspatercept (Reblozyl), first-line lower-risk MDS, Aug 28, 2023: COMMANDS (Platzbecker U, Lancet 2023; primary analysis Lancet Haematol 2024): 1.0 mg/kg SC q3wk (titrate to 1.75 mg/kg); 12-wk transfusion independence plus Hb increase 59% vs 31% epoetin alfa in ESA-naive, transfusion-dependent (2 to 6 RBC units/8 wk) lower-risk MDS with EPO <500 U/L, both RS-positive and RS-negative (not just RS+). Label is independent of RS status, but benefit over epoetin was clearest in RS-positive/SF3B1-mutant disease (RS-negative subgroup not significant; ESA or luspatercept both options).
• Ivosidenib (Tibsovo), R/R IDH1-mutant MDS, Oct 24, 2023: AG120-C-001 (Foran J, Blood 2024); CR rate 39%, transfusion independence 67%; 500 mg PO daily. Class warning: differentiation syndrome.
• Imetelstat (Rytelo), lower-risk transfusion-dependent MDS post-ESA, Jun 6, 2024: first-in-class telomerase inhibitor; IMerge phase 3 (Platzbecker U, Lancet 2024): 8-wk RBC TI 40% vs 15% placebo; 7.1 mg/kg IV q4wk. Class AEs: thrombocytopenia, neutropenia.
• Decitabine + cedazuridine (Inqovi), fully approved 2020: oral fixed combo bioequivalent to IV decitabine; useful for outpatient HMA delivery.
• Status of failed/withdrawn agents: eprenetapopt (APR-246) for TP53-mut MDS, phase 3 negative (2021); magrolimab (anti-CD47): ENHANCE (higher-risk MDS) stopped for futility Jul 2023, ENHANCE-2 (TP53-mut AML) Sep 2023; ENHANCE-3 (AML) halted Feb 2024 for increased mortality with an FDA full clinical hold; hematology development discontinued; pevonedistat + aza, phase 3 negative.

High yield (MDS)

• IPSS-M (2022) is the modern risk stratification (incorporates 31 mutations).
• Multi-hit/biallelic TP53 = very poor prognosis; consider early allo-HSCT if fit. Monoallelic TP53 alone behaves like TP53-WT; interpret by allelic state, VAF, karyotype, and blast burden.
• del(5q) to lenalidomide.
• SF3B1-mutated MDS-RS to luspatercept (preferred first-line per COMMANDS).
• Imetelstat (Jun 2024): new option for lower-risk MDS with TD anemia post-ESA.
• HMA (azacitidine) for higher-risk; allo-HSCT only curative.
• CMML is MDS/MPN overlap, separately classified.
• Therapy-related MDS: complex / -7 / -5q; alkylator ~5 to 7 yrs latency; topo II ~1 to 3 yrs.
• IDH1 mutation to ivosidenib.
• Vidaza, Inqovi, Reblozyl, Rytelo: modern MDS armamentarium.
• Germline predisposition (DDX41, RUNX1, GATA2) increasingly recognized: family screening plus donor avoidance.