Iadademstat

Iadademstat is Oryzon Genomics' oral, selective inhibitor of LSD1, an enzyme that leukemic cells depend on to stay in an immature, self-renewing state instead of maturing into normal blood cells. The lead program combines it with azacitidine in newly diagnosed AML patients too frail for intensive chemotherapy, where the Phase 2a ALICE trial reported an 81% overall response rate, 52% complete responses, and a median overall survival of approximately 9.3 months [1]. That response signal in unfit AML is the strongest clinical data any LSD1 inhibitor has produced in oncology, a class otherwise defined by mechanism-driven thrombocytopenia and stalled programs: GSK's GSK2879552 was formally stopped for futility in SCLC, and BMS/Celgene's pulrodemstat (CC-90011) was terminated in April 2025 for 'changed business objectives' despite some durable responses in neuroendocrine tumors [10]. Incyte's INCB059872 has no active trials on ClinicalTrials.gov and appears deprioritized. Oryzon holds FDA orphan drug and fast track designations for AML but has not yet started a registrational trial, and the company's ability to finance one is the central uncertainty for this asset. A parallel push into small cell lung cancer is running through academic Phase 1 combinations with checkpoint inhibitors and radiation, sponsored by NCI [2] and Yale [3] rather than by Oryzon itself. The commercial question is whether an 81% response rate in a single-arm Phase 2a can survive a randomized trial against azacitidine plus venetoclax (VIALE-A regimen: ~66% CR/CRi, 14.7-month median OS), the current standard for this exact patient population.

Iadademstat has never been approved anywhere. Oryzon holds FDA orphan drug designation for AML and, per company disclosures, fast track designation as well. The compound sits at Phase 2, anchored by the ALICE Phase 2a study in newly diagnosed AML patients unfit for standard 7+3 induction chemotherapy [1]. Oryzon has repeatedly guided toward launching a registrational AML trial but has yet to enroll one; the delay stems from both financing constraints at a small Spanish biotech and the shifting AML competitive picture, particularly the 2024 approval of revumenib for KMT2A-rearranged leukemia [4]. As of the FY2025 report, Oryzon held €28.4M in gross cash (net cash position €16.5M after €11.8M debt), with company guidance of operational runway through H1 2027 [11]. FY25 free cash outflow was €13.4M. Ongoing combination trials in SCLC (NCT06287775 through NCI [2], NCT07113691 at Yale with SBRT and atezolizumab [3]) and AML (NCT06357182 at OHSU adding venetoclax to iadademstat plus azacitidine [5]) are all Phase 1, sponsored by academic centers rather than by Oryzon. There is also a Phase 1 MDS study at Medical College of Wisconsin (NCT06502145) [8]. Next meaningful readout: mature ALICE follow-up and any registrational trial announcement from Oryzon. Oryzon has also updated the AML story with an internal ALICE-2 triplet dataset (iadademstat + azacitidine + venetoclax) showing 89% composite CR and 79% estimated 12-month OS at 8-month median follow-up, though this remains a single-arm early-phase readout [12].

LSD1, also called KDM1A, is an enzyme that erases specific chemical tags from histones, the proteins that DNA wraps around. Think of histones as spools with sticky notes attached; LSD1 rips off certain notes that would otherwise tell the cell to differentiate. In acute myeloid leukemia, especially cases driven by MLL/KMT2A gene fusions, the malignant cells hijack LSD1 to stay locked in a stem-cell-like state where they keep dividing instead of maturing into normal blood cells. Blocking LSD1 releases that block, and the leukemic cells start differentiating and dying [6]. The genetic case is solid: LSD1 knockdown differentiates MLL-fusion AML cells in mouse models, and iadademstat is the tightest-binding, most selective LSD1 inhibitor to reach the clinic [7]. The commercial case is less settled. Every LSD1 inhibitor tested so far has hit the same wall: on-target thrombocytopenia (low platelet counts) because normal blood cell production also uses LSD1. GSK's GSK2879552 was killed in SCLC after futility [6]. Imago's bomedemstat (now Merck) is still in play in myeloproliferative disease but is a different chemical scaffold. Iadademstat's Phase 2a signal in AML plus azacitidine, with manageable cytopenias in a population already prone to them, is the best mechanism validation the target has produced in humans so far [1].

The Phase 2a ALICE trial (NCT03895684) enrolled 36 newly diagnosed AML patients (median age 76) ineligible for intensive chemotherapy, dosing iadademstat on top of standard azacitidine [1]. For context, standard intensive induction in fit AML is the '7+3' regimen (7 days of cytarabine plus 3 days of an anthracycline like daunorubicin), which older or comorbid patients cannot safely receive; azacitidine, a hypomethylating agent given as a lower-intensity alternative, is what these 'unfit' patients are typically offered. Primary endpoints were safety and objective response rate. Published results in Lancet Haematology 2024 showed an 81% overall response rate with 52% CR/CRh in the efficacy-evaluable set, and a median overall survival of approximately 9.3 months across all patients (roughly 14.3 months in the CR/CRi subset) [1]. For comparison, azacitidine monotherapy in unfit AML delivers ~10-month median OS, and azacitidine + venetoclax (VIALE-A) delivers 14.7 months. Median duration of response is described as durable but a specific single figure was not clearly reported in the published paper; Oryzon has publicly characterized durations as 'above one year' but readers should treat this as approximate. Thrombocytopenia was the main adverse event, expected from the mechanism and manageable with dose modifications in most patients. The trial's weakness is what it is: single-arm, small, open-label. Crucially, the Lancet Haematology publication did not stratify response rates by cytogenetic or molecular subgroup (KMT2A/MLL fusion, IDH1/2, TP53, FLT3) in a way that would tell you whether the signal is driven by the ~15% of AML patients with KMT2A rearrangements (the mechanistic sweet spot for LSD1 inhibition) or is distributed across genotypes. This is a decision-critical design gap: if responses concentrate in KMT2A-rearranged disease, a registrational trial almost certainly needs biomarker enrichment. The de facto comparator remains azacitidine plus venetoclax; ALICE cannot tell you whether iadademstat plus azacitidine beats or matches it, and no head-to-head trial exists. A future registrational design almost certainly needs to be either a triplet against azacitidine plus venetoclax, or a randomized study in a niche where venetoclax underperforms, such as TP53-mutant AML. The current active trial list is academic and early: NCT07113691 (Yale, iadademstat + SBRT + atezolizumab in ES-SCLC, n=15, DLT endpoint) [3], NCT06287775 (NCI, iadademstat + atezolizumab or durvalumab in SCLC, n=45) [2], NCT06357182 (OHSU, iadademstat + azacitidine + venetoclax in AML, n=30) [5], NCT06502145 (MCW, iadademstat + hypomethylating agent in MDS, n=12) [8]. All Phase 1, none sponsored by Oryzon.

Our model estimates a 5% chance this drug is eventually approved. It starts from the historical base rate for Phase 2 drugs in this area (about 13%), then adjusts using ten facts about the trial and sponsor. What moves the number most: it is helped by a non-randomized design; it is held back by the sponsor's thin or weak approval record, smaller-than-typical enrollment for this phase, and weak or limited earlier-phase results. The other facts land near average for this stage, so they leave the estimate roughly where the base rate put it.

Efficacy risk: the ALICE signal came from a single-arm Phase 2a trial [1]. Cross-trial comparisons against azacitidine plus venetoclax are treacherous because unfit-AML populations vary in cytogenetic risk, TP53 status, and performance status. Median OS of ~9.3 months in ALICE is not obviously superior to azacitidine + venetoclax (14.7 months in VIALE-A), though populations may differ; without a randomized comparison this is unresolvable. Biomarker selection is essentially absent: ALICE was molecularly unselected and did not publish response rates stratified by KMT2A, IDH1/2, TP53, or FLT3 status, so there is no validated way to enrich a registrational trial. In SCLC, the LSD1 mechanism has failed once already at GSK, and academic Phase 1 combinations with checkpoint inhibitors are exploratory rather than confirmatory [6]. Safety risk: on-target thrombocytopenia is a known feature of every LSD1 inhibitor tested in humans [9]. In unfit AML, patients arrive with low platelets already, so the therapeutic window is narrow. Dose modifications helped in ALICE but constrain how aggressively the drug can be pushed in combinations, particularly with venetoclax which independently suppresses counts. Execution risk: Oryzon is a Spanish micro-cap with €28.4M gross cash and guided runway through H1 2027 at a ~€13.4M annual burn rate [11]. Every registrational program the company has floated over the past five years has been delayed. Without a large-pharma partner, financing a global Phase 3 in AML is an open question, and the equity-financing environment for single-asset small biotechs remains punishing. Commercial risk: even if approved, iadademstat plus azacitidine would slot behind azacitidine plus venetoclax unless data are clearly better on overall survival or in a defined subgroup. Payers will not pay for a branded triplet without a demonstrated survival win. The AML competitive field has densified with menin inhibitors (revumenib approved, ziftomenib pending) [4], IDH inhibitors, and expanded venetoclax combinations.

Worth watching, but with skepticism proportional to how long Oryzon has been guiding toward a registrational AML trial without starting one. The ALICE response data are the real signal: an 81% ORR in unfit AML is not noise, and the mechanism has a coherent biological story in MLL-fusion and hypomethylating-refractory disease [1][6] (hypomethylating-refractory = AML that has stopped responding to azacitidine/decitabine, the standard low-intensity option). The 9.3-month median OS is more equivocal - it beats azacitidine monotherapy but does not obviously beat aza+venetoclax. The question is whether Oryzon can translate an academic-scale Phase 2a into a registrational program before the AML field moves past them. Specific data points that would change the read: (1) Oryzon announcing a partnered or funded Phase 3 with a specified control arm and biomarker-defined population (KMT2A enrichment would be the mechanistically obvious play), (2) OHSU triplet (iadademstat + azacitidine + venetoclax) showing tolerable safety and preserved response depth (NCT06357182) [5] - Oryzon's own ALICE-2 readout [12] hints at 89% composite CR but is still single-arm early-phase, (3) any credible SCLC combination signal from the NCI or Yale trials [2][3], though these are Phase 1 and unlikely to move commercial value in the near term, (4) any published mutation-stratified reanalysis of ALICE that identifies a responder subgroup. Check back in 6-12 months for updated ALICE follow-up and any Oryzon corporate development activity. The stock (ORY.MC) and cash position are the leading indicators for whether the Phase 3 actually happens. If Oryzon runs out of runway before starting a registrational trial, iadademstat becomes an asset sale story rather than an approval story, and the LSD1 class remains stuck at 'promising Phase 2 signal, no approved drugs' for another cycle.