Emavusertib

Emavusertib is Curis's investigational oral IRAK4 inhibitor, paired with BeiGene's zanubrutinib (Brukinsa, a BTK blocker) in a Phase 2 chronic lymphocytic leukemia trial (NCT07271667, also called TakeAim CLL / CA-4948-203) that is just starting - Curis guides dosing of the first 5 patients by mid-2026 and initial data in December 2026 [1][9]. The primary endpoint for Cohort 1 is undetectable measurable residual disease (uMRD) rate, a deep-response readout that predicts longer remissions in CLL and increasingly drives label decisions. Target enrollment is 108 patients [1]. Important trial design nuance the original framing missed: the study enrolls patients already on BTKi monotherapy who have achieved partial remission but not complete remission or uMRD, then adds emavusertib on top to deepen response [1][9]. It is a salvage-deepening design, not a frontline head-to-head. The strategic logic: BTK inhibitors are the backbone of CLL treatment but eventually fail, and the failure modes often involve alternative NF-kB activation. IRAK4 sits in a parallel arm of the same survival signaling network, particularly in B-cell malignancies with MYD88 mutations, so blocking both kinases at once is a textbook combination hypothesis [2]. For Curis (CRIS, micro-cap), this is the program that has to work. The company's AML and MDS programs hit FDA partial clinical holds in 2022 over rhabdomyolysis cases and have struggled since [6]. The CLL combo, run with a BeiGene drug rather than the ibrutinib it tested in PCNSL, is the cleaner safety setup and the most commercially meaningful pivot left for the asset.

Novel small molecule. Emavusertib has never been approved anywhere [1]. It holds FDA Orphan Drug Designation in AML from prior development; no public CLL-specific designations have been announced [6]. The current CLL combo program (NCT07271667) is Phase 2 with the first 5 patients expected to be dosed by mid-2026 and initial data in December 2026 [9]. Three other active emavusertib trials are running in parallel: NCT03328078 (Phase 1 in relapsed/refractory primary CNS lymphoma, ibrutinib combo, Curis-sponsored) [3], NCT06439836 (Phase 1 with pembrolizumab in immunotherapy-refractory urothelial cancer, NCI-sponsored) [4], and NCT06696768 (Phase 1 with FOLFOX plus bevacizumab in metastatic colorectal cancer, NCI-sponsored) [5]. The two NCI-sponsored trials matter for one specific reason: they show external academic interest in IRAK4 inhibition beyond hematology, though they are early dose-finding and won't generate registrational data. Prior human evidence - anchor for the CLL bet. CA-4948-101 (TakeAim Lymphoma, NCT03328078) reported preliminary data at ASH 2024 in R/R PCNSL with prior BTK exposure: among 16 evaluable patients in combination with ibrutinib, 5 complete responses (3 PCNSL, 2 mantle cell lymphoma) and 1 partial response in a CLL patient [10]. Doses ranged 100-300 mg BID emavusertib with ibrutinib 560 mg QD. The combination was tolerated with no dose-limiting toxicities reported in that subset [10]. The single CLL PR is anecdotal but is the only public human signal of emavusertib activity in CLL and is the empirical seed under the Phase 2 design. Commercial structure: Curis is the sponsor of the Phase 2; BeiGene supplies zanubrutinib for the trial. Detailed economic terms (milestones, royalties, CLL rights on success) have not been disclosed in Curis's 8-Ks or earnings materials as of Q1 2026 [9], so any read-through to value distribution is speculative.

IRAK4 (interleukin-1 receptor-associated kinase 4) is an enzyme inside immune and B-cells that acts as the first kinase fired when toll-like receptors or IL-1 family receptors detect danger signals. In normal biology, it turns on NF-kB, the master switch that tells the cell to survive and divide. In cancer biology, that survival switch gets hijacked, most clearly in B-cell lymphomas carrying the MYD88 L265P mutation, which forces the Myddosome (a protein cluster that assembles around MYD88 and amplifies the danger signal) into a chronically active state and keeps IRAK4 driving survival signals around the clock [2]. That gives you a clean drug hypothesis: if a tumor cell depends on IRAK4-driven NF-kB to stay alive, blocking IRAK4 should starve it. The trickier question for CLL specifically is the MYD88 mutation rate. In Waldenström's macroglobulinemia, MYD88 L265P is present in >90% of patients. In ABC-DLBCL (activated B-cell-like diffuse large B-cell lymphoma, an aggressive subtype), the L265P rate runs ~30%. In CLL, by contrast, MYD88 mutations occur in roughly 2-4% of Caucasian patients and around 8% in some Asian cohorts [11]. So the genetic addiction argument applies to only a small slice of the CLL trial population. The bet is that combining IRAK4 blockade with BTK blockade hits both arms of B-cell NF-kB activation (BTK runs through the B-cell receptor, IRAK4 runs through TLR/IL-1R) and produces deeper responses than BTK inhibition alone, even in MYD88 wild-type disease. The PCNSL combination data ([10]) is the precedent the sponsor leans on. How strong is the case? Mixed. The genetic argument in MYD88-mutated lymphomas is solid. The CLL argument is more speculative because of the low MYD88 mutation rate. No IRAK4 inhibitor has won approval anywhere, so the target itself is not de-risked the way BTK is post-ibrutinib.

NCT07271667 plans to enroll 108 patients with CLL and other B-cell malignancies across two cohorts, all receiving emavusertib in combination with zanubrutinib [1][9]. Per Curis's Q1 2026 update, Cohort 1 enrolls patients already on BTKi monotherapy who achieved partial remission but have not achieved complete remission or uMRD - the design adds emavusertib to deepen response [9]. Curis is the sponsor; BeiGene supplies zanubrutinib. The trial is single-arm, with no head-to-head against zanubrutinib monotherapy [1]. The primary endpoint is uMRD rate in Cohort 1. uMRD (undetectable measurable residual disease) is typically measured by flow cytometry or next-generation sequencing at thresholds below one CLL cell per 10,000 healthy lymphocytes. The FDA has accepted uMRD as supportive evidence in recent CLL approvals because patients who reach it stay progression-free for years. Two design concerns. First, a single-arm trial with a deep-response endpoint cannot prove the combo beats zanubrutinib alone - Curis will have to compare against historical control data. The relevant benchmark is the rate at which BTKi-monotherapy patients in partial response convert to uMRD on continued monotherapy, which is low but not zero; the combo number has to land meaningfully above it. The writeup previously cited a 'teens' zanubrutinib monotherapy uMRD figure - that number applies to all-comer frontline CLL and is the wrong comparator for this salvage-deepening design, so it has been removed pending a verifiable comparator from BTKi-PR subsets. Second, mixing CLL with other B-cell malignancies complicates interpretation if the responder pool skews to one subtype, particularly the MYD88-mutated populations where the mechanism is strongest. The trial protocol as publicly disclosed does not stratify by MYD88 status [1]; whether MYD88 will be collected as a correlative for retrospective analysis is not stated in public materials. Enrollment pace also matters. Curis's prior trials have run slow, partly due to safety holds and partly due to limited site network. With first-patient dosing only beginning mid-2026 and 108 patients planned, full enrollment will likely take 18-24 months from that start.

Our model gives this drug a 15% chance of eventually reaching approval. That starting point comes from the historical approval rate for Phase 2 drugs in this area, which sits around 21%. The estimate is nudged up by a non-randomized trial design and more secondary endpoints than usual, but pulled down by the sponsor's weak approval record and limited earlier-phase results. The remaining factors fall close to average for this stage, so they leave the final number near where the base rate started.

Safety. The biggest specific risk is rhabdomyolysis. Curis's AML/MDS program was placed on partial clinical hold by FDA in April 2022 after reports of muscle breakdown in patients taking emavusertib [6]. The hold was adjusted and trials resumed, but the signal is real and mechanism-related thinking has not fully ruled out an on-target contribution. CLL patients are generally older with more comorbidities than the typical AML population, which raises the toxicity bar. Efficacy. CLL is overwhelmingly MYD88 wild-type - only roughly 2-4% of Caucasian CLL patients carry MYD88 mutations [11] - so the strongest genetic argument for IRAK4 dependence does not apply to most of the trial population. If the combo only adds depth-of-response in a small MYD88-mutant subset, the trial may miss the overall uMRD endpoint even with a real biological signal in a subgroup, and the lack of pre-specified stratification means a positive subset finding will be hard to defend regulatorily. Commercial. CLL combinations are crowded. Acalabrutinib plus venetoclax (AMPLIFY) was FDA-approved February 19, 2026 as the first all-oral, fixed-duration frontline CLL/SLL regimen [12]. Ibrutinib plus venetoclax already set high uMRD bars, and BeiGene is advancing zanubrutinib plus the BCL2 inhibitor sonrotoclax (BGB-11417). An IRAK4 add-on has to differentiate from a BCL2 add-on, and BCL2 inhibition is mechanistically cleaner and has a deeper evidence base. Execution and financial. Curis is a micro-cap with thin liquidity. Per the Q1 2026 10-Q, cash, cash equivalents and restricted cash totaled $20.8 million as of March 31, 2026, against $9.0 million of cash used in operations that quarter [7][9]. At that burn, the runway is roughly 7 quarters - implying cash exhaustion around late 2027 or early 2028 absent further financing. Curis raised approximately $8.8 million net via equity in March 2025 and ~$6 million in July 2025 to extend runway [9]. The December 2026 initial CLL readout is reachable on existing cash; full cohort 1 maturation in 2027 is not without an additional raise or partnership. A partnership or out-licensing on the CLL program would change the trajectory; a dilutive raise would weigh on CRIS.

Watch this for the biology. Curis is small enough that share price moves on press releases more than on drug economics, and the real question - whether IRAK4 inhibition adds anything to BTK inhibition in CLL - will get a first answer from the December 2026 initial 5-patient readout and then from cohort 1 in 2027, regardless of where CRIS trades. The signal to wait for: any uMRD conversions in the first 5 patients (Dec 2026), then the cohort 1 uMRD rate at the planned assessment window. Even one or two conversions from BTKi-PR to uMRD in the December 2026 readout would be a meaningful early signal because the patients enrolled have already failed to deepen on BTKi alone. A cohort 1 uMRD rate that materially exceeds what continued BTKi monotherapy delivers in the same partial-response population would be a real result; failure to convert any patients would suggest the IRAK4 hypothesis in CLL is wrong at this dose and schedule. Track whether Curis reports MYD88 status alongside responses - if they do, even retrospectively, the subset signal will dominate interpretation. The signal to ignore: any interim safety update without efficacy. Curis already knows where the toxicity floor is from prior trials, and re-confirming tolerability is not new information. Check back: December 2026 for the first 5-patient readout. Also worth tracking the NCI-sponsored urothelial (NCT06439836) and colorectal (NCT06696768) Phase 1 trials [4][5]. They read out faster and would expand the IRAK4 thesis beyond hematology if positive, which would matter more for the asset's long-term value than any single Curis-sponsored study. Watch the Curis cash position and any financing 8-Ks in 2026-2027 - the runway is tight enough that a bad raise could damage the equity even if the biology works.