OMX-0407 in pretreated solid tumors: a SIK inhibitor closed the week after its expansion readout window

What was tried

NCT05826600 was a Phase 1/1b dose escalation and cohort expansion study of OMX-0407, an oral salt-inducible kinase inhibitor, in patients with previously treated unresectable solid tumors. The sponsor was iOmx Therapeutics AG. It opened on 30 March 2023 and ran across 24 sites in Belgium, France and Spain as an open label, non-randomized, sequential design, enrolling 68 participants.

Escalation began at 20 mg daily, split as 10 mg twice daily, with dose steps set by a safety committee. Expansion used 100 mg orally twice daily, a fivefold increase over the starting total daily dose. Eligibility required confirmed advanced cancer, at least one prior line of systemic therapy for the tumor type under investigation, and ECOG performance status of 0 to 2 in escalation and 0 to 1 in expansion. Participants from the third cohort onward had to provide archival tumor material or undergo a screening core biopsy.

Two co-primary outcomes were registered: dose limiting toxicities over a four week cycle, and objective response rate assessed every 12 weeks in renal cell carcinoma, non-small cell lung cancer, urothelial cancer and angiosarcoma. Secondary outcomes covered maximum tolerated dose, adverse events by dose level, and pharmacokinetics.

The biological hypothesis

Salt-inducible kinases sit in the AMPK-related branch of the kinome and repress CRTC and HDAC-dependent transcription. In myeloid cells, SIK inhibition shifts macrophages away from an immunosuppressive program, which is the basis for treating SIK as an immuno-oncology target rather than a direct antiproliferative one. iOmx described OMX-0407 as a first-in-class, spectrum-selective kinase inhibitor hitting oncology-relevant tyrosine kinases alongside the SIKs, acting both by direct inhibition of tumor cell proliferation and by potentiating death receptor signaling while reshaping the tumor microenvironment.

The dual mechanism is the hypothesis and the problem. A spectrum-selective inhibitor lets one agent combine cell-intrinsic and immune effects, and it also leaves no clean way to attribute a response to a named kinase. The Open Targets record makes the evidence gap explicit. SIK3 has 2,385 disease associations, and its strongest is spondyloepiphyseal dysplasia at 0.582, carried by a genetic association score of 0.649. For neoplasm the score is 0.137, built from literature co-mention at 0.940 and animal model evidence at 0.614, with human genetic association at 0.050. Open Targets lists no known drugs or clinical candidates against SIK3. The genetics for this gene point at the skeleton, and the oncology case rests on mouse models and pathway reasoning.

What actually happened

The registry records primary completion and study completion on 2 April 2026, and gives the reason in one sentence: the study was terminated as part of strategic considerations and not based on safety concerns. The termination record was posted on 24 July 2026. No results are posted to ClinicalTrials.gov.

Enrollment tells part of the story. iOmx reported 20 patients in the completed escalation and set the expansion target at up to 86, for a maximum planned total of about 106. Actual enrollment closed at 68. Subtracting the reported escalation cohort leaves roughly 48 participants across the expansion, a figure computed from those two sourced inputs rather than a reported cohort count.

The publicly disclosed efficacy signal was thin before expansion started. The escalation package described a favorable safety profile, pharmacological activity in circulating blood cells, and one patient with a durable complete response after resistance to multiple prior chemotherapies. On that basis two indications were prioritized, renal cell carcinoma and angiosarcoma, with objective response rate as the endpoint and topline proof-of-concept data expected in early 2026. Completion landed on 2 April 2026, and termination followed.

Failure mechanism, best guess

The registry gives strategic considerations, so the archetype recorded here is strategic_reprioritization. The timing constrains what that phrase can mean. Expansion was built around objective response rate, topline data was guided to early 2026, completion occurred on 2 April 2026, and the program was closed. A sponsor that had seen a response rate worth financing in either indication would ordinarily disclose it while raising against it. A closure decided immediately after the readout window, with no expansion result disclosed, is consistent with a response rate that did not clear the bar, though no efficacy number has been published and none should be assumed.

Two design features made that outcome more likely. The first is a single-arm objective response rate endpoint across four tumor types with no prespecified enrichment marker. Angiosarcoma and renal cell carcinoma have very different response base rates to kinase inhibitors, so a pooled design leaves any observed response ambiguous between drug effect and tumor selection. The second is target attribution. An inhibitor that hits both tyrosine kinases and SIKs cannot tell a sponsor which activity produced a response, so a promising escalation signal does not narrow the next study.

Behind both sits the genetic deficit. A target whose genetic evidence for neoplasm scores 0.050, against 0.649 for a skeletal dysplasia, offers no independent human anchor when a single-arm response rate disappoints.

How to prevent this next time

No endpoint level data was posted, so no power or Bayesian calculation is supportable here. The usable levers are qualitative.

First, biomarker enrichment. The protocol already collected archival or fresh tumor tissue from the third cohort onward. A prespecified marker of SIK pathway inhibition in tumor or myeloid cells, tied to a go or no-go threshold, would have separated target engagement from absence of drug effect when the response rate came in low.

Second, indication discipline. Four tumor types across roughly 48 expansion patients yields cohorts too small to estimate a response rate with usable precision. Two cohorts of 24, each with a Simon two-stage boundary, would have produced an interpretable answer on the same enrollment.

Third, base rate adjustment. A first-in-class inhibitor against a target with no clinical precedent and no human genetic support for the indication should carry an explicitly lower prior probability of success, and the expansion should have been sized against that prior rather than against a single complete response.

Fourth, attribution by design. Where a compound is deliberately spectrum-selective, a pharmacodynamic readout for each intended axis is what converts a response into a reusable finding.

The single highest leverage change would have been to make expansion entry conditional on a prespecified SIK pathway engagement marker in tumor tissue, so a low response rate could be interpreted rather than merely absorbed.

What this means for similar programs

Spectrum-selective kinase inhibitors trade mechanistic clarity for breadth, and that trade should be priced into the plan rather than discovered at the readout. A single durable complete response in a 20 patient escalation is a reason to continue and not a basis for choosing indications. Where Open Targets genetic evidence for a target and disease pair is near zero while literature and animal model evidence is high, the program carries translational risk that only a biomarker-anchored design can retire. Terminations labelled strategic deserve the same scrutiny as efficacy terminations, since a closure timed to a readout window carries information the registry text does not.

Open questions

What was the objective response rate in the renal cell carcinoma and angiosarcoma cohorts, and how were the 48 expansion participants split across the four tumor types? What was the maximum tolerated dose, and did 100 mg twice daily achieve sustained SIK pathway inhibition in tumor tissue? What tumor type was the escalation complete responder, and did that response hold at closure? Will iOmx publish the expansion data?

Sources

1. • ClinicalTrials.gov, NCT05826600, full study record, retrieved 31 July 2026. https://clinicaltrials.gov/study/NCT05826600 - Open Targets Platform API v26.6.3, target SIK3 (ENSG00000160584), association scores for MONDO_0005070 and Orphanet_253, retrieved 31 July 2026. https://platform.opentargets.org/target/ENSG00000160584 - iOmx Therapeutics, Phase Ib initiation release reporting 20 patients in escalation and expansion of up to 86 patients in renal cell carcinoma and angiosarcoma. https://www.biospace.com/press-releases/iomx-therapeutics-initiates-phase-ib-with-omx-0407 - OMX-0407: a novel spectrum-selective small molecule kinase inhibitor in advanced or metastatic solid tumors. Journal of Clinical Oncology, 2025 ASCO Annual Meeting abstract 3103. https://- Feng S, et al. Roles of salt-inducible kinases in cancer (Review). International Journal of Oncology, 2023 (). https://- Sun Y, et al. Targeting TBK1 to overcome resistance to cancer immunotherapy. Nature, 2023 (). https://- Klaeger S, et al. The target landscape of clinical kinase drugs. Science, 2017 (). https://.