Bayesian Dose-Escalation Designs in Tigilanol Tiglate Phase I Trials: How EBC-46 Researchers Find a Safe Working Dose
Phase I trials of tigilanol tiglate use Bayesian and modified 3+3 dose-escalation schemes. This article explains how those statistical designs decide the next-cohort dose.
Phase I oncology trials answer a deceptively simple question: at what dose can a new agent be given without unacceptable toxicity, and what is the dose worth carrying forward into efficacy studies? For tigilanol tiglate (EBC-46), the question is sharpened by the agent's intratumoural delivery: dose escalation has to balance local tissue effects, systemic exposure, and tumour-volume-specific dosing, rather than the body-surface-area scaling used for systemic chemotherapeutics. Recent QBiotics-sponsored Phase I trials in soft-tissue sarcoma and head-and-neck cancers have applied modified Bayesian dose-escalation frameworks to navigate these constraints more efficiently than classic 3+3 schemes allow.
Why 3+3 is limiting for an intratumoural injectable
The traditional 3+3 design escalates dose in fixed steps and tests three patients per dose level, with rules for adding three more if a dose-limiting toxicity (DLT) is observed. It is robust and easy to interpret, but it is inefficient at finding the maximum tolerated dose (MTD), tends to under-dose patients in lower cohorts, and treats DLT data as a binary endpoint rather than borrowing information across dose levels. For an intratumoural agent like tigilanol tiglate, the toxicity profile is dominated by local effects (injection-site pain, ulceration, slough) whose grade depends as much on tumour volume and anatomical location as on absolute dose — a clinical reality the 3+3 framework does not capture well.
Bayesian Optimal Interval (BOIN) and continual reassessment
Modern Phase I designs increasingly use the Bayesian Optimal Interval (BOIN) or modified continual reassessment method (mCRM). Both treat DLT probability as a continuous parameter to be estimated, update the model after each cohort using a prior plus observed events, and propose the next dose by reference to a pre-specified target toxicity rate — typically 25–30% DLT. The advantage in tigilanol tiglate trials is that information from earlier dose levels informs decisions at higher levels, allowing the design to skip lower doses if early data are clean, or to hold a dose for an extra cohort if a single ambiguous event needs more observation.
Most tigilanol tiglate Phase I protocols also include a planned dose-de-escalation rule: if a DLT-rate posterior estimate exceeds the upper interval boundary, the next cohort is dosed at a lower level rather than the previous level being declared the MTD. This conservative behaviour matters because the toxicity signature includes wound-healing complications that may take weeks to fully evolve.
Tumour-volume-adjusted dosing
Unlike systemic dosing scaled to body weight or surface area, tigilanol tiglate trials dose by tumour volume — typically expressed as mg of tigilanol tiglate per cm³ of tumour. This conversion adds a structural assumption to the dose-escalation calculation: every cohort's nominal dose is a volume-normalised dose, and the protocol cap on total injected volume per session imposes a ceiling that becomes relevant for larger lesions. Quality-of-life endpoints and patient-reported pain scores are collected alongside DLT data, but the formal escalation rule operates on the volume-normalised dose.
DLT definitions are not trivial
How a Phase I trial defines a DLT shapes everything downstream. Tigilanol tiglate trials typically count Grade 4 toxicities or Grade 3 toxicities lasting more than seven days as DLTs, with specific carve-outs for expected local wound-healing reactions that resolve on schedule. This is a deliberate design choice: the agent's mechanism produces tumour necrosis and a wound bed that heals over several weeks. Treating every Grade 2–3 local reaction as a DLT would prematurely cap the dose. The protocol therefore distinguishes "on-mechanism" local effects from "off-target" toxicities — a distinction reviewed in detail in the published Phase 2 adverse-event profile.
Translating Phase I findings into Phase II
The output of a Bayesian Phase I design is not just an MTD but a posterior distribution over DLT probability at each tested dose. This richer summary lets the Phase II steering committee choose a recommended Phase II dose (RP2D) that balances DLT risk and projected efficacy, rather than mechanically picking the highest dose with <33% DLT events. For tigilanol tiglate, the RP2D has typically been one dose level below the formal MTD, reflecting a preference for predictable wound healing in larger pivotal cohorts.
Citations
3. QBiotics Group — Tigilanol tiglate clinical programme, 2026.
4. US FDA — Guidance for industry: oncology dose-finding (Project Optimus), 2024.
Related articles
- Pharmacokinetic Monitoring in Tigilanol Tiglate Phase I Trials: EBC-46 Drug Metabolism Considerations
- Phase III Pivotal Trial Design Considerations for Tigilanol Tiglate (EBC-46)
This article describes investigational clinical trial methodology. Tigilanol tiglate is an investigational pharmaceutical agent; readers should not interpret any of the above as medical advice or as a recommendation regarding any specific therapy.
