QCast Episode 60: Oncology Phase 1 Trial Design

Oncology Phase 1 trial design increasingly has to answer a broader question than how much treatment a patient can tolerate. In this QCast episode, co-hosts Jullia and Tom look at how dose optimisation, escalation design, simulation and early clinical data contribute to selecting a dose and regimen that can be taken forward with confidence. They also consider why the traditional maximum tolerated dose approach may be less informative for some targeted therapies and immunotherapies, particularly when toxicity emerges later or accumulates over repeated cycles.

The practical challenge is that these decisions depend on more than the statistical design alone. Dose-review meetings need current safety, laboratory, PK and PD data. Biomarker-driven eligibility can introduce screening and laboratory turnaround constraints, while expansion cohorts, combination regimens and more complex trial structures can increase both the scientific and operational demands on the study. Site capability also matters, particularly where trials involve intensive assessments, time-sensitive samples, oncology pharmacy support or urgent safety review.

🎧 Listen to the Episode:

Key Takeaways

Why dose optimisation goes beyond the maximum tolerated dose

A dose can meet an early DLT-based safety threshold and still prove difficult for patients to sustain over time. Repeated dose interruptions, reductions or discontinuations may show that the regimen is poorly suited to longer treatment, even if the initial escalation decision appeared acceptable. Dose selection therefore needs to consider longer-term tolerability alongside exposure, pharmacodynamic information and early evidence of activity.

How simulation can improve escalation decisions

Rule-based, model-assisted and model-based escalation approaches can behave differently under the same clinical assumptions. Simulation allows teams to test those differences before enrolment by examining outcomes such as correct dose selection, exposure to overly toxic or subtherapeutic doses and the effect of delayed toxicity. This helps determine whether the proposed design is appropriate for the expected behaviour of the treatment rather than relying on convention alone.

Why operational readiness affects dose review

A dose-escalation design can only use the information that is available when a review takes place. Delayed adverse event entry, unresolved queries, missing laboratory data or late PK and PD results can postpone decisions or leave the review team working with incomplete evidence. Biomarker testing and site capability add further dependencies, particularly when eligibility, treatment timing or safety review depends on fast turnaround and tightly coordinated trial activity.

Full Transcript

Jullia
Welcome to QCast, the show where biometric expertise meets data-driven dialogue. I’m Jullia.

Tom
I’m Tom, and in each episode, we dive into the methodologies, case studies, regulatory shifts, and industry trends shaping modern drug development.

Jullia
Whether you’re in biotech, pharma or life sciences, we’re here to bring you practical insights straight from a leading biometrics CRO. Let’s get started.

Tom
Now Phase 1 usually makes people think about one fairly straightforward question. How much of a new treatment can patients safely receive? Does that still describe what we’re trying to do in an oncology Phase 1 trial?

Jullia
Only partly. See, safety and tolerability are still central, but the decision has become broader than finding the highest dose patients can tolerate. We’re trying to understand the dose and regimen we can take forward, using safety alongside pharmacokinetics, pharmacodynamics and any early evidence of anti-tumour activity.

And oncology is unusual because these first-in-human studies generally involve patients with advanced cancer rather than healthy volunteers. That changes the clinical context around every dose decision.

Tom
And I suppose the type of treatment has changed the question as well. The traditional maximum tolerated dose approach came from cytotoxic chemotherapy, but we now have targeted therapies and immunotherapies. What does that change?

Jullia
It changes quite a lot. With some newer therapies, increasing the dose doesn't necessarily produce a corresponding increase in efficacy. Toxicities can also emerge later or accumulate over repeated treatment cycles.

So a dose might appear acceptable within the predefined dose-limiting toxicity, or DLT, window, but become difficult for patients to sustain over time. You may then see repeated dose interruptions, reductions or treatment discontinuations. That gives you a very different picture of whether the regimen is genuinely suitable.

Tom
That's an interesting distinction because that means a dose can technically pass the early safety assessment and still be a poor dose to take into a larger trial?

Jullia
Exactly, and that’s also reflected in current regulatory thinking. FDA's Project Optimus, for example, has reinforced the need to characterise and optimise oncology dosing rather than simply defaulting to the maximum tolerated dose.

That can involve exploring more than one dose and looking carefully at exposure, activity, safety and longer-term tolerability before settling on the dose and schedule you want to develop further.

Tom
So when a team actually starts escalating, what are its options? The 3+3 design is probably the one most will recognise.

Jullia
Yes, and 3+3 is still used. It's a rule-based approach where small cohorts are treated at successive dose levels and escalation depends on the number of DLTs observed.

But there are other options too. Model-assisted designs such as BOIN and model-based approaches such as CRM can use the accumulating toxicity information differently. They may give teams more flexibility or more explicit control over the probability of treating patients at doses that are too toxic.

Tom
Can you give me an example of where the choice of design might make a real difference rather than just being a statistical preference?

Jullia
So imagine toxicity tends to appear later than expected. Patients enter a cohort, complete the early part of the first cycle and the immediate safety picture looks acceptable. If the design and decision process don't deal well with incomplete toxicity information, the study could escalate while clinically important events are still developing.

Before the trial starts, you can simulate scenarios like that. You can ask how often a proposed design would select the correct dose, how many patients might receive doses that are too low or too toxic, and what happens if toxicity is delayed.

Tom
So simulation is really testing the behaviour of the design before you're relying on it with real patients?

Jullia
Yes, see, those operating characteristics help turn a choice between designs into something you can evaluate quantitatively. A method that looks attractive in theory may behave quite differently once you apply plausible assumptions about toxicity rates, delayed events, cohort sizes and enrolment.

The DLT window also has to be appropriate, and late toxicities need a defined place in decision-making.

Tom
Then there’s the information available at the actual dose review. I imagine it can control the pace of the trial?

Jullia
It can. See, a dose decision may depend on adverse events being entered promptly, relevant laboratory data being available, PK or PD results being reviewed and queries on escalation-critical data being resolved.

If the safety review meeting arrives and one site's data is incomplete, the statistical design can't compensate for that. You either delay the decision or make it with less information than intended. Early oncology studies have a very tight connection between data quality and operational cycle time.

Tom
Now once you've found a plausible dose, expansion cohorts can then start building evidence in particular tumour types or patient groups. Is there a risk that expansion begins too quickly?

Jullia
It’s possible. Expansion cohorts are valuable because they let you examine the selected dose in more patients, sometimes within a particular tumour type or biomarker-defined group, and look more closely at safety and early activity.

But you're also exposing more patients at that dose. If dose optimisation isn't mature enough, a large expansion programme can effectively lock in a dosing decision before you've fully understood it. So the purpose of each cohort and the criteria for continuing or stopping it therefore need to be clear.

And combinations add another layer because you're no longer looking at the new treatment in isolation. Overlapping toxicities, drug to drug interactions and schedule effects can all alter tolerability or exposure. A dose that worked as monotherapy might behave differently when another treatment is added.

That makes the rationale for the combination dose and schedule particularly important. You need enough information to understand what each component is contributing and whether the regimen patients actually receive is sustainable.

Tom
Biomarkers are also much more prominent in early oncology development now. They sound scientifically attractive because they can help identify the patients most likely to respond. Is it easy to underestimate what they do to the trial operationally?

Jullia
It’s very easy, yes. The biological rationale may be strong, but the assay has to work within the clinical process. If eligibility depends on a biomarker result, you need to know whether the sample can be collected, shipped, tested and reported within the screening window.

Prevalence matters too. A biomarker may give you a more targeted population, but if only a small proportion of screened patients qualify, recruitment changes significantly. You also need clarity on whether the biomarker is exploratory or actually being used to assign treatment.

Tom
So take a patient arriving at site who needs a biomarker result before enrolment, for example. What can derail that apparently simple step?

Jullia
Well, the sample might miss a shipping cut-off. The central laboratory turnaround might be longer than the protocol assumes. You could have differences between local and central testing, or problems with collection timing, storage or sample quality.

Any of those can affect whether the patient is eligible within the required timeframe. If you're using emerging measures such as circulating tumour DNA, interpretation and timing also need to be considered before enrolment begins.

Tom
We've also seen basket, umbrella and platform approaches become part of early oncology. Do they fundamentally change the principles we've been talking about?

Jullia
So the underlying principles stay similar, but the coordination burden increases. A basket trial might study one treatment across different tumour types, while an umbrella trial can evaluate several treatments within one tumour type. Platform designs can allow treatments or cohorts to enter and leave over time.

That flexibility can make learning more efficient, but decision rules need to be very clear. When cohorts are opening, pausing or closing at different points, version control around eligibility criteria, assessments and treatment rules becomes particularly important.

Tom
How much of that then comes down to whether sites can actually deliver the protocol?

Jullia
So site readiness is a major part of that. Early-phase oncology can involve intensive visits, frequent safety assessments, complex PK or PD sampling and time-sensitive treatment decisions. Sites need the clinical experience and operational capacity to manage those demands.

There can also be very practical requirements such as out-of-hours cover for urgent safety events, oncology pharmacy support, weekend sample shipping or tumour-specific imaging. If those assumptions aren't tested during feasibility, protocol deviations and delays become much more likely once recruitment starts.

Tom
If you had to leave listeners with a few takeaways before we finish, what should stay with them?

Jullia
First, Phase 1 oncology dose finding is increasingly about selecting a sustainable dose and regimen, not simply identifying the maximum tolerated dose.

Second, the escalation method needs to be judged by how it behaves under realistic conditions. Simulation and operating characteristics can expose problems before patients are enrolled.

And finally, scientific design and trial delivery can't really be separated. Biomarker turnaround, safety data entry, query resolution, site capability and review cadence can all determine whether the planned dose decisions are actually made using the evidence you intended.

Jullia
With that, we’ve come to the end of today’s episode on oncology phase 1 trial design. If you found this discussion useful, don’t forget to subscribe to QCast so you never miss an episode and share it with a colleague. And if you’d like to learn more about how Quanticate supports data-driven solutions in clinical trials, head to our website or get in touch.

Tom
Thanks for tuning in, and we’ll see you in the next episode.

About QCast

QCast by Quanticate is the podcast for biotech, pharma, and life science leaders looking to deepen their understanding of biometrics and modern drug development. Join co-hosts Tom and Jullia as they explore methodologies, case studies, regulatory shifts, and industry trends shaping the future of clinical research. Where biometric expertise meets data-driven dialogue, QCast delivers practical insights and thought leadership to inform your next breakthrough.

Subscribe to QCast on Apple Podcasts or Spotify to never miss an episode.