Could Accelerated Approval Guidance Usher In The Era of a Novel Trial Design in Oncology: Introducing TwICS

The recent FDA draft guidance on accelerated approval in oncology is firmly rooted in some of the limitations of single-arm studies. Challenges related to determining safety event attribution, the interpretation of time-to-event (TTE) endpoints and the assignment of treatment effects to combination regimens arm are some of the reasons the agency cites for its new draft guidance.

As a solution, the FDA has proposed that sponsors either conduct two smaller randomised control trials or a single large trial to support an accelerated approval application; these recommendations and their implications are summarised in a previous article here.

Could this draft guidance present an opportunity for a unique trial design?

By looking through the guidance, we can theorise the features that a trial would need to support accelerated approval; these are:

1. Randomisation: The agency notes concern with confounding when comparisons are made to external data.  
2. Safety database size and ability to assign events to the intervention.
3. Ability to determine response endpoints (such as ORR) in an appropriate manner.
4. Ability to determine treatment effects for components of combination regimens.
5. Maintenance of data integrity
6. That controls are representative of current standards of care.

Given the above, it is inviting to postulate that a Trial Within Cohorts (TwICS) design could satisfy these requirements.

The TwICS design:

The TwICS design is a pragmatic and adaptive approach that merges features of randomised controlled trials (RCTs) and observational cohort studies. In TwICS studies, an observational cohort (or registry) of patients with a disease of interest is first established. Eligible patients in this registry are subsequently randomly selected to receive an intervention or continue with the standard of care. As the randomisation of some is equivalent to the randomisation of all, this is comparable to a parallel randomised design. As in any parallel randomised design, endpoints are then assessed across both arms at pre-specified time points.  

Advantages of the TwICS Design

1. The trial is randomised, meaning comparisons with a control arm are significantly less complex than using an external control; furthermore, time-to-event endpoints are valid.
2. Safety database size & attribution would be relatively straightforward.
3. This design has the potential to accelerate patient recruitment for the entire clinical program significantly.
4. A single well-designed cohort or registry could support multiple trials.
5. The pragmatic nature of the design would ensure that current standards of care would be available as comparators.
6. Multi-arm studies are possible, allowing for the assessment of combination regimens.
7. This design provides an opportunity for proactive engagement with patients & supports groups and extended collaboration with academic institutions and therapeutic area experts.
8. This design allows for both the "one-trial" or "two-trial" approach as recommended by the FDA.

Challenges to the TwICS design:

Like anything in statistics & clinical research, there is no such thing as a free lunch. However, with some creativity, these challenges can be mitigated.

1.  There is a risk of bias & residual confounding if there is an imbalance between the intervention & control group. This can be mitigated by careful consideration of inclusion criteria for the observational cohort.
2. There is a risk of differential temporal assessment between the intervention and control groups. This can be mitigated by designating the observational cohort a low intervention study (LIS) and mandating follow-up assessments in the protocol.
3. There is a risk of differential endpoint assessment (e.g., criteria for progression) between the intervention & control arm. As previously stated, this could be mitigated by the designation of the observational cohort as a LIS or by establishing a response assessment charter as part of the protocol.
4.  There is the potential for operational complexity due to operational teams needing to gain experience conducting such studies. This can be mitigated through the proactive engagement of the operations team at the early study conception phase and by driving effective ideation workstreams.

How InovaSight can support Innovative approaches such as TwICS for accelerated approval & clinical development?

InovaSight offers a robust platform that empowers development teams to create trial prototypes and designs rapidly and precisely. The platform allows users to generate crucial assumptions regarding treatment effects, establish relevant endpoints, and optimise inclusion-exclusion criteria. Additionally, InovaSight can forecast potential changes in the standard of care and calculate the probability of technical success (PTS) for specific trial designs. The platform's ability to accurately identify and prioritise indications for a given drug and seamlessly connect teams to pertinent key opinion leaders (KOLs) renders InovaSight an indispensable resource for development teams. Furthermore, InovaSight is bundled with expert drug development & study design advisory, ensuring development teams can tap into a wide array of experience when developing a CDP.

‍