How can I objectively compare doses under FDA Project Optimus?

The FDA’s Project Optimus is changing the way sponsors think about dose selection.

For decades, oncology development relied heavily on the Maximum Tolerated Dose (MTD) paradigm. The assumption was simple: higher doses would deliver greater efficacy, as long as toxicity remained manageable.

Project Optimus challenges that paradigm.

The FDA now expects sponsors to identify the dose that optimizes the benefit and risk of a treatment. In practice, that means considering efficacy, safety, PK/PD, tolerability and increasingly patient experience together.

This sounds reasonable, until you look at the data.

Imagine a dose optimization study comparing 200 mg and 400 mg:

Endpoint	200 mg	400 mg
ORR	40%	44%
Grade ≥3 AEs	16%	24%
Treatment discontinuation	8%	13%

Which dose would you choose?

The higher dose shows higher efficacy but the lower dose appears less toxic and more tolerable. We could expect patient experience to favor the lower dose.

There is no obvious answer.

And that is precisely the challenge Project Optimus creates:

How can multiple outcomes be combined objectively to support dose selection?

Challenge 1: How can efficacy and safety be integrated quantitatively?

Most dose optimization programs already use exposure-response modelling.

Exposure-response analyses are essential. They help characterize how efficacy, toxicity and biomarker activity evolve with increasing exposure, but they don’t fully solve the decision problem.

Eventually, sponsors still need to answer questions such as:

• How much toxicity is acceptable for a modest efficacy gain?
• Is a small increase in ORR worth a steep increase of Grade ≥3 adverse events?
• How to integrate patient-reported outcomes in dose selection decisions?

In other words: How do we move from one-by-one marginal endpoints assessment to a single assessment of benefits and risks?

There exist utility-based approaches, such as Clinical Utility Index (CUI) or Multi-Criteria Decision Analysis (MCDA), that are often used to combine efficacy and safety into a single score.

But they introduce another significant challenge: weighting.

Challenge 2: How can doses be compared without arbitrary weights?

Most utility approaches rely on weights. For example:

• efficacy: 50%
• safety: 30%
• biomarker response: 20%

The dose with the highest score is selected.

But who decides the weights? Should sponsors, clinicians or patients decide? Should the importance of toxicity depend on disease severity?

Different stakeholders often arrive at different answers.

This is one of the reasons why sponsors increasingly look for approaches that avoid explicit weighting altogether.

Challenge 3: Is there an alternative to utility functions?

An alternative is to stop asking “How much efficacy is worth relative to toxicity?” and instead ask “Which outcomes are clinically more important?”

This is the principle behind Generalized Pairwise Comparisons (GPC) and Net Treatment Benefit (NTB) methodology.

Rather than assigning numerical weights, sponsors define a hierarchy of outcomes.

For example:

1. Overall survival
2. Grade ≥3 adverse events
3. Objective response rate
4. Patient-reported quality of life

Every patient receiving Dose A is then compared with every patient receiving Dose B.

The comparison starts with the most important outcome. If one patient performs better, the comparison ends. If not, the analysis moves to the next outcome.

The final result is a single metric summarizing how often one dose provides a clinically meaningful advantage over another across all prioritized outcomes.

This measure is called the Net Treatment Benefit (NTB).

Why is Net Treatment Benefit relevant for Project Optimus?

Project Optimus is fundamentally about multidimensional decision-making.

The FDA encourages sponsors to integrate:

• efficacy
• safety and tolerability
• pharmacokinetics
• pharmacodynamics
• patient experience

while making dose selection transparent and clinically meaningful. NTB aligns naturally with these objectives.

First, it provides a single measure of treatment benefit while preserving the clinical importance of individual outcomes.

Second, it avoids explicit utility weights. Sponsors do not need to agree that efficacy is worth exactly twice as much as toxicity. Instead, they define which outcomes matter most and let the analysis quantify the overall treatment advantage.

Third, NTB works at the patient level. The comparison is not performed on average outcomes alone, but across all patients receiving the competing doses.

Finally, the endpoint hierarchy can reflect what matters to different stakeholders, including patients and clinicians.

Beyond dose optimization: quantifying the totality of evidence

The same questions arise outside Project Optimus.

Sponsors often ask:

• Our primary endpoint was modest. What does the totality of the evidence tell us?
• Can efficacy, safety and quality of life be considered together?
• Is there a patient-centric way to quantify overall treatment benefit?

These are fundamentally the same questions that arise during dose optimization.

They are questions about integrating multiple outcomes in a transparent and clinically meaningful way.

Net Treatment Benefit is one approach designed specifically for these situations.

Frequently Asked Questions about Project Optimus and Net Treatment Benefit

What is FDA Project Optimus?

Project Optimus is an FDA initiative that aims to reform dose selection in oncology drug development. Rather than assuming the Maximum Tolerated Dose (MTD) is the optimal dose, the FDA encourages sponsors to identify doses that provide the best overall benefit-risk profile by considering efficacy, safety, pharmacokinetics, pharmacodynamics and patient experience.

Why is Maximum Tolerated Dose (MTD) no longer sufficient?

The MTD approach assumes that higher doses lead to greater efficacy until toxicity becomes unacceptable. However, many modern oncology therapies do not follow this pattern. Increasing the dose may produce limited additional efficacy while substantially increasing toxicity or negatively affecting quality of life. Project Optimus encourages sponsors to identify the dose that optimizes overall clinical benefit instead.

How do sponsors compare multiple doses under Project Optimus?

Sponsors typically compare doses using a combination of efficacy endpoints, safety data, exposure-response analyses, PK/PD measurements and patient-reported outcomes. The challenge is integrating these different outcomes into a transparent and clinically meaningful assessment that supports dose selection.

What is the Clinical Utility Index (CUI)?

The Clinical Utility Index (CUI) is a framework that combines multiple endpoints, such as efficacy and safety, into a single score using predefined weights. It can help compare treatment options or doses, but the results depend on how the weights are chosen and whether stakeholders agree on their relative importance.

What is Multi-Criteria Decision Analysis (MCDA) in drug development?

Multi-Criteria Decision Analysis (MCDA) is a structured decision-making approach that evaluates multiple outcomes simultaneously. Each outcome receives a weight reflecting its importance, and the weighted scores are combined to rank treatment options. MCDA is widely used in benefit-risk assessments but requires explicit assumptions about the relative value of different outcomes.

What is Generalized Pairwise Comparisons (GPC)?

Generalized Pairwise Comparison (GPC) is a statistical method that compares every patient in one treatment group with every patient in another group using a predefined hierarchy of outcomes. The comparison starts with the most clinically important endpoint and proceeds to lower-priority outcomes only if needed. This approach avoids assigning arbitrary numerical weights to endpoints.

What is Net Treatment Benefit (NTB)?

Net Treatment Benefit (NTB) is a summary measure derived from Generalized Pairwise Comparisons. It quantifies how often one treatment or dose provides a clinically meaningful advantage over another across multiple prioritized outcomes. NTB can integrate efficacy, safety and patient-reported outcomes into a single patient-centric measure.

How is Net Treatment Benefit different from utility-based approaches?

Utility-based approaches, such as CUI and MCDA, combine outcomes using numerical weights that specify the relative importance of efficacy, safety or other endpoints. NTB does not require these explicit weights. Instead, sponsors define a hierarchy of outcomes based on clinical importance, and the analysis evaluates which treatment performs better across all patients.

Can Net Treatment Benefit incorporate patient-reported outcomes?

Yes. Patient-reported outcomes, such as quality of life, fatigue or symptom burden, can be incorporated into the endpoint hierarchy used for GPC and NTB analyses. This allows sponsors to include the patient perspective alongside traditional efficacy and safety measures.

Is Net Treatment Benefit only useful for Project Optimus?

No. NTB can also be used to evaluate the totality of evidence in clinical trials, especially when multiple outcomes contribute to the overall treatment effect. Sponsors use NTB to assess efficacy, safety and quality of life together, support regulatory discussions, and provide a more patient-centric assessment of treatment value.