TL;DR – The FDA’s new push to rethink how optimal doses of cancer drugs are determined will lead to more effective drugs with fewer side effects, developed in a more capital-efficient way.

The FDA recently introduced Project Optimus as a direct response to drug companies shifting development focus away from chemo-like drugs towards more nuanced precision medicines. Unlike chemo, these tend to work more narrowly – acting only on pathogenic cells or proteins, while leaving normal ones alone. However, this also means that once the drug saturates all its targets, efficacy plateaus no matter how much more of it is administered. Side effects rise with higher doses, so naturally, when drug developers use high dosing in later-stage clinical trials, they run the risk of unacceptably high toxicity to the patient and, consequently, program failure.

Failure of a drug that would otherwise have provided material benefit to patients – but for dosing mistakes in clinical trials – is exactly what Project Optimus attempts to avoid. This push by the FDA is net good – for patients, biotech startups, big pharma, and investors alike. The new guidance forces drug developers to more deliberately consider the risk of taking the wrong dose forward into later-stage clinical trials and the huge cost of flaming out just before the finish line – both in capital invested and in losing out on potentially life-saving options for patients.

I was quoted briefly by the Wall Street Journal on this topic recently and wanted to flesh out a bit more the impact of – and response to – one of the momentous shifts in regulatory guidance of this decade. While dose-finding now requires more planning as a result of Project Optimus, biotech startups (and big pharma) are responding through clinical trial innovations:

• Adaptive study designs – statistically rigorous ways for selecting the next higher drug dose to test, using all the clinical data generated up to that point (unlike the 3- to 6-patient snapshots used in the traditional “3+3” design);
• Accelerated titration – moving faster through the required low, sub-therapeutic doses by enrolling a single patient in each (compared to the historical three patients);
• Backfilling – once a dose is shown to be safe and pharmacologically active, enrolling additional patients at that dose to generate more data; and
• Exposure-outcome modeling – determining the relationship between dose increases and their impact on tolerability and efficacy.

Some of these approaches offer finer control when setting the toxicity rate threshold. In a traditional “3+3” design, the rate is hard-coded at 33% – meaning two out of six patients are unable to tolerate a given dose. In contrast, adaptive designs can set any threshold, with the much lower ~22% being the most common, and can thus better avoid taking too toxic a dose forward into expensive later-stage trials. Other methods allocate the same number of patients more effectively into doses that are pharmacologically active and, as a result, more informative.

These approaches thus not only minimize the risk of selecting the wrong dose but can also be just as capital-efficient as the old way – simply by reallocating the same number of patients within a trial to greater effect. For any company, large or small, having more informative data for the same capital invested is always better…and is well worth the greater upfront planning.

But this is a long game. Biotech and pharma players are just starting to adjust to these changes now. Once they become the new normal, we will see better outcomes for companies, investors and, most importantly, patients.

I plan to use this space to go deeper on the developments which stand to have the greatest impact on innovation in biotech. I hope you’ll follow along and send me a note if you want to discuss anything I share here: mariana@venrock.com.