When FDA published the guidance, Optimizing the Dosage of Human Prescription Drugs and Biological Products for the Treatment of Oncologic Diseases in August 2024, I read it with great interest. At first, I thought that it was a draft guideline, as this was the first time I had seen a guideline addressing dose selection in the oncology setting. But I quickly realized that this final guidance implements the information gathered in the FDA’s Project Optimus.
First a word about Project Optimus, an initiative of the FDA’s Oncology Center of Excellence (OCE). This initiative sought to reform dose selection and optimization in oncology drug development. As an outcome of this initiative, the long-standing method of determining the maximum tolerated dose (MTD) in oncology trials, the 3+3 design that was developed for cytotoxic drugs, and using the MTD (or a dose close to it), is no longer the recommended method. In many cases, the MTD method had resulted in the administration of doses which, in the long-term, produced unacceptable toxicities. This is particularly true for some of the newer targeted agents. This has resulted the use of doses in pivotal trials that were inadequately characterized, resulting in more toxicity without improvements in survival or quality of life. The resulting need for dose and/or schedule reductions or drug discontinuation can, unfortunately, result in the missed opportunity for additional benefit to the patient.
To aid in understanding the new guidance, the FDA has defined ‘dosage’ as the combination of the drug dose and the schedule of treatment. Schedule refers to the recommended interval between doses and the duration of treatment. Dose refers to the quantity of the drug. Footnote 2 of the new guidance states: “… an optimized dosage is a dosage that can maximize the benefit/risk profile or provide the desired therapeutic effect while minimizing toxicity.”
One of the most important recommendations of the new guidance is that all data be considered when selecting the optimal dosage(s). I was not surprised to see the suggestion that more than one experimental dosage should be studied in registration/phase 2 trials. FDA’s selection of the term dosage optimization as opposed to dose determination or selection is intentional; the guidance outlines the multiple characterizations and considerations that are part of determining optimal dosage.
Multiple sources of information should be measured in early clinical studies and considered in dosage optimization. These include drug class, patient population, and the characterization of the pharmacokinetics (PK) following a single dose and at steady state for each dose evaluated. The PK plan needs to support population PK and dose, and exposure-response analyses during the clinical program. These are not one-time analyses and should be started early and updated as additional data become available. This iterative process will allow the early identification of subpopulations in which there may be differences in exposure.
Where appropriate, pharmacodynamic (PD) and pharmacogenomic data should be included in these analyses. If an intrinsic factor such as a genetic variation or organ impairment is relevant for the indication, the impact of these on PK, PD, safety, and activity should be evaluated early in development. If applicable, other dosing strategies such as the use of priming doses (i.e., for products that are associated with early-onset, serious, or life-threatening toxicities such as cytokine release syndrome) or intra-patient dose modifications should be explored. The potential for drug interactions with concomitant medications relevant to the indication should be evaluated early. And, for orally administered drugs, the effect of food on PK and safety should also be evaluated.
When selecting the doses for future study, safety, tolerability, and activity should be compared across the dosages tested and should include additional items such as duration of exposure, number of patients who received all planned doses, dose modifications, SAEs and toxicities of interest. Use of patient‑reported outcomes (PRO), and engaging with patients and advocacy groups is also recommended in the guidance.
It is in the sponsor’s best interest to discuss dosage optimization with FDA early in the clinical program, prior to phase 2. Sponsors should also consider applying to the Model-Informed Drug Development (MIDD) meeting program, if applicable. Sponsors are cautioned that having breakthrough designation or being part of an accelerated program does not mean that the optimal dosage selection step can be skipped in the clinical development of a product.
Not unexpectedly, this guidance recommends that multiple dosages be compared in early trials. A recommended trial design is the randomized, parallel dose-response trial, with treatment blinding/masking where feasible. A sufficient number of patients should be included for the assessment of safety and antitumor activity. An adaptive design can also be implemented, which would allow sponsors to stop enrollment in one or more arms based on evolving data. While a crossover design may not be optimal, if it is used, the analysis should be pre-specified and should state how safety and activity will be assessed to account for the crossover. In all designs, trial stopping rules for excessive toxicity should be pre-specified.
Lastly, there is not a one-size-fits-all solution. For a single product, different indications may require different optimal dosages. All relevant data, nonclinical and clinical, as well as dose- and exposure-response relationships, should be considered. Computer modeling may also be useful in this situation.
In summary, rather than using the old paradigm of determining an MTD as quickly as possible by evaluating escalating doses, FDA is asking sponsors to take a more deliberate approach. Evaluate a number of dosages, collect PK, PD, safety, and activity data, review all available nonclinical and clinical data, and select the appropriate dosages for further study.
