Dose selection is one of the most consequential decisions in early-phase oncology drug development. Carrying the wrong dose into Phase II can compromise efficacy, increase toxicity, and delay an entire program. The Recommended Phase 2 Dose, or RP2D, is the dose derived from Phase I data and intended for continued clinical development.
- What RP2D Means in the Context of Phase I Dose Optimization
- Why the Traditional MTD-Based Approach Is No Longer Sufficient?
- The FDA’s Project Optimus and Its Impact on RP2D Clinical Trials
- How RP2D Is Determined in Modern Trial Designs?
- Key Endpoints That Inform RP2D Selection
- Regulatory and Operational Considerations in RP2D-Driven Trial Design
- Conclusion
Recent evidence highlights why this decision requires more than early toxicity signals alone. A systematic review of Phase I oncology trials found that approximately 26.5 percent of studies identified new toxicities during dose expansion that were not observed during initial dose escalation.
This blog explains what RP2D means in modern clinical development, why traditional MTD-based approaches are no longer sufficient, and how RP2D trial design integrates PK, PD, safety, and early efficacy data.
What RP2D Means in the Context of Phase I Dose Optimization
The Recommended Phase 2 Dose (RP2D) is the dose selected at the conclusion of Phase I to be administered in Phase II. It represents the investigators’ and sponsors’ determination of the dose that best balances efficacy potential, safety, and tolerability for the intended patient population.
Historically, RP2D was treated as equivalent to the Maximum Tolerated Dose (MTD). For cytotoxic chemotherapies, this held reasonable logic: higher dose meant greater tumor kill, and the limit was toxicity. With the rise of molecularly targeted agents (MTAs) and immunotherapies, that assumption no longer holds. These agents can reach target saturation well below the MTD, meaning further escalation adds toxicity without adding benefit.
The result: RP2D is now a multidimensional determination, not simply the dose one step below the unacceptable toxicity level.
The Core Elements That Define RP2D
| Element | Description |
| Dose-Limiting Toxicity (DLT) | Grade 3 or 4 adverse events in cycle 1 that trigger dose de-escalation. |
| Pharmacokinetics (PK) | Drug exposure (AUC, Cmax) and clearance at each dose level. |
| Pharmacodynamics (PD) | Biological effect at the target (receptor occupancy, biomarker modulation). |
| Preliminary Efficacy | Objective response or disease control data from expansion cohorts. |
| Long-Term Tolerability | Cumulative toxicity, dose reductions, and treatment discontinuations across cycles. |
Why the Traditional MTD-Based Approach Is No Longer Sufficient?
The 3+3 rule-based dose escalation design was built around a straightforward dose-toxicity model. Enroll three patients, observe grade 3 or 4 toxicities, and escalate or de-escalate per a predefined rule set. The RP2D is the highest dose at which fewer than a prespecified proportion of patients experience a DLT.
Several structural limitations are documented when this approach is applied to modern therapeutics:
- Single-cycle toxicity window: The 3+3 design captures toxicity data only from cycle 1. For targeted agents administered chronically, cumulative toxicities appearing in cycles 3 through 6 are not factored into RP2D determination.
- Inadequate dose-response characterization: The design escalates until toxicity rather than until optimal biological activity. For agents with flat PK-PD relationships at higher doses, this results in an RP2D that may be unnecessarily high.
- Binary DLT endpoint: Grade 3 to 4 events are captured; grade 1 to 2 chronic toxicities that reduce patient adherence are not.
Research published by Quanticate (2025) found that in 54% of trials where the expansion phase was designed primarily to evaluate safety, new toxicities were identified that had not been observed during dose escalation alone. This figure shows how often the 3+3 approach produces an incomplete safety picture at the point RP2D is declared.
This limitation is not trivial. When the RP2D entering Phase II has an insufficiently characterized safety profile, sponsors face unplanned protocol amendments, dose-modification requirements, and potential FDA feedback that disrupt timelines.
The FDA’s Project Optimus and Its Impact on RP2D Clinical Trials
The FDA’s Oncology Center of Excellence launched Project Optimus in 2021 to reform dose optimization practices. The final guidance, issued in August 2024, formalizes the expectation that dose selection must be evidence-based rather than toxicity-driven alone.
Key requirements under Project Optimus for RP2D determination:
- Evaluation of at least two dose levels in a randomized dose-optimization stage before declaring RP2D.
- Integration of PK, PD, safety, and anticancer activity data as a collective basis for dose selection.
- Inclusion of patient-reported outcomes (PROs) to capture tolerability dimensions not reflected in CTCAE grading.
- Exposure-response analyses to characterize the dose-response relationship directly, rather than assuming it is monotone.
- Early FDA engagement through pre-Investigational New Drug (pre-IND) meetings and formal milestone discussions.
For sponsors developing MTAs or immunotherapies, a Phase I study designed around 3+3 escalation to MTD, followed by RP2D declaration one step below, is unlikely to satisfy regulatory expectations at submission review.
How RP2D Is Determined in Modern Trial Designs?
Phase I trials with a robust RP2D strategy operate in two structured sub-phases: dose escalation and dose expansion. Each contributes different types of evidence to the final decision.
Dose Escalation Phase
The escalation phase generates the safety envelope. Starting from a preclinically derived safe dose, cohorts receive incrementally higher doses according to predefined escalation rules. Modern designs used in RP2D clinical trials include:
- Bayesian Optimal Interval (BOIN) design: Assigns each dose level to one of three intervals (escalate, stay, or de-escalate) based on the observed DLT rate. Widely adopted in US oncology centers for its accuracy and operational manageability.
- Continual Reassessment Method (CRM): A fully model-based approach where the dose-toxicity curve is updated after each patient’s DLT data are observed. More dose-efficient than 3+3.
- Accelerated Titration Design: Allows single-patient cohorts during initial escalation, moving to three-patient cohorts upon DLT observation. Reduces the number of patients treated at sub-therapeutic doses.
The output is an MTD or Maximum Administered Dose (MAD), along with preliminary PK data and early safety characterization.
Dose Expansion Phase
The expansion phase refines RP2D selection and builds the evidence base for Phase II initiation. Cohorts are enrolled at one or more doses of interest to evaluate:
- Preliminary antitumor activity (objective response rate, disease control rate),
- PD biomarker modulation at the target.
- Longer-term tolerability and cumulative toxicity across multiple cycles,
- Subpopulation response patterns by histology or molecular marker,
Under Project Optimus, the expansion phase also supports randomized dose-comparison where more than one dose level warrants further evaluation.
Key Endpoints That Inform RP2D Selection
RP2D is not a single-endpoint determination. These endpoint categories should be prospectively specified in the trial protocol:
Safety and Tolerability
- DLT rate in cycle 1
- Grade 1 to 2 cumulative toxicities across cycles
- Dose reductions, interruptions, and discontinuations beyond cycle 1
Pharmacokinetics
- AUC, Cmax, drug half-life, and accumulation with repeated dosing.
- Dose linearity and saturation occur at higher dose levels
Pharmacodynamics
- Receptor occupancy or target inhibition across dose levels.
- Correlation between PK exposure and PD response.
Preliminary Efficacy
- Objective response rate at each dose level in expansion cohorts.
- Duration of response as a function of dose.
The objective is to identify the dose at which the PD endpoint is maximally and consistently achieved without adding incremental toxicity burden.
Regulatory and Operational Considerations in RP2D-Driven Trial Design
Sponsors preparing Phase I programs under current regulatory expectations need to account for the following:
- Protocol Design Rigor. The RP2D declaration must be supported by a prospectively planned analysis strategy. Ad hoc post-hoc justification is scrutinized by the FDA during the IND and NDA, or Biologics License Application (BLA), review stages. Protocol amendments that retroactively change DLT definitions raise questions about the scientific integrity of the declared dose.
- Biomarker Planning. For MTAs, a predefined PD biomarker strategy is a regulatory expectation. Sponsors must specify which biomarkers define optimal biological dosing, how samples will be collected, and how PD results will factor into RP2D decision criteria before the study begins.
- Adaptive Design Integration. Smooth Phase I/II designs, in which RP2D is declared at an interim decision point, and Phase II enrollment, require pre-specified decision criteria to avoid inflating Type I error. These designs should be discussed and aligned at the pre-IND or Scientific Advice stage well before enrollment begins.
Conclusion
The Recommended Phase 2 Dose (RP2D) clinical trial has evolved from a straightforward toxicity threshold into a multidimensional clinical and pharmacological judgment. The FDA’s Project Optimus framework, finalized in 2024, codifies the expectation that dose selection must integrate PK, PD, safety, and efficacy data within prospectively planned analyses.
Sponsors who design Phase I programs around these expectations reduce regulatory uncertainty, strengthen the data package entering Phase II, and position their programs for more defensible submissions. The scientific and operational rigor applied at the RP2D stage is one of the clearest predictors of late-phase program success.
