Biotech and pharma sponsors struggle when dose decisions are made late, rest on narrow evidence, or require protocol amendments that burn time and capital. The fastest way to change that trajectory is to make dose clarity a design objective across early development and to operationalize how your teams generate, interpret, and act on exposure–response signals.This playbook lays out how to build that capability and execute it at sites, as well as how to measure progress in ways that matter to executive teams. It draws on current best practices, with select elements summarized in a recent IQVIA white paper on combination therapy dose optimization in oncology.
Build the capability that turns models into dose decisions
Dose optimization often falters when study strategy and daily delivery come apart. In combinations, overlapping toxicities can mask which agent is driving adverse events, exposure–response patterns can plateau or invert as doses rise, and sites may follow different sampling or imaging schedules. These issues distort the dose–response signal and make escalation choices unreliable.Designing effective and efficient combination dose-finding trials is explored in great detail in a white paper prepared by IQVIA, a clinical research organization. Drawing on the key design, operational, and analytic principles laid out in the paper, the recommendations below translate them into a sponsor playbook that emphasizes decision quality and feasibility. Before prescribing any tactics, align on what “good” looks like for sponsor teams and put dose clarity at the center. That means connecting preclinical and clinical PK and PD, and using mechanistic insights only where they explain observed variability. With inputs aligned, clinical development and quantitative pharmacology teams can estimate exposure–response across patient types and defend dose choices in protocol and at submission.How do you do that? Start by choosing designs that map the therapeutic surface rather than a single axis. For dual‑novel combinations, a two‑dimensional approach reduces blind spots. First, optimize each agent as a monotherapy in the same population, then run parallel cohorts that escalate agent A while holding B constant, and separately escalate B while holding A constant. This sequence exposes interaction effects that single‑axis designs can miss and shortens the path to a defensible dose before expansion.When adding a novel agent to an approved therapy, control what you can to keep interpretation clean. Maintain the approved agent at its effective dose so you can attribute efficacy and safety to the investigational drug without ambiguity. Where immune‑related responses are delayed or variable, organizations use randomized dose‑finding and patient‑reported outcomes to sharpen selection. Define timing windows and backfilling rules so cohorts remain informative as new data arrive.Keep in mind that with many targeted and immuno‑oncology agents, higher does not always mean better. Efficacy may plateau before toxicity escalates, and the optimal dose can sit below the maximum tolerated dose (MTD). Planning dose ranges and escalation rules that reflect this biology reduces amendments and protects tolerability. Anchor designs to the decision questions that matter to governance. If your objective is to establish a recommended phase 2 dose, build contrasts that resolve exposure–response on the intended timeline. If your objective is to advance a combination, first optimize monotherapy, then use parallel or staggered escalation to reveal interaction effects without overwhelming operational capacity. This alignment turns modeled insights into decisions your review committees can endorse.
Operationalize for signal quality across sites and time
Strong designs fail when inputs are noisy, late, or inconsistent. The most common sources of noise are practical: sample draws that miss protocol windows, imaging intervals that vary by site, lab methods that differ across regions, and delays that push interim analyses past cohort decision points. Each of these erodes the interpretability of the exposure–response relationship.One surefire way of minimizing that risk is through feasibility planning that matches the cadence of dose decisions. Align site selection, enrollment targets, and central lab capacity with the pace of safety lead‑ins, escalations, and backfilling. Restrict the number of initial sites where appropriate, then pre‑approve the next wave so activation can scale quickly once lead‑in endpoints are met, or the recommended phase 2 dose is determined.Write protocols that convert analytics into controlled actions. Include screening rules to drop unsafe or clearly ineffective doses and enrichment rules that preserve power in remaining cohorts. Codify backfilling to expand informative cohorts and enroll molecularly defined subtypes when they clarify dose selection. Define decision thresholds and model validation steps in the protocol and governance checklists so everyone knows what information will trigger escalation, expansion, or pause.Resource biostatistics and data management to the pace of the study. Dose‑finding produces frequent interim looks. If programming, cleaning, and data locks trail cohort timelines, decisions arrive too late to guide the next move. Assign clear owners for safety narratives, exposure–response analyses, and data readiness to keep reviews on schedule.Match endpoints to the decision horizon. Complement early readouts, such as objective response rate, with measures such as progression‑free survival or response duration where feasible. For immuno‑oncology, plan for delayed toxicities and atypical response patterns. Specify windows and patient‑reported outcomes that inform dose choices without over‑reacting to early noise.Finally, partner with intent. Combination dose‑finding adds design and delivery complexity that experienced CROs can help navigate, including two‑dimensional escalation, parallel arms, and multi‑agent optimization at scale. Keep pharmacometrics, protocol governance, and product strategy in‑house to preserve accountability while using external delivery capacity to protect timelines and quality.
Make it measurable with milestones and portfolio metrics
Programs move faster when milestones are explicit and measurable. Translate your design into gates that turn modeling into selection:
Exposure–response clarity thresholds that trigger escalation, enrichment, backfill, or de‑escalation.
Safety stability criteria before expansion and before combining agents.
Validation readouts that greenlight the dose recommended for phase 2.
Balance ambition with feasibility from the start. Dose ranges and the number of combinations directly affect total sample size, resource demand, and the robustness of outcomes. The IQVIA analysis captures this trade‑off succinctly: calibrate dosing ranges and combination breadth so your study remains powered and interpretable within real operational limits.Consider seamless phase II/III strategies where they fit. You can use scenario planning to right‑size options before the first patient is enrolled, then lock choices into the protocol language and governance checklists so the study executes to plan. One study published in the Journal of the National Cancer Institute (2023) suggests that such designs can reduce the sample size required for dose optimization while preserving decision quality. Track indicators that connect operational rigor to clinical and financial results:
Leading indicators: time from first patient dosed to RP2D determination, proportion of cohorts requiring backfill, rate of dose‑limiting toxicities per escalation step, time from data cut to decision meeting.
Lagging indicators: amendment rates due to dosing, discontinuation rates attributable to toxicity at the chosen dose, and post‑marketing dose changes, where applicable.
Context matters across the portfolio. According to the Kuic Research report, combinations that include immuno‑oncology agents are projected to reach at least 18 billion dollars globally by 2028. In crowded categories, faster, clearer dose decisions become a differentiator that compounds across programs.
Regulatory expectations confirm the shift away from MTD
Current guidance encourages sponsors to move beyond the maximum tolerated dose as a default. For a clinician’s perspective on why legacy MTD paradigms underperform for targeted and IO agents, and why AI for dosing is still early but advancing, read this article on optimizing oncology drug dosing.The emphasis is on characterizing dose–response and safety early, generating randomized assessments across a range of doses, and beginning designs with data‑ and model‑informed scenario planning. The intent is to improve tolerability and overall regimen quality while minimizing the time impact of optimization through fit‑for‑purpose study design and delivery. These themes, summarized in the IQVIA white paper, align scientific rigor with practical execution and reinforce the case for building this capability now.For sponsors, the implication is straightforward. Dose selection is no longer a late‑stage clean‑up. It is a front‑loaded requirement that demands scientific, operational, and governance readiness across phase 1 and phase 2.
Commit now to a right‑first‑dose capability
Sponsors that institutionalize this approach will make faster, clearer, and more credible dose decisions. The near‑term actions are concrete:
Start with one or two priority assets and lock dose decision criteria into the protocol.
Harmonize sites and labs to protect the exposure response signal, and staff biostatistics and data management to the cadence of cohort reviews.
Engage experienced partners for complex escalation while keeping pharmacometrics and governance in-house.
Track a focused set of metrics that tie investment to decision speed, tolerability, and amendment risk.
Remember that each well‑run study improves the next. Over a few cycles, dose optimization becomes a habit, and your organization learns faster, de‑risks registrational paths, and strengthens the case for its therapies in the clinic and the market.
