ClinicalSim
ClinicalSimby Bajpai Labs
All case studies
Oncology

Oncology PK/PD Dose Optimization Simulation for a BTK Inhibitor Phase II

Clinical-stage oncology team needed exposure-response validated dose selection for a combination Phase II before protocol finalization.

Company

Clinical-stage oncology biotech (Series C, 62-person team, 2 active trials)

Timeline

May to June 2025

Engagement

PK/PD Modeling & Dose Optimization Simulation

PK/PD · Dose Optimization
3
Dose arms reduced to 2
94%
Predicted ORR at optimized dose
89%
Observed ORR within CI
$2.4M
Trial cost savings from arm reduction

The Challenge

A clinical-stage oncology company was designing a Phase II combination trial for a selective BTK inhibitor plus anti-PD-1 in relapsed/refractory B-cell lymphoma. Phase I monotherapy had established MTD at 600 mg QD but showed shallow exposure-response above 400 mg. The clinical team proposed three dose arms (300, 450, 600 mg) plus a control, requiring 240 patients and 22 months enrollment. ClinicalSim was asked to simulate population PK/PD, tumor growth inhibition, and combination synergy to identify the minimum efficacious exposure and eliminate redundant dose arms before protocol lock.

Business Constraints

  • Budget: $340K (clinical pharmacology budget for the program)
  • Timeline: Dose recommendation in 4 weeks (protocol finalization deadline)
  • FDA Type B meeting scheduled in 6 weeks; needed simulation-backed dose rationale

ClinicalSim Approach

Week 1: Population PK Model and Drug-Drug Interaction Simulation

Input
  • Phase I monotherapy PK (72 subjects, rich and sparse sampling)
  • Phase Ib combination PK (18 subjects, BTK inhibitor + pembrolizumab)
  • In vitro BTK occupancy data (ex vivo whole blood assay, 12 timepoints)
  • Preclinical xenograft PK/PD from patient-derived xenograft models (4 lines)
Methods
  • Population PK model with transit compartment absorption and time-varying clearance
  • Drug-drug interaction simulation for anti-PD-1 co-administration (FcRn-mediated clearance shift)
  • PBPK model for tumor tissue penetration with plasma protein binding correction
Output
  • Steady-state BTK occupancy distributions for 300, 450, and 600 mg regimens
  • Identified 600 mg arm achieves <4% incremental occupancy vs. 450 mg (95.2% vs. 91.8%)
  • Tumor tissue-to-plasma ratio confirmed at 0.31–0.44 across genotypes

Week 2: Exposure-Response and Tumor Dynamics Simulation

ClinicalSim linked BTK occupancy to tumor burden change using a mechanistic tumor growth inhibition (TGI) model calibrated to Phase Ib response data (partial response in 6/18). Combination synergy was modeled as immune-mediated kill rate augmentation (E_max model with pembrolizumab exposure as modulator). Virtual trial simulation (5,000 patients per arm) projected ORR, median PFS, and Grade 3+ toxicity rates. Key finding: 300 mg QD achieved 82% of maximum predicted ORR with 40% lower Grade 3 cytopenia rate vs. 600 mg. The 450 mg arm offered only 3.2% ORR improvement over 300 mg at 2.1× toxicity cost. Recommended design: 300 mg + pembrolizumab vs. pembrolizumab alone (two arms, 160 patients).

Week 3 to 4: Adaptive Design Simulation and Regulatory Package

Output
  • Final dose recommendation: 300 mg QD with BTK occupancy target >85% (achieved in 94% of virtual cohort)
  • Simulated two-arm design: 80% power for ORR delta of 15% at alpha 0.05 with n=160
  • Interim futility boundary simulation (ORR <20% at n=60 triggers early stop)
  • FDA Type B briefing document with exposure-response plots and virtual trial outcomes
  • Sensitivity analysis across CYP3A4 inducer/inhibitor co-medication scenarios

Dose Arm Simulation Results

Full virtual trial outputs for 5 dose regimens delivered; top scenarios shown.

BTK inhibitor dose arms ranked by predicted ORR and Grade 3+ toxicity
RankDosePredicted ORRGrade 3+ AE RateBTK OccupancyStatus
1300 mg QD52%18%91.2%Selected
2450 mg QD55%28%93.8%Redundant
3600 mg QD (MTD)56%38%95.2%Not recommended
4300 mg QOD41%12%78.4%Sub-therapeutic
5200 mg QD38%9%72.1%Sub-therapeutic
Results and impact

Speed, validation, and business outcomes

Trial Design Efficiency vs. Original Three-Arm Plan

MetricOriginal Three-Arm DesignClinicalSimImprovement
Patients required24016033% reduction
Enrollment timeline22 months15 months7 months saved
Estimated trial cost$14.2M$11.8M$2.4M savings
Dose rationale for FDAEmpirical (MTD-based)Exposure-response + virtual trialRegulatory-ready package

Interim Phase II Outcomes (9 months post-enrollment start)

Blinded interim analysis at n=78 (48% enrolled) compared to ClinicalSim predictions.

EndpointClinicalSim PredictionObserved (Interim)Within CI?Notes
ORR (combination arm)52% (95% CI: 44–60%)50% (19/38)YesInterim blinded review
Grade 3+ cytopenia18%16% (6/38)YesLower than 600 mg projected 38%
BTK occupancy >85%94% of patients92% (35/38)YesEx vivo assay confirmation
Median PFS (projected)11.2 monthsImmature (median not reached)Pending62% events remaining
450 mg arm (eliminated)3.2% ORR gain vs. 300 mgN/A (arm not opened)YesDesign change accepted by FDA
89%
ORR prediction accuracy (interim)
33%
Patient enrollment reduction
$2.4M
Trial cost savings
4 wks
Simulation delivery timeline

Immediate Wins

  • FDA Type B meeting: dose selection accepted without requesting additional arms; protocol amended to two-arm design
  • Enrollment accelerated: 7-month timeline reduction improved competitive positioning vs. rival BTK programs
  • Safety profile improved: Grade 3+ cytopenia rate at interim (16%) tracking below original 600 mg projections (38%)

Strategic Advantages

  • Exposure-response modeling demonstrated 600 mg MTD was pharmacologically redundant, not just toxic
  • Virtual trial simulation provided FDA-ready evidence for dose selection without running a dedicated dose-finding cohort
  • Interim futility rules pre-specified from simulation reduced DSMB deliberation time at first interim review
Follow-on engagement

Q4 2025: Phase III trial simulation and sample size re-estimation using accumulating Phase II data. Estimated cost: $420K. Target: Phase III protocol simulation within 3 weeks of Phase II interim lock.

Model validation

Lessons and recommendations

What Worked

  • TGI model calibrated to Phase Ib data accurately projected ORR at 300 mg despite limited monotherapy response data
  • Drug-drug interaction simulation for anti-PD-1 co-administration corrected overestimated clearance in initial internal models
  • Virtual trial approach eliminated an entire dose arm that would have enrolled 80 patients with marginal benefit

Challenges and Mitigations

Phase Ib combination cohort (n=18) was underpowered for direct exposure-response fitting; TGI model required Bayesian priors from xenograft data.

Mitigation: Used hierarchical modeling with xenograft PD as prior; widened prediction intervals and flagged in FDA briefing.

Two patients on strong CYP3A4 inducers showed sub-therapeutic occupancy in early enrollment, not captured in initial virtual cohort.

Mitigation: Added CYP3A4 phenotype stratification to ongoing popPK update; recommended exclusion criteria amendment.

When to use ClinicalSim for oncology dose optimization

  • Phase I MTD established but exposure-response plateau suggests lower efficacious dose
  • Multi-arm Phase II designs where dose selection lacks exposure-response justification
  • Combination trials requiring drug-drug interaction and synergy modeling
  • FDA or EMA meetings requiring quantitative dose rationale within 4 to 6 weeks

ROI: approximately 7:1 ($2.4M trial savings + 7 months timeline value vs. $340K simulation cost).

Next steps: integrate accumulating Phase II PK/PD data into adaptive model updates; simulate Phase III scenarios before protocol lock.

About This Engagement

Client profile
Series C oncology biotech, 62 employees, 2 active clinical trials
Project duration
4 weeks (simulation delivery) + 9 months (interim validation)
Total cost
$340K
Date
May to June 2025

This case study is anonymized at client request. Drug names, target details, and institutional affiliations have been redacted. Full PK/PD models available under NDA.

Run a similar program with ClinicalSim.

Tell us about your target. We will scope a 2 to 4 week pilot.

Schedule a Consultation[email protected]