Transcript open

Optimal phase 2 dose selection based on the relationship
between exposure and target occupancy
Mona Alameddine(1), Christophe Boetsch(1), Ronan Carnac(1), Patricia Sanwald-Ducray(1) and Nicolas Frey(1)
(1) Roche Pharma Research & Early Development, Clinical Pharmacology, Roche Innovation Center Basel.
Introduction
Objective
Target occupancy (TO) is a useful metric to confirm drug engagement with
the target in early clinical phase and to potentially predict the expected
clinical response in the target patient population.
To support optimal dose selection for Phase 2, using simulations from a
population pharmacokinetic (popPK) model and an exposure-TO model built
on data collected in Phase 1.
For most first in class compounds, the relationship between TO and clinical
response is largely unknown in early development resulting in challenges for
optimal Phase 2 dose selection.
Methods
Using Phase 1 PK data collected in single and multiple ascending doses in healthy volunteers, a popPK model was built using NONMEM ® (Version 7.2.0).
Using Phase 1 PET data in healthy volunteers an exposure - TO model was built using Phoenix® NLME (Version 1.2). By combining those 2 models, 24-h
steady-state TO time profiles were simulated for 7 different doses.
For each dose, the percentages of hourly TO measurements for three non-overlapping predefined ranges of TO (10-30%, 30-50%, 50-80%) were calculated
and used to select the dose strength and the number of doses to be investigated in Phase 2 to ensure optimal coverage of selected TO ranges and to
adequately characterize the TO-clinical response relationship.
Results
Rich pharmacokinetic (PK) data from 95 individuals were adequately described by a two-compartment population PK model with first order absorption and
elimination. Body weight significantly impacted Inter-compartment clearance and peripheral volume while age significantly affected peripheral volume. A direct
Emax model (Fig.1). adequately described the PK-TO relationship using PK and PET scan data from 9 individuals. TO simulations (Fig.2) were performed using
a distribution of covariate from a historical trial in a similar patient population in order to properly account for the covariate effect on PK (Table 1). Furthermore,
by using graphical illustration (Fig.3), it was determined that with the expected between patient variability in Phase 2 at least 3 doses should be investigated to
get an appropriate coverage of the overall TO range compared to 2 doses.
Predicted TO (%)
Fig.1 Target Occupancy/Exposure relationship
20-50% inhibition
60-80% inhibition
Concentration (ng/mL)
Fig.1 TO/Exposure relationship was successfully characahterized and best described using
an Emax model
Table 1
Dose
Percentage of hourly simulated measurements under TO ranges within the dosing
interval
10-30% TO
30-50% TO
50-80% TO
Dose 1
52
1
0
Dose 2
69
25
1
Dose 3
29
53
18
Dose 4
10
48
42
Dose 5
4
34
61
Dose 6
1
15
80
Dose 7
0
1
60
Fig.3 Coverage of Target Occupancy achieved with 2 doses compared to that
achieved with 3 doses
Fig.2 Simulated Target Occupancy time profiles at Steady State
Fig 2: The optimal dose for each TO range was selected based on the highest number of
hourly simulated time points within each TO range. shaded area that represents the 95%
prediction Interval, (See Table 1)
Fig.3: Histograms illustrate the coverage of TO comparing two versus three doses. It is
evident that exploration of the full TO range is best supported by the investigation of 3 rather
than 2 doses.
Conclusions
Phase 1 data was successfully leveraged and translated into useful graphical representations to support the identification of optimal doses and number of
doses to be evaluated in Phase 2 to characterize the TO-efficacy relationship to ensure appropriate dose selection for confirmatory trials.