Disease Models Overview and Case Studies

1
Disease ModelsOverview and Case Studies

• Joga Gobburu
• Pharmacometrics
• Office Clinical Pharmacology,
• Office of Translational Sciences, CDER, FDA

2
Pharmacometrics Survey

• Between 2000-2006, 72 NDAs needed Pharmacometrics
  Reviews/Analyses
• For each of the Pharmacometrics Reviews, the
  customers were asked to rate the impact on
  approval related and labeling decisions
• Pivotal Decision would not have been the same
  without Pharmacometrics analysis
• Supportive Decision was well supported by the
  Pharmacometrics analysis
• No Contribution No need for the Pharmacometrics
  analysis

3
Impact of Pharmacometrics Analyses 2000-2004
Pivotal Regulatory decision will not be the same
without PM reviewSupportive Regulatory decision
is supported by PM review
Bhattaram et al. AAPS Journal. 
2005 7(3) Article 51. DOI  10.1208/aapsj070351
4
Pivotal Regulatory decision will not be the same
without PM reviewSupportive Regulatory decision
is supported by PM review
Impact of Pharmacometrics Analyses 2005-2006
DCPDivision of Clinical Pharmacology _at_survey
pending in 1 case
5
NDA1 Approval of monotherapy oxcarbazepine in
pediatrics for treating partial seizures using
prior clinical data

• FDA/Sponsor pursued approaches to best
• utilize knowledge from the previous trials to
• assess if monotherapy in pediatrics can
• be approved without new controlled trials

6
NDA2 Establishment of biomarker-outcome
relationship allowed more efficient future trial
design

• The sponsor was pursuing an accelerated approval,
  for drug to prevent a life-threatening disease,
  based on a biomarker even though clinical
  endpoint analysis failed in two pivotal trials

7
NDA2 Establishment of biomarker-outcome
relationship allowed more efficient future trial
design
Relative risk of the disease event
Hazard ratio10.0 (95 CI 2.5-30.0) pRatio of biomarker level to baseline
8
NDA3 Insights into trial failure reasons will
lead to more efficient future trials
Severe Baseline Disease Responders
Mild Baseline Disease Non-Responders
9
Females seem to be more sensitive to QT
prolongation
Slope
Slope
Slope
Slope
10
Need/Opportunities for Innovative Quantitative
Methods in Drug Development
Optimal design to show disease modifying
effects?
Good marker(s) of survival benefit in cancer
patients?
Maximize the change of success of a 2yr obesity
trial?
Given 85 of depression trials fail, how to
improve success?
Best dose for a 26wk trial based on 12 wk data?
Providing solutions for these issues calls for
efficient use of prior knowledge
11
Manage and Leverage Knowledge
Information

• Biomarker-Endpoint
• Time course
• Drop-out
• Inclusion/Exclusion
• criteria (Trial)

Placebo Disease Models

• Parkinsons
• Obesity, Diabetes
• Tumor-Survival
• Rheumatologic condition
• HIV
• Epilepsy
• Pain

Knowledge
We are referring to such diverse quantitative
approach(es) as Disease Modeling
12
Core Development Strategy for Testosterone
Suppressants
IC50
Reporter Gene Assay
- Early screening of compounds based on IC50
value. - High thrput method to filter thousands
of compounds - Based on prior experience, a few
potential entities will be selected for the next
phase
Preclinical
PKPD data
Disease Model
- In vitro IC50 as a guide for preclinical dose
selection - Animal models to measure all
possible biomarkers e.g. GnRH, LH, T and Drug
conc.
Clinical Trial Simulation
Dose optimization in cancer patients
PKPD data
- Invitro and preclinical data for clinical dose
and regimen selection - Clinical development plan
- Pilot study for dose optimization thr
innovative trial designs
Pivotal trial
----2 mo-----
----2 mo-----
----2 mo-----
----3 mo-----
---------12 mo--------------
Actual execution time.- it does account for time
spent accumulating resources.
From Pravin Jadhav, VCU/FDA
13
Obesity

• Obesity trials are large, over 1-2 yrs and
  fraught with challenges due to high drop-out rate

Dr. Jenny J Zheng Dr. Wei Qiu Dr. Hae Young Ahn
14
Obesity
Model Qualification
Baseline Body Weight 3000 patients
15
Patients with small weight loss drop-out
Drop-out patients
Remaining patients
0-12
36-52
12-24
24-36
16
Obesity Time Course of Placebo Effect
17
Value to Drug Development

• Effective use of prior data for designing future
  registration trials
• Might lead to alternative dosing considerations
• Titration vs. fixed dose
• Could lead to increased trial success
• Allows of designing useful shorter duration
  trials for future compounds for screening and
  initial dose range selection

18
Diabetes

• How to reliably select doses for registration
  trials based on abbreviated dose finding trials
• Need arose from an EOP2A meeting
• Work in progress No patient population and
  drop-out models yet.

Drs. Vaidyanathan, Ahn, Yim, Zheng, Wang,
Gobburu, Powell, Sahlroot, Orloff
19
Pivotal Trial Dose Selection Anti-Diabetic

• Sponsor conducted 12 wk dose ranging trial in
  diabetics
• Key Regulatory Question
• What is a reasonable dose range and regimen for
  the pivotal trial(s)?
• Challenge
• Estimate of effect size on HbA1c at 26 wks not
  available. Effect size on FPG available.

20
FPG-HbA1c relationship from historic
studies employed to estimate effects on HbA1c of
the new compound
Drug Conc.
FPG
HbAlc
Time (Week)
Jusko et al
21
Biological relationship between FPG-HbA1c bridged
information gap


Drug X (other) in 28 patients
Hybrid dataset in 100 patients
Drug X (Sponsor) in 72 patients
22
Value to Drug Development

• More informed dose/regimen selection
• Could lead to increased trial success
• Quantitative analysis was critical
• Effective use of prior data for predictions
• Supports conduct of useful shorter duration
  trials for future compounds

23
Disease Models Challenges

• Data Management
• How to best maintain an efficient database?
• Analysis
• How to best conduct meta-analysis?
• Identify and fill gaps (time-varying biomarkers
  in survival models)?
• Inter-disciplinary collaboration
• Biologists, Pharmacologists, Statisticians,
  Disease Experts, Quantitative Clinical
  Pharmacologists, Engineers need to come together
  to develop these models as a team.