Transcript CIPA Compound Selection Overview

CIPA COMPOUND
SELECTION OVERVIEW
Thomas J Colatsky
Director, Division of Applied Regulatory Science
OCP/OTS/CDER
US Food and Drug Administration
Drug Sets Needed
1) Training set
2) Validation set
3) Test set
http://www.freevector.com/site_media/preview_images/FreeVector-BeatingHeart-Graphics.jpg
Training Set: Purpose
http://www.freevector.com/site_media/preview_images/FreeVector-BeatingHeart-Graphics.jpg
• Evaluate the performance of the consensus
model (O’Hara-Rudy human ventricular myocyte)
• Identify any components that need to be improved
or revised to reflect known activities
•
•
•
•
Ion channels
Cellular homeostasis
Drug-channel interactions
Other (e.g. adrenergic tone?)
• Use revised model to assess:
• Candidate risk metrics
• Simulation protocols needed to evaluate relative
clinical proarrhythmia (TdP) risk
Training Set Requirements
• Must be well characterized in both (a) clinical outcomes of
TdP risk and (b) ion channel pharmacology
• Need both multi-channel and hERG specific drugs
• Range from no risk to high risk
• Scale TBD - use 5 Redfern categories as possible starting point?
• INaL, ICaL, IK1, IKr, IKs = currents of most interest
• Confirm (?)
• Need pharmacology data on each (mostly IC50s now)
• Kinetic information on channel block?
• Overlap with EJ Park’s protocol research?
• Consensus model(s) for drug-channel interaction?
• TARGET: 4-5 well-characterized drugs per risk category
for a total of 20-25 drugs
• Training quality will reflect completeness of data sets – if
incomplete, more data will be needed for a second round
Validation Set: Purpose
http://www.freevector.com/site_media/preview_images/FreeVector-BeatingHeart-Graphics.jpg
• Determine how well the fully trained model rank
orders drugs with varying proarrhythmic (TdP) risk
• Confirm adequacy of candidate ion channels
• Identify whether model, metrics and/or simulation
protocols need to be further improved or revised
• Finalize and make available:
• Consensus model (executable)
• Ion channels pharmacology requirements
• Risk definitions and scaling
• Risk metric performance characteristics
Validation Set Requirements
• Must be well characterized in both (a) clinical outcomes of
TdP risk and (b) ion channel pharmacology
• Need both multi-channel and hERG specific drugs
• Range from no risk to high risk
• Need pharmacology data on each of the required channels
• Are INaL, ICaL, IK1, IKr, IKs still the candidates of most interest?
• Standardized voltage clamp protocols
• TARGET: 4-5 well-characterized drugs per risk category
for a total of 20-25 drugs (could be fewer)
Test Set: Purpose
http://www.freevector.com/site_media/preview_images/FreeVector-BeatingHeart-Graphics.jpg
• Try to “break” the model
• Use relatively unknown drugs for which outcome data
(clinical/non-clinical) exist somewhere
• Individual companies can run the model and compare results
• CIPA In Silico group to be notified of “fails” and try to fix
• Intended to confirm the adequacy of the model,
pharmacology, metric and risk scale
• Finalize and make available:
• Consensus model and its performance characteristics
• Ion channels pharmacology requirements
• Risk definitions and scaling
• Risk metric performance characteristics
Test Set Requirements
• Must be well characterized in both clinical
outcomes of TdP risk and ion channel
pharmacology.
• Should include as many drugs as the
pharmacology community needs to gain comfort
and prove the concept.