The IVIVC Toolkit™ for WinNonlin - Pharsight
The IVIVC Toolkit™ for WinNonlin - Pharsight
The IVIVC Toolkit™ for WinNonlin - Pharsight
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
<strong>The</strong> <strong>IVIVC</strong> Toolkit <strong>for</strong><br />
<strong>WinNonlin</strong>®: Product<br />
Introduction and<br />
Demonstration<br />
Jason Chittenden<br />
Product Marketing Manager<br />
All contents Copyright ©2006 <strong>Pharsight</strong> Corporation. All rights reserved. This presentation may only be viewed <strong>for</strong> personal<br />
use. No part of this document may be reproduced or transmitted in any <strong>for</strong>m or by any means, electronic or mechanical,<br />
including photocopying, <strong>for</strong> any purpose, without the express written permission of <strong>Pharsight</strong> Corporation. <strong>Pharsight</strong>®,<br />
<strong>WinNonlin</strong>®, <strong>IVIVC</strong> Toolkit <strong>for</strong> <strong>WinNonlin</strong>, <strong>WinNonlin</strong> AutoPilot, WinNonMix®, <strong>Pharsight</strong> Knowledgebase Server,<br />
PKS, PKS Reporter, Drug Model Explorer®, DMX®, Trial Simulator and TS2 are trademarks or registered trademarks of<br />
<strong>Pharsight</strong> Corporation. All other brand and product names are trademarks or registered trademarks of their respective holders.
About Our Presenter<br />
• Jason is a graduate of:<br />
– Cali<strong>for</strong>nia State Polytechnic University in Pomona<br />
• B.S. in Chemical Engineering<br />
– University of Cali<strong>for</strong>nia at Irvine<br />
• M.S. in Biochemical & Chemical Engineering<br />
• Previously:<br />
– Simulation Sciences- Senior Engineer<br />
– Alliente- Software Development Engineer<br />
– Simulations Plus- Senior Scientist / Product Manager <strong>for</strong> GastroPlus<br />
• At <strong>Pharsight</strong>:<br />
– Product Marketing Manager<br />
– Located in Cary, NC<br />
25 October 2006<br />
2<br />
© 2006 <strong>Pharsight</strong> Corporation. All rights reserved.
Welcome<br />
• About today’s audience:<br />
– Pharmacokineticists<br />
– Formulators<br />
– Pharmacologists<br />
– Quality Assurance<br />
– Academics<br />
– Managers, Directors, CEOs<br />
• Flow of Discussion<br />
– <strong>WinNonlin</strong><br />
– <strong>IVIVC</strong> Toolkit<br />
– Sample Problem<br />
25 October 2006<br />
3<br />
© 2006 <strong>Pharsight</strong> Corporation. All rights reserved.
What is <strong>WinNonlin</strong>?<br />
• Industry leading PK/PD data analysis tool<br />
• Intuitive and easy-to-use spreadsheet style interface<br />
• Powerful and fast algorithms<br />
– Noncompartmental analysis of both blood and urine data using methods <strong>for</strong> rich or sparse datasets<br />
– Three methods <strong>for</strong> AUC calculation: linear trapezoidal, linear/log trapezoidal, and linear up/log down<br />
– Model Fitting / Parameter Estimation<br />
• Extensive built-in library of predefined models<br />
• User defined models<br />
– Deconvolution, semicompartmental modeling, nonparametric superposition, crossover design,<br />
bioequivalence<br />
– Analysis using general linear modeling approaches and ANOVA<br />
– Descriptive statistics, including weighted descriptive statistics<br />
• Flexible data management<br />
– Quick data export to NONMEM®, SAS, S-PLUS®, SigmaPlot® and Excel<br />
– Excel compatible workbooks, including <strong>for</strong>mulas, functions, data import/export, missing value codes, sort,<br />
trans<strong>for</strong>mations and worksheet merge<br />
– One-click export of results, plots, and tables to Microsoft Word®<br />
– Units definition and conversions, including the ability to specify preferred output units, units <strong>for</strong> a dosing<br />
regimen, and unit conversions<br />
– Support <strong>for</strong> Excel, SAS Transport, and NONMEM file <strong>for</strong>mats<br />
25 October 2006<br />
4<br />
© 2006 <strong>Pharsight</strong> Corporation. All rights reserved.
New <strong>IVIVC</strong> Toolkit <strong>for</strong> <strong>WinNonlin</strong><br />
• Extension of <strong>WinNonlin</strong> functionality<br />
• Separately licensed<br />
• Stand-alone tools enable <strong>IVIVC</strong> workflows<br />
– Deconvolution<br />
• Numerical<br />
• Wagner-Nelson<br />
• Loo-Riegelman<br />
– Convolution<br />
– Levy Plot<br />
• <strong>IVIVC</strong> Wizard facilitates particular <strong>IVIVC</strong> workflow<br />
25 October 2006<br />
5<br />
© 2006 <strong>Pharsight</strong> Corporation. All rights reserved.
New <strong>IVIVC</strong> Toolkit <strong>for</strong> <strong>WinNonlin</strong><br />
• <strong>IVIVC</strong> Wizard: Two-stage <strong>IVIVC</strong> development<br />
– Polyexponential UIR from IV or IR data<br />
– Deconvolution over individual subjects<br />
– Arbitrary <strong>IVIVC</strong> models could include<br />
• Dose non-linearity<br />
• Absorption window<br />
• Time scaling, shifting<br />
– New Convolution functionality<br />
– Levy plot<br />
– Internal and External Validation of <strong>IVIVC</strong> model<br />
• Convenience<br />
– Save settings <strong>for</strong> sharing between colleagues<br />
– Make a change in base data and refresh the entire analysis<br />
25 October 2006<br />
6<br />
© 2006 <strong>Pharsight</strong> Corporation. All rights reserved.
Game Plan<br />
• Dissolution Data<br />
– Choose data to model, data <strong>for</strong> external validation<br />
– Select a <strong>for</strong>mulation as a target (to compare to prediction)<br />
– We will interpolate using a dissolution model<br />
• Study Data<br />
– Use reference <strong>for</strong>mulation to get Unit Impulse Response (UIR)<br />
– Deconvolve PK profiles using UIRs to get average fraction absorbed <strong>for</strong> each<br />
<strong>for</strong>mulation<br />
• Build Correlation<br />
– Analyze Levy and Correlation plots to learn about possible correlation structure<br />
– Select and fit a correlation model<br />
• Evaluate Correlation Model<br />
– Examine prediction errors <strong>for</strong> Cmax and AUC<br />
• Predict<br />
– How similar is a new <strong>for</strong>mulation to our target?<br />
25 October 2006<br />
7<br />
© 2006 <strong>Pharsight</strong> Corporation. All rights reserved.
In Vitro Data<br />
Dissolution Profiles of CR <strong>for</strong>mulations<br />
1.2<br />
1.0<br />
0.8<br />
0.6<br />
0.4<br />
Form=CR1<br />
Form=CR2<br />
Form=CR3<br />
Form=CR4<br />
Form=CR5<br />
Form=CR6<br />
Form=CR7<br />
0.2<br />
0.0<br />
0 2 4 6 8 10 12 14 16 18<br />
Time (h)<br />
25 October 2006<br />
8<br />
© 2006 <strong>Pharsight</strong> Corporation. All rights reserved.
In Vivo Data<br />
Average PK Profiles <strong>for</strong> CR Formulations<br />
700<br />
600<br />
500<br />
400<br />
300<br />
200<br />
Form=CR1, Study=Study 1<br />
Form=CR2, Study=Study 1<br />
Form=CR3, Study=Study 1<br />
Form=CR4, Study=Study 2<br />
Form=CR5, Study=Study 3<br />
Form=CR6, Study=Study 3<br />
Form=CR7, Study=Study 4<br />
100<br />
0<br />
0 5 10 15 20 25 30<br />
Time (h)<br />
25 October 2006<br />
9<br />
© 2006 <strong>Pharsight</strong> Corporation. All rights reserved.
In Vivo Data<br />
Average PK Profiles <strong>for</strong> Reference Formulation<br />
1000<br />
900<br />
800<br />
700<br />
600<br />
500<br />
400<br />
300<br />
Form=ref, Study=Study 1<br />
Form=ref, Study=Study 2<br />
Form=ref, Study=Study 3<br />
Form=ref, Study=Study 4<br />
200<br />
100<br />
0<br />
0 5 10 15 20 25 30<br />
Time (h)<br />
25 October 2006<br />
10<br />
© 2006 <strong>Pharsight</strong> Corporation. All rights reserved.
<strong>IVIVC</strong> Wizard: In-Vitro Tab<br />
• Locate Workbook and Worksheet with dissolution data<br />
• Map dissolution data<br />
– One dissolution profile per <strong>for</strong>mulation<br />
• Model dissolution data<br />
– Using new sigmoidal models: Weibull, Hill, Double Weibull, Makoid-Banakar<br />
– Can choose weighting model<br />
– Provides <strong>for</strong> smooth interpolation while building/running models<br />
• Partition <strong>for</strong>mulations<br />
– Internal validation and modeling<br />
– External validation<br />
– Target (pivotal or brand), also used <strong>for</strong> external validation<br />
25 October 2006<br />
11<br />
© 2006 <strong>Pharsight</strong> Corporation. All rights reserved.
<strong>IVIVC</strong> Wizard: In-Vivo Tab<br />
• Locate Workbook and Worksheet with study data<br />
• Map study data<br />
– Which <strong>for</strong>mulation was used?<br />
– Which columns identify unique patients?<br />
• Specify reference <strong>for</strong>mulation<br />
– Used to fit Unit Impulse Response (UIR) function<br />
– Dosing route (IV-bolus, IV-infusion, Immediate Release/Solution)<br />
• Fit UIR<br />
– Choose weighting<br />
– Choose model selection<br />
– Will fit polyexponential to each patients reference profile<br />
– With IR reference, we can fit time lag and/or strip 1 st order ka<br />
– With IV reference, we can use bolus or infusion<br />
25 October 2006<br />
12<br />
© 2006 <strong>Pharsight</strong> Corporation. All rights reserved.
<strong>IVIVC</strong> Wizard: In-Vivo Tab<br />
• Deconvolution<br />
– Numerical deconvolution of each profile (that isn’t reference) using UIR <strong>for</strong> that<br />
subject<br />
– Results (Fraction Absorbed, Fabs) averaged per <strong>for</strong>mulation<br />
– Fabs based on administered dose <strong>for</strong> that <strong>for</strong>mulation<br />
• Using other data<br />
– User can provide average Fabs per <strong>for</strong>mulation and UIRs per subject<br />
– Bypasses fitting UIR and Deconvolution<br />
– Must strictly adhere to specified <strong>for</strong>mats<br />
– Why? To use WN, LR or other deconvolution approach with the <strong>IVIVC</strong> Wizard<br />
25 October 2006<br />
13<br />
© 2006 <strong>Pharsight</strong> Corporation. All rights reserved.
<strong>IVIVC</strong> Wizard: Correlation Tab<br />
• Exploratory Plots<br />
– Correlation Plot (Fabs vs. Fdiss)<br />
– Levy Plot (Tvivo vs. Tvitro)<br />
• Correlate deconvolved Fabs with Fdiss<br />
– Built-in equation: Fabs = Fa*Fdiss(tscale*time-tshift)-Foffset<br />
• Fdiss modeled or interpolated from raw data<br />
• Absorption cutoff optional<br />
– Custom equation built on user ASCII models<br />
• Only using modeled Fdiss<br />
• Can incorporate dose<br />
• Validation Grid<br />
– Runs correlation and convolution on dissolution data<br />
– Shows prediction error (%,ratio) of Cmax and AUC <strong>for</strong> each <strong>for</strong>mulation and<br />
average over all <strong>for</strong>mulations (internal and external) – compared to observed<br />
data <strong>for</strong> each <strong>for</strong>mulation.<br />
25 October 2006<br />
14<br />
© 2006 <strong>Pharsight</strong> Corporation. All rights reserved.
<strong>IVIVC</strong> Wizard: Prediction Tab<br />
• Locate Workbook and Worksheet with dissolution data<br />
• Map dissolution data<br />
– One dissolution profile per <strong>for</strong>mulation<br />
• Specify dose per <strong>for</strong>mulation<br />
• Model Dissolution Data<br />
– Or use interpolated raw data (with built-in models)<br />
• Predict PK<br />
– For each test <strong>for</strong>mulation<br />
• Run correlation<br />
• Convolve with UIRs from target <strong>for</strong>mulation<br />
– Uses individual UIRs, then compares mean of each.<br />
• Show prediction error (%,ratio) of Cmax and AUC – compared to observed data <strong>for</strong> each<br />
<strong>for</strong>mulation.<br />
25 October 2006<br />
15<br />
© 2006 <strong>Pharsight</strong> Corporation. All rights reserved.
<strong>IVIVC</strong> Wizard: Project Workspace<br />
• Load and Save projects<br />
• Colored bullets highlight status of project<br />
– Red – not completed<br />
– Yellow – needs to be updated<br />
– Green – complete<br />
• Designed so Prediction Tab can be used independently<br />
of model development workflow<br />
25 October 2006<br />
16<br />
© 2006 <strong>Pharsight</strong> Corporation. All rights reserved.
<strong>IVIVC</strong> Wizard: Usage Scenarios<br />
• Develop <strong>IVIVC</strong><br />
• Assist Formulation design<br />
• Evaluate dissolution experiment<br />
• Set dissolution specifications<br />
• Investigate absorption<br />
• Explore correlation structure <strong>for</strong> 1-stage <strong>IVIVC</strong><br />
25 October 2006<br />
17<br />
© 2006 <strong>Pharsight</strong> Corporation. All rights reserved.
<strong>IVIVC</strong> Toolkit <strong>for</strong><br />
<strong>WinNonlin</strong>® Product<br />
Demonstration<br />
To view the demonstration,<br />
please click on the ‘Replay<br />
Event Now’ link <strong>for</strong> this event,<br />
available on our website at:<br />
http://www.pharsight.com/events/<br />
eventsonline_archive.php
New <strong>IVIVC</strong> Toolkit <strong>for</strong> <strong>WinNonlin</strong><br />
• Benefits<br />
– Incrementally leverage your existing investment in <strong>WinNonlin</strong><br />
– Powerful <strong>IVIVC</strong> exploration and development capability<br />
– Visualize the in-vitro in-vivo relationship with new plotting features<br />
– Improve collaboration between Clinical Development and Formulation<br />
– Easy application of correlations to predict PK from new in vitro data<br />
25 October 2006<br />
19<br />
© 2006 <strong>Pharsight</strong> Corporation. All rights reserved.
Questions<br />
???<br />
Submit questions by<br />
clicking on the question<br />
mark icon on the floating<br />
toolbar located on the<br />
lower right side of your<br />
screen<br />
Or contact <strong>Pharsight</strong><br />
Sales later at<br />
sales@<strong>Pharsight</strong>.com