dear all forum members, greetings! our company has purchased the IVIVC toolkit this year .As per my understanding of IVIVC the basic steps to perform IVIVC are 1.) fitting the in vitro data 2.) then generate the UIR & deconvolve the in vivo data 3.) these UIR’s will estimate the fraction of drug absorbed but i am not able to understand below listed functions in it 
1.) How is the dissolution model selection to be made ? 2.) What is the logic behind increasing the number of UIR exponentials? 3.) How to choose the correlation model amongst the three options given? thanks in advance
Hi Stat06, Your questions are fundamental to performing valid IVIVC and I would suggest you attend some training or read up some more about model fitting in general and specifically correlation of IVIVC. 1.) How is the dissolution model selection to be made ? > just like any other model fitting and selection process, reviewing visual diagnostics like goodness of fit as well as AIC etc to choose the model that best describes your data. For in vitro dissolution experiments there is the further complication as you may try to optimise your experiment there as best as you can to match observed in vivo behaviour. 
2.) What is the logic behind increasing the number of UIR exponentials? > this equates to the number of apparent compartments observed in your reference in vivo, if you see 1 compartment then you need only one exponential, for 2 compartments you will need the macro constants to describe those (A & B, which are the intercepts and alpha and beta, the slopes. 3.) How to choose the correlation model amongst the three options given? again visual inspection and goodness of fit diagnostics. the Levy plots will help you here, e.g. the line crosses at 0 then you could disregard Tshift from your model as your dissolution model is closely resembling any Tlag observed with the oral data. 
Simon
dear Simon, greetings! thanks a lot for your reply 
I am trying to understand the theory as well as application of IVIVC form the user guide and example giuide given along with the software to me. Can you refer some literature regarding this topic? Also i a have a query & i guess you can help me out with this one. The FDA has issued some new guidelines for MR dosage forms using QBD.In the example given they have advised to go for IVIVR less robust than IVIVC(according to the guidelines given) .Is there any difference between these two (IVIVC & IVIVR) or its just semantics? I am attaching the example pdf file as well for your reference . thanks stats 06 [file name=QbD_example_MR.pdf size=2302867]/extranet/media/kunena/attachments/legacy/files/QbD_example_MR.pdf[/file]QbD_example_MR-20120612.pdf (2.2 MB)
dear Simon, greetings! thanks a lot for your reply I am trying to understand the theory as well as application of IVIVC form the user guide and example giuide given along with the software to me. Can you refer some literature regarding this topic? Also i a have a query & i guess you can help me out with this one. The FDA has issued some new guidelines for MR dosage forms using QBD.In the example given they have advised to go for IVIVR less robust than IVIVC(according to the guidelines given) .Is there any difference between these two (IVIVC & IVIVR) or its just semantics? I am attaching the example pdf file as well for your reference . thanks stats 06 [file name=QbD_example_MR-20120612.pdf size=2302867]Certara | Drug Development Solutions (2.2 MB)
You should find links some of the references we used in the product documentation e.g. Pharmaceutical Product Development: in Vitro-in Vivo Correlation (Drugs and the Pharmaceutical Sciences) by Chilukuri, Dakshina Murthy (Editor), and Sunkara, Gangadhar (Editor), and Young, David (Editor) additionally you may find the following useful; www.uv.es/~mbermejo/DissolutionC.pdf there is some computer based training at www.tsrlinc.com/resources/modern-biopharmaceutics/ Regarding your query on the FDA MR guidelines, I haven’t had a chacne to read that yet but I think that Figs 10 and 11 illustrate why they went for IVIVR, i.e. a RELATIONSHIP which is less strong than a CORRELATION, so yes it is partly semantics. but also indicates that the data is insufficient to justify and validate a true IVIVC Simon 
