Thank you Linda for your valuable inputs. In an ivivc model, under correl. Tab, we get AbsScale, Tscale, and under “Generate plots” tab, we get the Fabs, Fdiss, Tvivo & Tvitro. What I meant was particularly calculation of “Diss” from “Diss(TscaleTvivo)". I believe Diss is not equal or same as Fdiss or is it so ? As the way I know Tvitro=TscaleTvivo, therefore I tried to calculate Diss from predicted dissolution profile (from the “Fit dissolution” tab) by multiplying the interpolated Fdiss(or predicted) with Tscale & Tvivo (obtained from the plot). I matched the time of vitro vs. vivo (i.e. Tvitro corresponding to Tvivo, e.g. Tvitro= 2 hr corresponding to Tvivo= 4 hr, Tscale = 0.83, Fdiss (at 2 hr) = 0.26; here, taking the Tvivo(=4 hr)Tscale(=0.83)Fdiss(at 2 hr)(=0.26)). However, I am unable to get the Fabs as calculated by WNL at the same time point. Moreover, I tried verifying the Tvitro=TscaleTvivo, I could not match Tvitro even (multiplying Tscale with Tvivo). For interpolation what you have provided is a way to calculate the interpolated (predicted) Fdiss but still the mystery of "Diss(TscaleTvivo)” remains. Can you kindly help me out where I might be going wrong ? Regards Arindam
Dear Arindam, You have "What I meant was particularly calculation of ‘Diss’ from 'Diss(TscaleTvivo)'", Recall that one can’t calculate ‘Diss’ by itself. One can calculate the function Diss evaluated at the time TscaleTvivo. Also, you have "taking the Tvivo(=4 hr)Tscale(=0.83)Fdiss(at 2 hr)(=0.26)", so it looks like you are trying to multiply by Diss and not treating it as a function. Also, you are using Fabs and Fdiss from the Levy Plot (from Generate Plots) to work this out, and I would do it a different way, as follows. I don’t have your project, so I am using the project that we ship with Phoenix, from: C:\Program Files (x86)\Pharsight\Phoenix\application\Examples\Projects\WinNonlin\EX08_IVIVC_Workflow.phxproj. It uses the same correlation function as you, Fabs = AbsScale * Diss(TscaleTvivo), and I want to check that for a certain Tvivo, it correctly calculates Fabs, and I will need to calculate Diss(TscaleTvivo) to do so. When the Correlation model is run, it generates an output called Correlation.Correlation.Predicted Data. Since the model used was Fabs = AbsScale * Diss(TscaleTvivo), the Predicted column in this output is Fabs, and the Time column in this output is Tvivo. Let’s try to verify the second line of this output, which has: Function = 1, Formulation = CR01, Time = 0.72 (this is Tvivo), Predicted = 0.65189 (this is Fabs). So first compute TscaleTvivo. From the Correlation.Correlation.Final Parameters output, we get the final parameter estimates for the correlation model: AbsScale 1.0204429 Tscale 0.23615882 So TscaleTvivo = 0.23615882 * 0.72 = 0.17003435 Now to compute Diss(TscaleTvivo), we want Diss(0.17003435). As I posted before, this is done by linear interpolation of the in vitro predicted dissolution values. So go to Vitro.Dissolution.Predicted Data, and try to find the time 0.170034345. It is not there, but it is in between these two time values: Form Function Time Predicted CR01 1 0.160804 0.61895 CR01 1 0.180905 0.662243 Using the linear interpolation equation I posted before: Diss(T) = y1 + ((T - t1)/(t2 - t1)) * (y2 - y1) Diss(0.17003435) = 0.61895 + ((0.17003435 - 0.160804)/(0.180905 - 0.160804)) * (0.662243 - 0.61895) Diss(0.17003435) = 0.63883 Then Fabs = AbsScale * Diss(Tscale*Tvivo) = 1.0204429 * 0.63883 = 0.65189 This exactly matches the line of output above that we were trying to match: Function = 1, Formulation = CR01, Time = 0.72 (this is Tvivo), Predicted = 0.65189 (this is Fabs). I hope this helps. Many of the people who monitor the forum will be on vacation for the next week or until the New Year, so it may take some time to get further responses. Regards, Linda