Bioequivalence and Bioavailability Forum • Correlation / IVIVC Wizard / WinNonlin

Samaya B
☆

India,
2014-07-31 14:48
(3917 d 11:00 ago)

Posting: # 13324
Views: 7,268

Correlation / IVIVC Wizard / WinNonlin [Software]

Dear All,
Good afternoon!

I was performing IVIVC using IVIVC Wizard_ WinNolin for the first time using 3 different Test formulation with slow, medium and fast release rate) and got stuck in third tab (Correlation). It would be very helpful if anyone can explain exact meaning of following correlation models and which correlation model is to be chosen while performing IVIVC?

2. Fabs = AbsScale * Diss (Tscale * Tvivo)
3. Fabs = AbsScale * Diss (Tscale * Tvivo - Tshift)
4. Fabs = AbsScale * (Diss (Tscale * Tvivo - Tshift) – AbsBase)

Kinldy advice!

Thank you!
Regards,
Samaya

Edit: I added numbers to the models. You forgot to mention the first one:

Fabs = Diss (Tvivo)

[Helmut]

Helmut
★★★

Vienna, Austria,
2014-07-31 16:15
(3917 d 09:33 ago)

@ Samaya B
Posting: # 13326
Views: 6,155

IVIVC Toolkit – supported correlation models

Post reply

Hi Samaya,

❝ I was performing IVIVC using IVIVC Wizard…

I only know the IVIVC Toolkit. Which version of Phoenix/WinNonlin are you working with?

❝ It would be very helpful if anyone can explain exact meaning of following correlation models…

The same time-scale in vitro and in vivo is used. I never have seen an example in the “real world”. Maybe that’s the reason why you didn’t mention it.
Time scaling of in vivo part. Most commonly applied model.
Like #2 but modeling a lag-time in vivo.
Like #3 but with an intercept.

❝ … and which correlation model is to be chosen while performing IVIVC?

It is your job to find out (and validate) which model fits your data best. The IVIVC Toolkit provides an entire arsenal of methods doing so. ;-)

If you are not sure, consider registering at Pharsight’s Extranet, where you could upload your project-file for further exploration.

—
Dif-tor heh smusma 🖖🏼 Довге життя Україна!
Helmut Schütz

The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes

SDavis
★★

UK,
2014-08-01 18:13
(3916 d 07:35 ago)

@ Samaya B
Posting: # 13330
Views: 6,034

Correlation / IVIVC Wizard / WinNonlin

Post reply

Helmut's right that you might be better discussing these specifics on the phoenix forum by posting your project if at all possible. Certainly as a start I would be looking at the levy Plots, the internal and external prediction errors etc.

Simon

—
Simon
Senior Scientific Trainer, Certara™
[link=https://www.youtube.com/watch?v=xX-yCO5Rzag[/link]
https://www.certarauniversity.com/dashboard
https://support.certara.com/forums/

Samaya B ☆ India, 2014-08-06 15:18 (3911 d 10:31 ago) @ SDavis Posting: # 13345 Views: 5,881	Correlation / IVIVC Wizard / WinNonlin Post reply
	Thank you Helmut for explaining models briefly. And yes I will get myself registred at phoenix forum. Thanks Simon! Regards, Samaya

JMCardot
☆

France,
2014-08-10 18:18
(3907 d 07:30 ago)

@ Samaya B
Posting: # 13357
Views: 5,869

Correlation / IVIVC Wizard / WinNonlin

Post reply

Dear Samaya,

May I add a point, the choice of the model will depend of your data.

First you have to understand your data and not only to press a "wizard" and to wait for results. Phoenix is a great software but if you input wrong information/incorect setting at the begining results are going to be nonsense.

If you use a time scaling that means for example that in vivo is slower than in vitro but with a similar shape/kinetic. In this case time scaling is simple. If the shape of the curve in vitro and in vivo is different that means that you do not reflect the same type of kinetics ... that is more problematic except if you know why.

Some key points:

which model used for fitting the data before time scaling.
what is the number of points that you have in vitro and in vivo. If you have only 3 points in vitro (close to 20, 50 and 80% dissolved) forget it, you cannot have a good estimation of the shape of the dissolution curve.
Do not over-parametrize your time scaling equation even if you have the impression that fit is better => if you have 3 points and 3 parameters fit is perfect but not informative.

As stated by Simon you have a lot of information on Phoenix forum.

Best regards,

JM