2 replies to this topic

[mjPK]

Hi 

 

I am trying to fit a model where I have concentration vs time data in two of the compartments. I built this model graphically since I need a two-way micro-parameter clearance from the tumor compartment. I am doing an IV bolus dose of 1mg at time zero and fitting 72 hours of observation time. The data in the initial phases is noisy but I would expect the model to work through that a bit. 

 

When I fit, I get saturation behavior from my tumor compartment which diverges greatly from the observation data, but good agreement of the plasma data. I am not sure how to resolve here. 

 

Another peculiar behavior is that my observations for the tumor are very low vs plasma as expected. The model misses from the very beginning predicting much higher values. Orders of magnitude higher. 

 

I have attached the file, and the model I am referring to is oddly named "One_Compartment_Nano_with_tumor_obs"

 

I would welcome any guidance on this to get a better behavior when having a model with two sets of observations. 

 

Thanks in advance

Attached Files

•  pk_proj_20SEP23.phxproj   1.67MB   64 downloads

Edited by mjPK, 20 September 2023 - 12:14 PM.

[bwendt@certara.com]

Hi,

 

I think you started the right way: first model the plasma data and then add tumor data. You could setup the combined model in a slightly different way, using a peripheral compartment instead of an urine compartment.

Please see a few steps in the attached project file that shows you a good starting point for your work. You would need to improve the fit for the tumor data and I would suggest you include concentration data from other tissues.

 

Bernd

 

 Plasma_Tumor.phxproj   3.21MB   107 downloads

[mjPK]

Hi,

 

I think you started the right way: first model the plasma data and then add tumor data. You could setup the combined model in a slightly different way, using a peripheral compartment instead of an urine compartment.

Please see a few steps in the attached project file that shows you a good starting point for your work. You would need to improve the fit for the tumor data and I would suggest you include concentration data from other tissues.

 

Bernd

 

Plasma_Tumor.phxproj

 

 

Thanks for the reply. I was able to get a good fit by changing the model structure as you suggested.