Build a cell-cycle based tumor growth model

Dear nmusers: Recently I am trying to build a cell-cycle based tumor growth model, i.e. tumor size = G1+ S + G2 + M. Meanwhile, I have quite a few simultaneous biomarkers to characterize the cycle behaviors. Those biomarkers could be calculated based on the population of G1, S, G2, and M. I have the following puzzles needed to be addressed: 1) In my data file, what is the CMT number I should give to those observations (such as tumor size and many biomarkers)? By the way, I need CMT to specify/initialize the corresponding compartment. Since I am not quite sure how to use L2 and PCMT, I tried many times, all failed except the following foolish way: I create one more compartment to hold the tumor size: DADT(iTUMROR) = DADT(iG1) + DADT(iS) + DADT(iG2) + DADT(iM) and in my data file, I put the CMT compartment created (iTUMOR) into the observation of tumor size. However, for all other biomarkers, there is no way to write this kind of DADT equations. How could I solve this in a smart way? 2) Since some biomarker data are collected based on certain time point, saying t0. In other words, they are relative fold-changes compared to the vehicle at t0. Therefore, I need to save the intermediate population of G1, S, G2 and M at t0, as constants, which will be used to scale the later-on populations so that the model predictions are comparable with the observations. However, in the both block $DES and $ERROR, there is no way to save these intermediate populations as global variables or constants. I guess, I have to use MSFO to separate the simulation to many sections? But it am not sure how to do it. Your thoughts and feedback are really appreciated! Thanks a lot! Feng

Dear Feng, a smart way could be to (actually) write down equations for the other biomarkers. The simplest way to do this is is to assume that when cells have evolved in a given phase (say G1) for a particular duration, they move to the following one (S). By doing such, you may also solve your second issue setting initial conditions to all your derivative equations. Best regards, Benjamin o set a phase duration and to write that when cells in G1 stayed for the entire duration, they move to D, and so on... Il 17-07-2008 10:49, "Feng Yang" <[EMAIL PROTECTED]> ha scritto: > Dear nmusers: > Recently I am trying to build a cell-cycle based tumor growth model, i.e. > tumor size = G1+ S + G2 + M. Meanwhile, I have quite a few simultaneous > biomarkers to characterize the cycle behaviors. Those biomarkers could be > calculated based on the population of G1, S, G2, and M. I have the following > puzzles needed to be addressed: > > 1) In my data file, what is the CMT number I should give to those > observations (such as tumor size and many biomarkers)? By the way, I need CMT > to specify/initialize the corresponding compartment. Since I am not quite > sure how to use L2 and PCMT, I tried many times, all failed except the > following foolish way: > I create one more compartment to hold the tumor size: > DADT(iTUMROR) = DADT(iG1) + DADT(iS) + DADT(iG2) + DADT(iM) > and in my data file, I put the CMT compartment created (iTUMOR) into the > observation of tumor size. > > However, for all other biomarkers, there is no way to write this kind of DADT > equations. How could I solve this in a smart way? > > 2) Since some biomarker data are collected based on certain time point, saying > t0. In other words, they are relative fold-changes compared to the vehicle at > t0. Therefore, I need to save the intermediate population of G1, S, G2 and M > at t0, as constants, which will be used to scale the later-on populations so > that the model predictions are comparable with the observations. > > However, in the both block $DES and $ERROR, there is no way to save these > intermediate populations as global variables or constants. I guess, I have > to use MSFO to separate the simulation to many sections? But it am not sure > how to do it. > > Your thoughts and feedback are really appreciated! Thanks a lot! > > Feng > > >

Feng, There is no need to worry about compartment numbers to match your observations. The CMT data item is only needed when you have to identify a compartment for an AMT. With NONMEM VI it is no longer necessary to use CMT and AMT to initialize compartments. If you are still using NONMEM V it is time you updated your system. I use my own named data item called DVID to identify different types of observations e.g. $INPUT ID TIME DVID DV L2 where DVID could range from 1 to 4 to define the 4 different observations. You could use it in $ERROR like this. IF (DVID.EQ.1) THEN Y=A(1)+EPS(1) ENDIF ... IF (DVID.EQ.4) THEN Y=A(4)+EPS(4) ENDIF I am not really sure how to reply to your other questions because I dont know which things you are measuring. Do you have observations of G1, S, G2 and M? Do you also have a measurement of tumour size? Please give us an idea of what differential equations you think you need to describe the biomarkers you mention. Nick Feng Yang wrote: > Dear nmusers: > > Recently I am trying to build a cell-cycle based tumor growth model, i.e. tumor size = G1+ S + G2 + M. Meanwhile, I have quite a few simultaneous biomarkers to characterize the cycle behaviors. Those biomarkers could be calculated based on the population of G1, S, G2, and M. I have the following puzzles needed to be addressed: > > 1) In my data file, what is the CMT number I should give to those observations (such as tumor size and many biomarkers)? By the way, I need CMT to specify/initialize the corresponding compartment. Since I am not quite sure how to use L2 and PCMT, I tried many times, all failed except the following foolish way: > > I create one more compartment to hold the tumor size: > DADT(iTUMROR) = DADT(iG1) + DADT(iS) + DADT(iG2) + DADT(iM) > > and in my data file, I put the CMT compartment created (iTUMOR) into the observation of tumor size. > > However, for all other biomarkers, there is no way to write this kind of DADT equations. How could I solve this in a smart way? > > 2) Since some biomarker data are collected based on certain time point, saying t0. In other words, they are relative fold-changes compared to the vehicle at t0. Therefore, I need to save the intermediate population of G1, S, G2 and M at t0, as constants, which will be used to scale the later-on populations so that the model predictions are comparable with the observations. > > However, in the both block $DES and $ERROR, there is no way to save these intermediate populations as global variables or constants. I guess, I have to use MSFO to separate the simulation to many sections? But it am not sure how to do it. > > Your thoughts and feedback are really appreciated! Thanks a lot! > > Feng -- Nick Holford, Dept Pharmacology & Clinical Pharmacology University of Auckland, 85 Park Rd, Private Bag 92019, Auckland, New Zealand [EMAIL PROTECTED] tel:+64(9)373-7599x86730 fax:+64(9)373-7090 http://www.fmhs.auckland.ac.nz/sms/pharmacology/holford

Feng, I don't really understand what you're trying to do , but, it might be helpful to note that you are not required to use the compartment prediction in the Y = statement For example, if you have $DES DADT(1) = .... DADT(2) = .... DADT(3) = ... it is permitted to have $ERROR Y = A(1)/6 + LOG(A(2)) + A(3)/S3 + EPS(1) without using F (which comes from a specific compartment). So, you may be able to construct the data file/model such that it doesn't matter which compartment the observation is assigned to. In NMVI, there is a new syntax for initializing compartments IF(A_0FLG.EQ.1) then A_0(2) = 1 END IF would this help with initializing the compartments? WRT your second question, I can't say that I really understand, but you can set up your own common as verbatim code and put whatever you want in it. $DES " FIRST " COMMON /MYVARS/ VAL1, VAL2, VAL3 " VAL1 = THETA(1) " VAL2 = A(1) TVAL1 = VAL1 TVAL2 = VAL2 DADT(1) = ... . . . but, always be aware of the limits of verbatim code, NMTRAN does not look at it, and will not properly take the derivative of any _expression_ inside. If you're want DOUBLE PRECISION numbers, you can just let NMTRAN generate the declarations for you, otherwise it gets a little more complicated. you can put whatever you want into it, it will be save between calls to DES, and it available in $PK and $ERROR (you'll need to put in the same COMMON statements in $PK or $ERROR). Mark Sale MD Next Level Solutions, LLC www.NextLevelSolns.com 919-846-9185 > -------- Original Message -------- Subject: [NMusers] Build a cell-cycle based tumor growth model From: "Feng Yang" <[EMAIL PROTECTED]> Date: Thu, July 17, 2008 4:49 am To: > > [email protected] > > Dear nmusers: Recently I am trying to build a cell-cycle based tumor growth model, i.e. tumor size = G1+ S + G2 + M. Meanwhile, I have quite a few simultaneous biomarkers to characterize the cycle behaviors. Those biomarkers could be calculated based on the population of G1, S, G2, and M. I have the following puzzles needed to be addressed: 1) In my data file, what is the CMT number I should give to those observations (such as tumor size and many biomarkers)? By the way, I need CMT to specify/initialize the corresponding compartment. Since I am not quite sure how to use L2 and PCMT, I tried many times, all failed except the following foolish way: I create one more compartment to hold the tumor size: DADT(iTUMROR) = DADT(iG1) + DADT(iS) + DADT(iG2) + DADT(iM) and in my data file, I put the CMT compartment created (iTUMOR) into the observation of tumor size. However, for all other biomarkers, there is no way to write this kind of DADT equations. How could I solve this in a smart way? 2) Since some biomarker data are collected based on certain time point, saying t0. In other words, they are relative fold-changes compared to the vehicle at t0. Therefore, I need to save the intermediate population of G1, S, G2 and M at t0, as constants, which will be used to scale the later-on populations so that the model predictions are comparable with the observations. However, in the both block $DES and $ERROR, there is no way to save these intermediate populations as global variables or constants. I guess, I have to use MSFO to separate the simulation to many sections? But it am not sure how to do it. Your thoughts and feedback are really appreciated! Thanks a lot! Feng