Junghoon Lee (1), Konstantinos Biliouris (2), Jeremy Hing (1) , Stuart Shumway (1), Dinesh De Alwis (1)
(1) Merck Research Laboratories, Merck & Co. Inc., Rahway, NJ 07065, U.S.A.
(2) Chemical Engineering Department, University of Minnesota, Twin Cities, Minneapolis, MN 55455, U.S.A.
Objectives: The tumor-growth model by Simeoni et al. [1] has been widely used to understand and characterize the drug effect observed in xenograft models. Two improvements to the Simeoni model are proposed to model the effects of targeted therapeutics.
Methods: Phase-specificity of targeted therapeutics is modeled by introducing four additional compartments, representing gap 1 (G1), synthesis (S), gap 2 (G2), and mitosis (M) phases of the cell cycle. The relative transition rates among the four compartments are fixed and adopted from the literature, whereas the cell-death rates from the compartments are empirically determined to closely recapitulate the exponential-to-linear growth characteristics of the Simeoni model. Finally, the drug effect on cell death from the S-phase compartment is modeled with a sigmoidal function instead of a proportional one.
Results: The model is validated against data from xenograft bearing mice treated with gemcitabine and drug X known to be targeted. The transition from the proportional to the Emax model for the drug effect on cell death successfully captures both the non-proportional effect of drug X on tumor growth and the proportional effect of gemcitabine during the first 30 days of treatment. In a 60-day simulation, the proposed model, unlike Simeoni model, predicts that tumor regression is affected not only by the total amount of administered drug but also by the treatment schedule.
Conclusions: Our model suggests that the efficacy of targeted therapeutics may exhibit treatment-schedule dependency due to their phase-specificity; a characteristic that Simeoni model fails to capture. In addition, validation using drug-X data shows that the drug effect on cell death is non-linear for some drugs and the Simeoni model needs to be adjusted. The use of literature (and possibly in-vitro) values for some model parameters can help limit the number of additional estimated parameters to two, keeping the proposed model practical for drug development settings. Further validation of the proposed model with longer-term data is needed to confirm the predicted dependency on treatment schedule.
References:
[1] M. Simeoni et al., Cancer Res., 64: 1094-101 (2004).
Reference: PAGE 22 (2013) Abstr 2916 [www.page-meeting.org/?abstract=2916]
Poster: Oncology