Sreenath M. Krishnan (1), Emilie Schindler (1), Gaia Schiavon (2), Lena E. Friberg (1)
(1) Department of Pharmaceutical Biosciences, Uppsala University, Uppsala, Sweden. (2) The Institute of Cancer Research: Royal Cancer Hospital, London, UK.
Objectives: Three-dimensional (3D) and density-based tumor metrics have been suggested to better discriminate tumor response to treatment than the traditional unidimensional (1D) metrics for gastro-intestinal stromal tumor (GIST) which often exhibit non-uniform size changes. This study aims to characterize the longitudinal 1D, 3D and density responses of liver metastases in imatinib-treated GIST patients and quantify the inter-individual (IIV) and inter-lesion variability (ILV).
Methods: Data were obtained from a retrospective, non-interventional study involving 77 GIST patients treated with oral imatinib at a starting dose of 400 mg/800 mg q.d. (74/3) [1]. Maximum trans-axial diameter (MTD), software-calculated segmented volume (Vs), calculated ellipsoidal volume (Ve) and density data were collected from 137 lesions (1-2 per patient) at baseline and at least one post-baseline visit with median follow-up time of 360 days. Tumor growth inhibition (TGI) models with constant, exponential or Gompertz growth were explored to describe MTD, Vs and Ve data. Indirect response models (IDR) with inhibition of the production or stimulation of the loss of response were investigated for tumor density. IIV and ILV were tested in all parameters. All models were combined into a joint model to explore correlations.
Results: TGI models with an exponential growth and a linear dose-driven drug effect that washes out over time [2] best characterized the MTD, Vs and Ve data. A mixture model described the bimodal distributions of the baseline data and accounted for ILV. The model-predicted doubling time was typically lower for Ve (1.3 years) and Vs (1.5 years) than for MTD (8.7 years). The growth rate constants and drug effect parameters were associated with large IIV but no ILV. An IDR model with stimulation of the loss of response adequately described tumor density data. Large IIV (120 %CV) and ILV (53 %CV) were identified in the drug effect. High correlations (>99%) were estimated between MTD, Vs and Ve model parameters. Correlations between the density model parameters and MTD, Vs and Ve model parameters were low (<20%).
Conclusions: The developed models adequately described the longitudinal 1D, 3D and density data of liver metastases in imatinib-treated GIST patients. All tumor responses were associated with large IIV. ILV was identified in the drug sensitivity of density only. In a next step, the model-predicted tumor metrics will be evaluated as predictors of overall survival.
Acknowledgements: This work was supported by the Swedish Cancer Society.
References:
[1] Schiavon G. et al. Eur J Cancer (2014) 50:972-80.
[2] Claret et al. J Clin Oncol (2009) 27: 4103-4108.
Reference: PAGE 25 () Abstr 5934 [www.page-meeting.org/?abstract=5934]
Poster: Drug/Disease modeling - Oncology