2013 - Glasgow - Scotland

PAGE 2013: Drug/Disease modelling
Sonya Tate

Tumour growth inhibition modelling and prediction of overall survival in patients with metastatic breast cancer treated with paclitaxel alone or in combination with gemcitabine

Sonya C Tate (1), Valerie Andre (1), Nathan Enas (2), Benjamin Ribba (3) and Ivelina Gueorguieva (1)

(1) Eli Lilly and Company, Erl Wood, UK, (2) Eli Lilly and Company, Indianapolis, USA, (3) INRIA, Grenoble, France

Objectives: The aim of this study was to develop a modelling framework to a) characterise the tumour growth inhibitory (TGI) effects of paclitaxel and gemcitabine in metastatic breast cancer (MBC) patients, and b) investigate the predictive potential of change in tumour size (CTS) on overall survival (OS).

Methods: A randomised phase III trial was performed to evaluate the clinical benefit of gemcitabine for MBC patients receiving concomitant paclitaxel therapy [1]. Data from the study were collated and inclusion criteria were applied to align the WHO tumour size dataset with RECIST 1.0. A sequential modelling approach was applied to first evaluate tumour growth dynamics and to subsequently predict OS from CTS. The PK/TGI model was developed to incorporate the effect of combination therapy. A survival model was used to characterise OS as a function of various covariates, including CTS at weeks 1 to 12, baseline tumour size, treatment group, ECOG status, age and ethnicity.

Results: Of the 598 patients enrolled, 486 patients contributed 1477 measurements meeting the inclusion criteria. Survival data were available for 446 of those patients, of which 376 had died at the last follow up. A PK/TGI model which was previously used to assess tumour growth dynamics in a range of tumour types [2,3,4] was successfully applied to MBC. To model drug combination, we included two distinct routes of tumour shrinkage, one by paclitaxel and another by gemcitabine. Drug exposure was incorporated into the model by simulating AUC0-24 from published PK models [5,6]. The median predicted CTS at week 8 was -11% from baseline for the paclitaxel arm and -26% for the gemcitabine/paclitaxel arm. OS was described using a survival model with a Weibull distribution, incorporating CTS at week 8 as the best predictor of OS. Additional covariates included in the survival model were ECOG status and baseline tumour size.

Conclusions: We have extended an existing clinical TGI model to incorporate combination therapy and successfully applied this model to analyse tumour size data in MBC patients treated with paclitaxel alone or in combination with gemcitabine. Notably, using this combination therapy model, we have found that CTS at week 8 was a good predictor of OS in accordance with previous studies for NSCLC and thyroid cancer [3,7]. These results support the use of this modelling approach in future clinical studies which incorporate combination therapy in a parallel design.

This work is supported by the DDMoRe project.

References:
[1] Albain, K S et al. J Clin Oncol (2008) 26(24):3950-3957
[2] Claret, L et al. J Clin Oncol (2009) 27(25):4103-4108
[3] Claret, L et al. Cancer Chemother Pharmacol (2010) 66(6):1141-1149
[4] Houk, B E et al. Cancer Chemother Pharmacol (2010) 66(2):357-371
[5] Zhang, L et al. J Pharmacokinet Pharmacodyn (2006) 33(3):369-393
[6] Karlsson, M O et al. Drug Metab Dispos (1999) 27(10):1220-1223
[7] Wang, Y et al. Clin Pharmacol Ther (2009) 86(2):167-174




Reference: PAGE 22 (2013) Abstr 2924 [www.page-meeting.org/?abstract=2924]
Oral: Drug/Disease modelling
Top