2021 - Online - In the cloud

PAGE 2021: Drug/Disease Modelling
Anyue Yin

Anti-cancer treatment schedule optimization based on tumor dynamics modelling incorporating evolving resistance

Anyue Yin (1,2), Johan G.C. van Hasselt (3), Henk-Jan Guchelaar (1,2), Lena E. Friberg (4)# , Dirk Jan A.R. Moes (1,2)# #these authors share senior authorship

(1) Department of Clinical Pharmacy and Toxicology, Leiden University Medical Center, Leiden, the Netherlands, (2) Leiden Network for Personalized Therapeutics, Leiden University Medical Center, Leiden, the Netherlands, (3) Division of Systems Biomedicine and Pharmacology, Leiden Academic Centre for Drug Research (LACDR), Leiden University, Leiden, The Netherlands, (4) Department of Pharmacy, Uppsala University, Uppsala, Sweden.

Introduction: Evolutionary mechanism driven by intra-tumor heterogeneity and the evolving adaption of tumor cells to the selection pressure of treatment is acknowledged as a key factor related to the development of treatment resistance[1-3]. To improve the anti-cancer treatment outcome, it may be important to take this mechanism into consideration when designing treatment strategies. The quantitative modeling approach could support the decision making process.

A clinical genetic biomarker that can capture the tumor heterogeneity and monitor the evolving treatment resistance is circulating tumor DNA (ctDNA)[4,5]. Monitoring tumor-specific genetic alternation can facilitate precise adjustment of treatment by selecting treatment targeting on the newly developed actionable mutation[4,6]. Such adaptive treatment suppresses the proliferation of resistant tumor clones. Considering evolutionary dynamics, suppressing the emergence of resistance by applying intermittent treatment has also been proposed[7-9].

Objectives: 1)To develop a model incorporating evolving cancer resistance to characterize tumor dynamics based on a dataset of metastatic colorectal cancer (mCRC) patients. 2)To explore novel treatment schedules with simulation in order to improve the clinical outcome of patients.

Methods: A dataset containing longitudinal tumor sizes and mutant KRAS levels in ctDNA was identified from a previous study where patients with mCRC were treated with the anti-EGFR inhibitor panitumumab[10].

A mathematical model was developed to describe the obtained data. Tumor tissue was assumed to consist of multiple clonal sub populations[2]. One clonal population (TS) was sensitive to anti-EGFR inhibitor (D1). Another clonal population (TR1) harbors KRAS mutation and was resistant to D1. KRAS mutation was assumed to be acquired during treatment. When the treatment was suspended, a back transfer process was incorporated. A hypothetical treatment (D2) that targets TR1 and a third clonal population (TR2) that is resistant to both D1 and D2 were also incorporated. The shedding rates of ctDNA depended on the size of TR1 and TR2, and Hill equations were applied to describe the delayed emergence (or ability to detect) of mutant genes in ctDNA. The parameter values were obtained by fitting to the data or informed by literature values. A sensitivity analysis was performed to evaluate the impact of parameter values.

Dosing schedules, including continuous D1, intermittent D1 with different on- and off-dosing duration, and adaptive schedules with D1 and D2 guided by ctDNA results were evaluated using simulations in R. One hundred virtual patients with wild-type KRAS initially were simulated for each schedule. Median progress-free survival (PFS), period when tumor size stayed below baseline level (TTS<TS0), and time when mutant genes became detectable (Tmutant_test) (>5 fragments/ml) were evaluated. 

Results: The developed model could capture the reported time-courses of tumor burden and mutant KRAS level in ctDNA obtained from 28 mCRC patients[10].

Five designs of intermittent schedule prolonged median PFS and median TTS<TS0 compared with the continuous schedule. The simulated regimen with D1 administered for 8 weeks and suspended for 4 weeks prolonged median PFS from 36 weeks to 44 weeks. The median TTS<TS0 was prolonged from 52 weeks to 60 weeks. Extending the treatment holiday caused inferior results. The mutant KRAS became detectable before disease progression.

The simulated adaptive regimens showed to further prolong median PFS to 56-72 weeks and TTS<TS0 to 114-132 weeks under different designs. A survival plot also illustrated better clinical outcomes under intermittent and adaptive treatment. Median Tmutant_test of the hypothetical second mutation was higher than median PFS but lower than TTS<TS0.

Sensitivity analysis showed that although the predicted PFS (tumor dynamics) and Tmutant_test (ctDNA dynamics) were affected when parameter values varied, the intermittent and adaptive schedule still resulted in better treatment outcomes (PFS) than continuous treatment.

Conclusions: A model incorporating evolving cancer resistance was developed and captured tumor dynamics and mutation concentrations observed in patients. Compared with continuous anti-cancer treatment, intermittent and adaptive schedules were predicted to better suppress the evolving resistance and suggested a potential improved clinical outcome. An external clinical pilot study is required to validate the results.

[1] Zhao, B., M.T. Hemann, and D.A. Lauffenburger, Modeling Tumor Clonal Evolution for Drug Combinations Design. Trends Cancer, 2016. 2(3): p. 144-158.
[2] Beerenwinkel, N., et al., Cancer evolution: mathematical models and computational inference. Syst Biol, 2015. 64(1): p. e1-25.
[3] Carr, T.H., et al., Defining actionable mutations for oncology therapeutic development. Nat Rev Cancer, 2016. 16(5): p. 319-29.
[4] Wan, J.C.M., et al., Liquid biopsies come of age: towards implementation of circulating tumour DNA. Nat Rev Cancer, 2017. 17(4): p. 223-238.
[5] Siravegna, G., et al., Clonal evolution and resistance to EGFR blockade in the blood of colorectal cancer patients. Nat Med, 2015. 21(7): p. 795-801.
[6] Nangalia, J. and P.J. Campbell, Genome Sequencing during a Patient's Journey through Cancer. N Engl J Med, 2019. 381(22): p. 2145-2156.
[7] Zhang, J., et al., Integrating evolutionary dynamics into treatment of metastatic castrate-resistant prostate cancer. Nature Communications, 2017. 8(1).
[8] Rowe, M., et al., The use of intermittent enzalutamide dosing in the treatment of metastatic castrate-resistant prostate cancer. Journal of Clinical Oncology, 2020. 38(6_suppl): p. 81-81.
[9] Gatenby, R.A. and J.S. Brown, Integrating evolutionary dynamics into cancer therapy. Nat Rev Clin Oncol, 2020. 17(11): p. 675-686.
[10] Diaz, L.A., Jr., et al., The molecular evolution of acquired resistance to targeted EGFR blockade in colorectal cancers. Nature, 2012. 486(7404): p. 537-40.

Reference: PAGE 29 (2021) Abstr 9823 [www.page-meeting.org/?abstract=9823]
Oral: Drug/Disease Modelling