The influence of toxicity constraints in models of chemotherapeutic protocol escalation

Abstract

The prospect of exploiting mathematical and computational models to gain insight into the influence of scheduling on cancer chemotherapeutic effectiveness is increasingly being considered. However, the question of whether such models are robust to the inclusion of additional tumour biology is relatively unexplored. In this paper, we consider a common strategy for improving protocol scheduling that has foundations in mathematical modelling, namely the concept of dose densification, whereby rest phases between drug administrations are reduced. To maintain a manageable scope in our studies, we focus on a single cell cycle phase-specific agent with uncomplicated pharmacokinetics, as motivated by 5-Fluorouracil-based adjuvant treatments of liver micrometastases. In particular, we explore predictions of the effectiveness of dose densification and other escalations of the protocol scheduling when the influence of toxicity constraints, cell cycle phase specificity and the evolution of drug resistance are all represented within the modelling. For our specific focus, we observe that the cell cycle and toxicity should not simply be neglected in modelling studies. Our explorations also reveal the prediction that dose densification is often, but not universally, effective. Furthermore, adjustments in the duration of drug administrations are predicted to be important, especially when dose densification in isolation does not yield improvements in protocol outcomes.

The Unpluggable in Pursuit of the Undruggable: Tackling the Dark Matter of the Cancer Therapeutics Universe

“…Indeed, it is becoming understood that toxicity can often be a reliable predictor of anticancer drug benefit (81, 82), consistent with a role for normal cell interaction in tumor viability and progression. Other drug classes that may not affect tumor response or survival in the metastatic setting – e.g., protease inhibitors, or antagonists of G-protein-coupled receptors (83) – could still provide first-in-class targets by which to block metastasis in the adjuvant setting, thus improving survival.…”

“…Similarly, the judicious adjuvant use of cytotoxic chemotherapy to induce stromal toxicity and thus trigger micrometastatic apoptosis due to paracrine loop disruption – for example, in the relatively unexplored context of high-grade prostate cancer – remains as rationally justified as more costly molecularly targeted initiatives. Indeed, it is becoming understood that toxicity can often be a reliable predictor of anticancer drug benefit ( 81 , 82 ), consistent with a role for normal cell interaction in tumor viability and progression. Other drug classes that may not affect tumor response or survival in the metastatic setting – e.g., protease inhibitors, or antagonists of G-protein-coupled receptors ( 83 ) – could still provide first-in-class targets by which to block metastasis in the adjuvant setting, thus improving survival.…”

Early treatment gains for antibiotic administration and within human host time series data

“…A shortcoming of the model (4.1) is that the interaction terms are specific. D’onofrio (2010) generalized the model (4.1) to allow for more flexible functions representing the interactions. In Appendix A we investigate models with a more general immune response.…”

Optimal Therapy Scheduling Based on a Pair of Collaterally Sensitive Drugs

“…The underlying dynamics of resistance development has previously been studied using cell populations consisting of treatment sensitive and resistant types, using either genotypic or phenotypic classifications [18]. Additionally, others have justified their choices of detailed cellular heterogeneities using: (i) stages in evolutionary structures [19,20], (ii) phases of cell cycle [21,22,23,24], or (iii) spatial distribution of irregular therapy effect [25,26]. Among these, researchers (including [18,22,23,27,28]) have studied the effect of a pair of collaterally sensitive drugs as we propose here, using the Goldie-Coldman model or its variations [19,28,29,30].…”