Paradoxical behavior of time to RECIST progression as a metric for treatment effect

e14022 Background: RECIST time to progression (RTTP) is the cornerstone of widely used surrogates for patient benefit such as RECIST progression-free survival (RPFS), which in turn is used either for accelerated regulatory approval or to study the effects of factors such as drug exposure and drug target expression levels on RPFS. While one might expect increased antitumor effect to result in longer RTTP and RPFS, we will show that this is often not the case due to a fundamental mathematical property of the RECIST definition of disease progression. Methods:We use a simple mathematical model that treats patient tumor burden dynamics as an evolutionary process consisting of a decreasing drug-sensitive subpopulation of tumor cells and an independently growing drug-resistant subpopulation. This model has successfully described tumor dynamics in NSCLC, melanoma (CPT Pharmacometrics Syst. Pharmacol. (2017) 6, 29–39), and ovarian cancer (internal data). In the presence of drug, we represent the net fitness of the sensitive cells in terms of an exponential decay or “kill” rate k > 0, and the net fitness of the resistant cells as an exponential growth rate g > 0, and the initial fraction of drug-resistant cells by 0 ≤ f ≤ 1. For a given virtual patient (parameter set), we simulate RTTP as the time after start of treatment at which tumor burden V(t) is 20% higher than its lowest value (per RECIST criteria). We then vary the kill rate k and check whether RTTP increases or decreases with increasing k. Results: There is a large set of parameter values for which RTTP is predicted to vary inversely with increasing kill rate k. These parameter values are attained clinically in ovarian cancer patients taking alisertib and paclitaxel (internal data). Conclusions: While RECIST progression criteria may be useful for patient care to identify when a patient is losing her response, care must be taken in interpreting RTTP as a metric of treatment effect due to the real possibility that a stronger treatment effect results in shorter RTTP. We recommend that alternative metrics for patient benefit, such as a generalization of Days Gained (Neal et al PLoS ONE 8(1)e51951), be evaluated. We will present results of these evaluations currently in progress.