RT Journal Article SR Electronic T1 Phase i trials in melanoma: A framework to translate preclinical findings to the clinic JF bioRxiv FD Cold Spring Harbor Laboratory SP 015925 DO 10.1101/015925 A1 Eunjung Kim A1 Vito W. Rebecca A1 Keiran S.M. Smalley A1 Alexander R.A. Anderson YR 2015 UL http://biorxiv.org/content/early/2015/03/02/015925.abstract AB One Sentence Summary By integrating mathematical modeling with in vitro experiments we show that treatment-induced autophagy modulates melanoma treatment response, and that adding outcomes data from a companion clinical trial allows implementation of phase i trials (virtual, informed clinical trials) for melanoma combination therapy.Abstract The combination of chemotherapy and an AKT inhibitor in patients with metastatic solid tumors including melanomas was tested in a recently completed phase 1 clinical trial. Our experiments showed that such regimens differentially induce autophagy in melanoma cells and autophagy modulates the response to treatment. Motivated by these observations, we formulated a mathematical model comprised of a system of ordinary differential equations that explains the dynamics of the response of melanoma cells to different mono and combination therapies. Model parameters were estimated using an optimization algorithm that minimizes differences between predicted cell populations and experimentally measured cell numbers. The model predicts that the combination therapy treatment protocol used in the trial is effective in short term tumor control but that the treatment will eventually fail, although smarter schedules can be applied to extend response. To move this model forward into a more clinically relevant setting, we implemented a phase i trial (a virtual/imaginary yet informed clinical trial), where a genetic algorithm was used to generate a cohort of virtual patients that captured the diversity of disease response observed in a comparable clinical trial. Simulated clinical trials with the cohort and sensitivity analysis defined parameters that discriminated virtual patients having more favorable versus less favorable outcomes. These analyses established the relevance of selecting patients based on rates of tumor growth and autophagic flux. Finally, the model predicts optimal therapeutic approaches across all virtual patients, laying the foundation for phase i-informed clinical melanoma trials. The specific melanoma model developed here is just one example of the much broader potential of the phase i framework, which can be applied to almost any parameterized cancer model.