15 November 2024
Adam Palmer
Department of Pharmacology
UNC Chapel Hill
Some cancers are cured by combination therapy, but there is limited understanding of how cures are achieved, partly because it involves shrinking tumors below the limit of detection. Mathematical models have illuminated important concepts in tumor drug response, but past models have been too simplistic to explain real clinical trial results in humans. I will present a mathematical model of tumor heterogeneity which produces clinically lifelike simulations of curative treatments for the most common blood cancer, large B-cell lymphoma. A key step in making clinically realistic models is borrowed from population-pharmacokinetics, which uses parameter distributions to model the diversity of drug kinetics in human populations. Here we adapt this approach to model tumor kinetics, using it at two different scales: cell-to-cell and patient-to-patient variation. What results is a 'bottom-up' mechanistic model of how heterogeneous tumors respond to multiple drugs, which quantitatively reproduces data from human trials including survival distributions, tumor kinetics during therapy, the optimal number of cycles of chemotherapy, and the prognostic effect of tumor growth rate. This model also explains why nearly all clinical trials over the past 20 years failed to improve cure rate in large B-cell lymphoma, and accurately predicted the success of the first new drug combination to improve cure rate over this period. This theory quantitatively explains how drug combinations cure some cancers and can be used to aid the design of new regimens and clinical trials.