Evaluation of bone metastases heterogeneity in metastatic prostate cancerTurk, Maruša (Avtor) Jeraj, Robert (Mentor) metastatic prostate cancercomputational modellingtumour heterogeneityinterpatient treatment response heterogeneityintrapatient treatment response heterogeneityresistanceThe treatment response of metastatic castration-resistant prostate cancer (mCRPC) to systemic therapies is affected by the evolution of resistance. Diverse levels of therapy resistance lead to heterogeneity of the treatment response, with some metastatic lesions, or even parts of lesions, that respond and some that do not respond. Evaluating treatment response heterogeneity is an important element based on which treatment strategies could be optimized. This thesis aimed to investigate the evolution of resistance and treatment response heterogeneity in mCRPC patients during their treatment course. To examine this objective, we developed a simple top-down computational model built on molecular imaging data of mCRPC patients treated with first-line (mCRPC-1) or subsequent-line (mCRPC-2) therapy. Ordinary differential equations simulated the dynamics of individual lesions, their response and the formation of new lesions. The model included only key kinetic parameters describing lesion growth, intrinsic and acquired resistance, the incidence of new lesion development and treatment efficiency. By examining the evolution of resistance and treatment response heterogeneity, the proportions of lesions comprising entirely of drug-resistant cells and nonfavourable responding lesions were extracted. Regarding mCRPC-1 patients, the proportion of nonfavourable responding lesions was low (median=0.45) even after 30 months of therapy, while the proportion of lesions comprising entirely of drug-resistant cells was high (median=0.96) after 18 months of therapy. Most of the mCRPC-2 lesions had a high proportion of nonfavourable responding lesions (median=0.68) and a high proportion of lesions comprising entirely of drug-resistant cells (median=1) were predicted after 5 months of therapy. Together, the findings revealed that some lesions with favourable treatment responses comprised entirely drug-resistant cells. This work demonstrates that our computational model can identify therapy-resistant lesions before they show clinical evidence of progression. These lesions should receive additional or new therapy before they show progression. However, our results require further confirmation in independent clinical studies.20222022-11-04 07:15:06Doktorsko delo/naloga142379VisID: 130004COBISS_ID: 126350339sl