Abstract

Active surveillance (AS) is a well-established clinical strategy for managing prostate cancer (PCa) with low to intermediate risk. However, current AS protocols have limitations, including delayed detection of tumor progression and a lack of personalized monitoring plans. To overcome these issues, we propose advancing AS towards a predictive patient-specific paradigm using a computational model. The proposed model utilizes longitudinal imaging and clinical data collected during standard AS for individual patients. It describes PCa growth in terms of tumor cell density dynamics, considering tumor cell mobility and net proliferation. The model is initialized with the first MRI scan and then parameterized to minimize the discrepancy between model predictions and tumor cell density at subsequent scans. The calibrated model is used to forecast tumor progression up to a future MRI scan. Preliminary results from an analysis of eight PCa patients enrolled in AS show promising agreement between the model predictions and actual tumor growth. These findings suggest that the proposed computational model could serve as a promising tool for predicting tumor progression in AS, aiding in the early identification of patients who may require treatment. If successful, this approach has the potential to improve the effectiveness of AS by providing personalized monitoring plans and reducing both overtreatment and undertreatment of PCa. Joint work with: G. Lorenzo, C. Wu, J.P. Yung, J.F. Ward, H. Gomez, A. Reali, T.E. Yankeelov, A.M. Venkatesan, T.J.R. Hughes