According to the American Cancer Society, prostate cancer (PCa) is the most common newly-diagnosed cancer and the second leading cancer-related cause of death among men in the US in 2019. The current clinical protocols of PCa are based on two key strategies: regular screening of men over fifty and patient triaging in risk groups. Prostatic tumors are usually detected at an early organ-confined stage, which poses a low to intermediate risk to the patient and may not produce any symptoms or require treatment for long time. However, most patients are prescribed a radical treatment immediately after diagnosis (e.g., surgery or radiation therapy), which may adversely impact their lives without necessarily prolonging their longevity or quality of life. Moreover, some patients who initially delay their treatment ultimately succumb to PCa due to an inaccurate initial diagnosis. Thus, while regular screening enables the detection of the majority of tumors at an early and mild stage, the limited individualization of patient monitoring and treatment has led to significant rates of both overtreatment and undertreatment.

To overcome these fundamental limitations in the clinical management of PCa, we propose the use of a mechanistic mathematical model for which we can perform computer simulations to forecast patient-specific tumor growth. This computational model is based on a set of partial differential equations that describe the main mechanisms involved in organ-confined PCa growth. The model is parameterized using longitudinal multiparametric magnetic resonance (MR) images and the available clinical data for each patient. Tumor growth is simulated over the patient's prostate extracted from T2-weighted MR images. We use isogeometric analysis to accurately and efficiently address the computational challenges arising in this application.

Our preliminary simulation results show that our computational technology can predict tumor growth and associated serum Prostate Specific Antigen (PSA, a key biomarker in clinical management of PCa) with reasonable accuracy. We also explore the potential of model parameters and variables to characterize tumor aggressivity. Thus, we believe that our imaging-based modeling approach could be a promising tool capable of being implemented in current PCa protocols to assist physicians in the clinical management of PCa.

Citation Format: Guillermo Lorenzo, Thomas J. Hughes, Alessandro Reali, Hector Gomez, Thomas E. Yankeelov. An image-based mechanistic computational model for early prediction of organ-confined untreated prostate cancer growth [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 5483.