An imaging‑based immersed isogeometric framework for organ‑level forecasting of prostate cancer growth
Please login to view abstract download link
Prostate cancer (PCa) is a major global health burden for men, ranking among the most frequently diagnosed and deadliest cancers. Although screening and therapeutic options have improved, current clinical practice still relies heavily on generalized protocols and observational follow‑up, which contributes to persistent problems of over- and undertreatment. In response to these challenges, we develop a computational framework capable of generating individualized forecasts of PCa growth. Our approach is grounded in biomechanistic modelling and implemented through isogeometric methods [1,2]. Our work focuses on patients managed under active surveillance, where disease is monitored after diagnosis to delay or avoid unnecessary treatment. For each patient, we reconstruct the prostate anatomy directly from T2‑weighted MRI and estimate the parameters governing tumor growth by calibrating the model to serial multiparametric MRI examinations using a nonlinear least‑squares approach. The framework also incorporates longitudinal serum PSA measurements, enabling the model to reproduce and anticipate trends in this key biomarker of PCa. The numerical formulation relies on the finite cell method [2], an immersed approach in which the prostate geometry extracted from MRI is encoded as a level‑set function and embedded within a structured hexahedral mesh aligned with the voxel grid. This strategy avoids generating complex boundary‑fitted meshes and allows MRI segmentations to be integrated seamlessly into the computational domain. The use of truncated hierarchical splines further enables targeted refinement around the tumor region. Our simulations capture patient‑specific PCa behavior with high fidelity, offering a practical and efficient tool for personalized tumor forecasting. Thus, while further validation is required, our approach has the potential to support clinical decision‑making for improved management of PCa.