Abstract Bone metastasis (BM) is a lethal consequence for advanced prostate and renal cancer patients, contributing significantly to morbidity, elevated mortality risk, and substantial healthcare burden. In vivo biological experiments provide evidence of the cellular mechanisms supporting cancer progression and therapy response, but they are limited in addressing the multi-parameter complexity of BM in a time-cost effective manner and pose ethical concerns about harm and distress. In silico computational models, instead, can explore an unlimited number of experimental combinations avoiding time and resource consumption, providing a valuable alternative to investigate disease biology and microenvironment interactions. Accordingly, this work introduces digital twins to monitor BM progression, response to therapy and impact on bone mechanical properties. By combining three-dimensional multiphoton microscopy and in silico modeling, we generated in vivo inspired, spatially explicit, multicellular A(BM)2s (Agent-Based Models of BM) replicating tumor growth, angiogenesis, and bone resorption. We rigorously retrieved driving coefficients using in vivo data from prostate and kidney tumors and calibrated them through a random forest regressor algorithm to adhere to the in vivo growth rate. We further conducted robust double verification by simulating both the anti-angiogenic effects of cabozantinib and the anti-resorptive effects of zoledronic acid. Our results highlight the predictive character of our A(BM)2 in anticipating therapeutic outcomes and increasing our understanding of the complex dynamics of BM. In parallel, we developed a digital twin for mechanical testing of long bones and vertebrae that integrates finite element analysis with micro-CT-derived bone geometry and material properties. Computational compression tests revealed increased fragility in trabecular bone following treatment with the bone-targeting agent Radium-223, whereas three-point bending tests showed no fragility in cortical bone post-treatment. However, they did reveal significantly increased fragility after osteolysis occurred, indicating that these complementary approaches provide a comprehensive understanding of bone mechanical integrity. Overall, our digital twin approach provides a transformative platform to dissect tumor progression, predict therapy response, and evaluate the consequences on bone mechanics. By enabling these insights in silico, we aim to significantly reduce dependence on animal models, directly supporting and strengthening the 3Rs principle. Citation Format: Luca Marsilio, Stefan Maksimovic, Elisa Serafini, Alice Maccarini, Sergio Barrios, Pietro Cerveri, Stefano Casarin, Eleonora Dondossola. From animals to computational oncology: Digital twins as ethical enablers of next-generation bone metastasis research abstract. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 1486.
Marsilio et al. (2026) studied this question.