A multiscale framework for in silico thrombus generation and photoacoustic simulations

Abstract

Thrombus composition and microstructure play a critical role in determining the treatment success for thrombus-related diseases such as ischemic stroke and deep vein thrombosis. However, no in vivo diagnostic method can fully capture thrombus microstructure yet, hindering personalized treatment. Photoacoustic imaging is uniquely positioned to provide information on thrombi composition as it relays optical absorption information from diffuse photons at acoustic propagation depths. Computational modeling enables systematic exploration of microstructural effects on imaging signals, offering insights into developing improved in vivo diagnostic techniques. However, no photoacoustic simulation platform can model microstructural features within centimeter-scale phantoms at reasonable computational cost. In this work, we present recursive iterative fibrin network emulation (REFINE), a topology-driven framework for generating in silico thrombi that replicate their key microstructural traits. Unlike existing methods, REFINE enables controlled, recursive optimization of thrombus topology, making it suitable for accurate photoacoustic modeling and potentially powerful for biomechanical analyses beyond this study. These digital thrombi are embedded into a multiscale photoacoustic simulation platform that bridges microscale acoustic modeling with macroscale thrombus geometries, enabling efficient and realistic simulation of photoacoustic signal responses. We created unique representation of thrombi microstructure for various compositions and porosities. Our simulation framework effectively links microstructural features to macroscale imaging outcomes, in agreement with previous empirical studies. Our simulation results demonstrate that thrombus microstructure significantly affects photoacoustic spectral responses and can be reliably modeled in silico. These findings highlight the potential of a multiscale photoacoustic simulation approach as a powerful tool for characterizing tissue microstructure and demonstrate the utility of our computational framework for in silico thrombi analysis and the development of diagnostic imaging strategies.