Experimental validation of a virtual imaging framework for three-dimensional functional optoacoustic tomography

Multiple-wavelength optoacoustic tomography (OAT), also known as photoacoustic computed tomography, enables estimation of functional parameters such as blood oxygen saturation based on wavelength-dependent optical absorption of chromophores. However, accurate reconstruction of functional images remains challenging due to difficulties in precisely characterizing imaging systems and the properties of the imaging target. A realistic and fully controllable virtual imaging framework offers a cost-effective and ethical approach for systematically investigating inversion behavior and supporting the development and evaluation of computational methods. Although a virtual imaging framework for three-dimensional (3D) quantitative OAT has been established, direct experimental validation has not yet been demonstrated. In addition, to the best of knowledge, direct comparison between simulated and experimental data for OAT systems employing a spherical measurement geometry, such as the one used in this study, has not been reported. This study presents an experimental validation of the framework through comparison of simulated and experimental measurements and the corresponding reconstructed images. A cylindrical gelatin phantom containing eight tubes filled with ionic aqueous solutions mimicking blood at oxygen saturation levels of approximately 56.5%, 64.5%, 85.2%, and 90.1% was constructed and imaged using the LOIS-3D system (TomoWave Inc., Houston, TX) at illumination wavelengths of 700 nm and 800 nm. A numerical phantom matching the fabricated physical phantom was generated, assigned experimentally measured optical and acoustic properties, and virtually imaged using a configuration replicating the experimental setup, including modeling of the spatial impulse response and electro-acoustic impulse response of the transducers. Both experimental and simulated data were reconstructed using the universal backprojection method and an optimization-based reconstruction method. Oxygen saturation levels were then estimated from the reconstructed initial pressure distributions using linear spectral unmixing. The results demonstrate consistent correspondence between experimental and simulated data at both the measurement level and the reconstructed image level. These findings support the use of the framework for reliable method development and systematic performance assessment in 3D OAT, including functional imaging applications.