Inverse scattering in biological samples via beam propagation.

Abstract

Multiple scattering limits optical imaging in thick biological samples by scrambling sample-specific information. Physics-based inverse-scattering methods aim to computationally unscramble this information often by using nonconvex optimization solvers. However, their inherent nonconvexity often leads to highly sample-dependent performance and inaccurate reconstructions, particularly in strongly scattering specimens. Here, we introduce a novel inverse-scattering framework based on multislice beam propagation (MSBP) that robustly achieves high-quality scatter correction and label-free volumetric imaging across a diverse range of scattering biological samples. We rigorously benchmarked imaging performance across multiple MSBP solver implementations using both scattering calibration phantoms and biological specimens. We found that an amplitude-only cost function in the inverse solver, combined with angular and defocus diversity in the scattering measurements, enabled volumetric, label-free imaging with high-quality and subcellular-level scatter correction. Together, these results establish a foundation for the reliable application of inverse scattering to achieve biologically interpretable three-dimensional imaging in increasingly thick, multicellular samples, thus introducing a new paradigm for deep-tissue computational imaging.