Organoid and spheroid research has significantly increased in recent years, offering crucial insights into developmental biology and oncology. The potential to reduce animal experimentation is a key driver. However, challenges exist in culturing these structures. Spheroids can be grown with or without extracellular matrices, while organoids are often cultured in Matrigel, a complex and variable substance that introduces irreproducibility. The mechanical impact of matrix scaffolds is often poorly understood. Consequently, Matrigel-free organoid growth methods are being actively pursued to address these issues. Contact-free levitation is desirable for sample handling. While holographic optical tweezers can rotate single cells, they are unsuitable for larger cell clusters due to overheating concerns. Ultrasound techniques offer a promising alternative. Standing bulk acoustic waves (BAWs) levitate samples by pushing them into low-pressure regions, enabling controlled rotation. Surface acoustic waves (SAWs) can also be used, and various techniques employing acoustic microstreaming have been explored. However, for compatibility with scanning-based imaging modalities like OCT, BAW operation is preferable due to its ability to support tilting the sample into various stationary orientations and simplified device scaling for larger samples. Acoustic levitation offers the advantage of scaffold-free confinement, making the system open to various assays like light irradiation, chemical addition, or mechanical probing. Optical coherence tomography (OCT) is a high-speed, label-free imaging technique well-suited for biological samples. While offering high resolution, mature samples can become optically dense, resulting in limited penetration depth and shadow artifacts. Shadow removal algorithms have been developed, but remain challenging for highly attenuating structures. 3D optical coherence refraction tomography addresses this by controlling the incident beam angle, but is limited in angular range and requires sample immobilization in agarose gel. This paper introduces ULTIMA-OCT to overcome these limitations.

Several studies have explored the use of ultrasound for manipulating biological samples, particularly for levitation and controlled rotation. Methods employing standing bulk acoustic waves (BAWs) and surface acoustic waves (SAWs) have been developed, demonstrating capabilities in rotating various organisms and cells. Holographic optical tweezers have been employed for single-cell manipulation, but their limitations regarding heat generation at larger scales necessitate alternative approaches. The paper references specific studies exploring BAWs, SAWs, and acoustic microstreaming for sample rotation and manipulation, highlighting the existing techniques and their limitations concerning sample size, stability, and compatibility with existing imaging techniques like OCT. The review also addresses previous attempts to improve OCT imaging of optically dense samples using shadow removal algorithms and methods to control incident beam angle, concluding that these are insufficient to address the challenges.

ULTIMA-OCT uses a 3D-printed acoustic manipulation chamber to levitate and reorient samples for multi-angle OCT imaging. The chamber features four piezoelectric transducers and four reflectors, generating standing bulk acoustic waves (BAWs) to create stable trapping positions. By modulating the transducer voltages, the sample is rotated stepwise in a controlled manner. The system is designed to accommodate a range of sample sizes and shapes, demonstrated using zebrafish embryos and melanoma spheroids. The acoustic chamber is characterized by its octagonal cross-section, enabling the generation of intersecting standing waves. The method is shown to be adaptable for samples with varying degrees of asymmetry, achieving both step-wise reorientation and controlled sustained rotation, depending on the specific acoustic settings. Dark-field microscopy is used for real-time monitoring during acoustic manipulation. The chamber is compatible with existing OCT systems, enabling multi-angle imaging through the bottom coverslip. The process involves acquiring optical coherence microscopy (OCM) scans at multiple orientations, where the exact angles are not precisely known a priori. A model-based algorithm addresses this uncertainty, reconstructing reflectivity, attenuation, and refractive index maps. The algorithm incorporates constraints and regularization techniques, including total variation (TV) and Tikhonov regularization, as well as positivity and object support constraints, to account for the uncertainties and to handle artifacts such as shadowing. The optimization parameters include reflectivity, attenuation, refractive index contrast, rotation (using unit quaternions), and translation. The algorithm uses a gradient-based optimization approach, utilizing a multiscale strategy involving low-resolution initial reconstructions followed by refinement at high resolution.

ULTIMA-OCT successfully levitated and reoriented both zebrafish embryos and melanoma spheroids. OCM imaging revealed significant shadow artifacts in single-angle images of both types of samples, particularly pronounced in the zebrafish eye and yolk sac, and in melanoma spheroids with high melanin content. Multi-angle imaging using ULTIMA-OCT mitigated these artifacts. The model-based algorithm effectively fused the multi-angle data, generating 3D reconstructions of reflectivity, attenuation, and refractive index maps for zebrafish embryos, with significantly improved penetration depth. The reconstruction results demonstrated the ability to visualize structures otherwise obscured by shadow artifacts in single-angle OCT/OCM images, resulting in a clearer and more complete 3D representation of the sample. The refractive index values obtained were consistent with those reported in the literature. The reconstructions were achieved with a relatively small number of viewing angles, suggesting efficiency of the approach. Furthermore, the system maintained stable trapping of samples during the imaging process, avoiding motion artifacts. The use of a large sample chamber helped to minimize the impact of reflections from chamber walls. The study demonstrated that the system is capable of handling samples of different sizes and shapes, including asymmetric samples.

ULTIMA-OCT offers several advantages over traditional OCT imaging methods. The contact-free levitation eliminates the need for sample immobilization in gel, avoiding potential artifacts and facilitating access for other assays. The multi-angle imaging significantly enhances penetration depth, revealing structures hidden in single-view images. The model-based reconstruction algorithm handles the inherent uncertainties in the viewing angles, providing accurate 3D reconstructions. The study successfully demonstrated the potential of ULTIMA-OCT in overcoming limitations faced with imaging optically dense specimens. The technique addresses the need for a versatile platform to handle samples with varying shapes and sizes and to enable multi-modal imaging and advanced experiments. The methodology presented here is adaptable to other microscopy techniques, expanding the potential applications.

ULTIMA-OCT provides a novel and effective approach to multi-angle OCT imaging of optically dense samples. The 3D-printed acoustic chamber is a simple and cost-effective addition to existing OCT systems. The model-based reconstruction algorithm successfully addresses the challenges associated with unknown viewing angles and shadow artifacts. The technique offers significant advantages in terms of sample accessibility, enhanced penetration depth, and the generation of comprehensive 3D reconstructions. Future research could focus on optimizing the chamber design and exploring applications in long-term monitoring of live samples. Furthermore, integrating ULTIMA-OCT with other imaging modalities would broaden its scope.

The reconstruction of the attenuation and refractive index maps has limitations. The model assumes angular independence of scattering and absorption, which might not always hold true, potentially affecting the accuracy of the attenuation map. The accuracy of the refractive index map, especially outside the eye regions, depends on the angular coverage and the specimen's characteristics. The relatively small number of viewing angles used in this study could be further optimized for enhanced accuracy. While biocompatibility was discussed, long-term studies on live samples are still needed to fully assess the effects of the acoustic trapping. The current setup may also benefit from further improvements in axial resolution and handling of large samples.