SBIR-STTR Award

In this Phase I application, we seek to establish the feasibility of an advanced deterministic approach for solving the radiative transport equation (RTE) for use within a commercially viable small animal optical imaging system. To date, small animal optical imaging using fluorescence and bioluminescence has been confined to non-tomographic planar imaging. However, internal heterogeneties and the non-geometrical propagation of light substantially reduce the effectiveness of these methods for imaging of intra-tissue sources of fluorescence or luminescence. While there have been attempts at predicting light propagation for tomographic small animal imaging using the diffusion approximation, the small volumes and heterogeneities present in mice provide conditions where diffusion theory is not valid. Realizing this, transport based solutions of the RTE have been idenfitied as a promising alternative. However, approaches to date have relied on numerical methods which do not possess the accuracy or efficiency required for effective small animal image reconstruction. In the proposed research, the capabilities and infrastructure of an established commercial radiation transport system provided by Radion Technologies will be leveraged for use in small animal tomographic image reconstruction. Through a combination of third order accurate spatial differencing, robust acceleration methods and the use of arbitrary tetrahedral elements, the proposed approach is well suited for accurately and efficiently modeling both transport and diffusive regimes. This technology will be applied towards modeling fluorescent light generation using frequency-domain photon migration (FDPM) measurements pioneered by the Photon Migration Laboratory at the Texas A&M University. The specific aims of this application are (1) to quantitatively evaluate performance of the proposed approach for forward predictions of FDPM measurements at the excitation and emission (fluorescent) wavelengths; (2) to adapt this approach for image reconstruction, including the extension of an existing adjoint solution method and development of a process control driver; (3) to reconstuct fluorophore absorbtion cross section mappings from FDPM measurements using (a) weighted back projection and (b) inverse optimization algorithms; and (4) to validate this approach through fluorophore concentration image reconstruction within cross sections of a mouse phantom. Success will be measured on the ability of the proposed approach to accurately reconstruct fluorophore concentrations while having a computational efficiency suitable for ultimate commercial implementation. If successful, Phase 2 will seek to further develop this process towards commercialization and to demonstrate RTE-based imaging of (i) peptide targeted fluorescent contrast agents in xenograft mice with metastatic cancer and (ii) GFR expression in transgenic mice

In the proposed Phase II research, a commercially viable software reconstruction platform, employing the solution of the radiative transport equation (RTE), will be developed for accurate and robust in vivo imaging of fluorescent targets in small animals. Although small animal optical tomography systems have recently been introduced, current systems do not employ fluorescence or bioluminescence, which have been confined to planar projections. In addition, current optical tomography systems rely on the use of time independent measurements and reconstruction algorithms based on the diffusion approximation. However, it is well known that the small volumes and heterogeneities in mice present conditions where the diffusion approximation is not valid. Realizing this, transport based solutions of the RTE have been researched as a promising alternative. However, approaches to date have relied on numerical methods which do not have the accuracy or efficiency required for commercial deployment. In the Phase I research, Transpire's algorithms for solving the RTE were successfully applied towards a proof-of-concept process, employing BCM's expertise in frequency-domain photon migration (FPDM), for reconstruction of fluorophore concentrations in a 3-D mouse phantom. Phase II research will extend this work to develop an RTE based, tomographic reconstruction software platform for FDPM photon migration. The specific aims are: (1) to develop an optimized 3-D RTE solver for modeling excitation and emission light propagation in forward and adjoint modes, streamlined for small animal imaging conditions with non-contact, non-matching mediums; (2) to develop an adaptive non-linear algorithm employing forward and adjoint RTE solutions for fluorescence yield reconstructions in small animal volumes; (3) to enable hybrid microCT/optical imaging by developing automatic grid generation from microCT images; (4) to integrate the above components into an automated software system for FDPM tomographic reconstruction; and (5) to quantitative demonstrate commercial viability through imaging of a peptide targeted NIR and visible fluorescent contrast agent in xenograft mice. Successful completion of Phase II will result in an automated software product, when combined with BCM's expertise in hardware technologies for frequency-domain fluorescence small animal imaging, can result in a commercially viable small animal imaging system. The proposed Phase II research will substantially advance the state-of-the-art in optical small animal imaging, which could enhance drug discovery as well as contribute significantly to new understanding of disease processes. Through preclinical validation, the proposed research can ultimately be migrated towards clinical imaging in areas such as lymph node mapping and tumor margin definition in melanoma.