Robust reconstruction of fluorescence molecular tomography based on adaptive adversarial learning strategy

Objective.Fluorescence molecular tomography (FMT) is a promising molecular imaging modality for quantifying the three-dimensional (3D) distribution of tumor probes in small animals. However, traditional deep learning reconstruction methods that aim to minimize the mean squared error (MSE) and iterative regularization algorithms that rely on optimal parameters are typically influenced by strong noise, resulting in poor FMT reconstruction robustness.Approach.In this letter, we propose an adaptive adversarial learning strategy (3D-UR-WGAN) to achieve robust FMT reconstructions. Unlike the pixel-based MSE criterion in traditional CNNs or the regularization strategy in iterative solving schemes, the reconstruction strategy can greatly facilitate the performance of the network models through alternating loop training of the generator and the discriminator. Second, the reconstruction strategy combines the adversarial loss in GANs with the L1 loss to significantly enhance the robustness and preserve image details and textual information.Main results.Both numerical simulations and physical phantom experiments demonstrate that the 3D-UR-WGAN method can adaptively eliminate the effects of different noise levels on the reconstruction results, resulting in robust reconstructed images with reduced artifacts and enhanced image contrast. Compared with the state-of-the-art methods, the proposed method achieves better reconstruction performance in terms of target shape recovery and localization accuracy.Significance.This adaptive adversarial learning reconstruction strategy can provide a possible paradigm for robust reconstruction in complex environments, and also has great potential to provide an alternative solution for solving the problem of poor robustness encountered in other optical imaging modalities such as diffuse optical tomography, bioluminescence imaging, and Cherenkov luminescence imaging.

Air-Filled Microbubbles Based on Albumin Functionalized with Gold Nanocages and Zinc Phthalocyanine for Multimodal Imaging

Microbubbles are intravascular contrast agents clinically used in diagnostic sonography, echocardiography, and radiology imaging applications. However, up to date, the idea of creating microbubbles with multiple functionalities (e.g., multimodal imaging, photodynamic therapy) remained a challenge. One possible solution is the modification of bubble shells by introducing specific compounds responsible for such functions. In the present work, air-core microbubbles with the shell consisting of bovine serum albumin, albumin-coated gold nanocages, and zinc phthalocyanine were prepared using the sonication method. Various physicochemical parameters such as stability over time, size, and concentration were investigated to prove the potential use of these microbubbles as contrast agents. This work shows that hybrid microbubbles have all the necessary properties for multimodal imaging (ultrasound, raster-scanning microscopy, and fluorescence tomography), which demonstrate superior characteristics for potential theranostic and related biomedical applications.

Monitoring cancer cell surface receptor expression during anti-angiogenesis therapy in vivo

Concurrent administration of cancer therapeutics with tumor vasculature targeting treatment has been shown to improve overall survival in multiple human cancer types, as such combinations aim to destroy different compartments of tumors. Anti-angiogenesis therapeutics designed to inhibit tumor induced vessel sprouting have also been shown to re-model the tumor vasculature through a transient vessel normalization effect, which leads to improved perfusion of oxygen and drug in tumor. However, the effects that this normalized vasculature has on the availability of cancer receptor, such as EGFR, is unknown. Herein, we examined the use of MRI-PAFT to estimate cancer surface receptor availability in response to anti-angiogenesis therapy, using MRI-coupled paired agent fluorescence tomography. Bevacizumab treated tumors showed increase in RA compared to control tumors, but this was not statistically significant.

Noninvasive quantification of target availability during therapy using paired-agent fluorescence tomography

Immuno-oncological treatment strategies that target abnormal receptor profiles of tumors are an increasingly important feature of cancer therapy. Yet, assessing receptor availability (RA) and drug-target engagement, important determinants of therapeutic efficacy, is challenging with current imaging strategies, largely due to the complex nonspecific uptake behavior of imaging agents in tumors. Herein, we evaluate whether a quantitative noninvasive imaging approach designed to compensate for nonspecific uptake, MRI-coupled paired-agent fluorescence tomography (MRI-PAFT), is capable of rapidly assessing the availability of epidermal growth factor receptor (EGFR) in response to one dose of anti-EGFR antibody therapy in orthotopic brain tumor models.

Methods: Mice bearing orthotopic brain tumor xenografts with relatively high EGFR expression (U251) (N=10) or undetectable human EGFR (9L) (N=9) were considered in this study. For each tumor type, mice were either treated with one dose of cetuximab, or remained untreated. All animals were scanned using MRI-PAFT, which commenced immediately after paired-agent administration, and values of RA were recovered using a model-based approach, which uses the entire dynamic sequence of agent uptake, as well as a simplified “snapshot” approach which requires uptake measurements at only two time points. Recovered values of RA were evaluated between groups and techniques. Hematoxylin & eosin (H&E) and immunohistochemical (IHC) staining was performed on tumor specimens from every animal to confirm tumor presence and EGFR status.

Results: In animals bearing EGFR(+) tumors, a significant difference in RA values between treated and untreated animals was observed (RA = 0.24 ± 0.15 and 0.61 ± 0.18, respectively, p=0.027), with an area under the curve - receiver operating characteristic (AUC-ROC) value of 0.92. We did not observe a statistically significant difference in RA values between treated and untreated animals bearing EGFR(-) tumors (RA = 0.18 ± 0.19 and 0.27 ± 0.21, respectively; p = 0.89; AUC-ROC = 0.55), nor did we observe a difference between treated EGFR(+) tumors compared to treated and untreated EGFR(-) tumors. Notably, the snapshot paired-agent strategy quantified drug-receptor engagement within just 30 minutes of agent administration. Examination of the targeted agent alone showed no capacity to distinguish tumors either by treatment or receptor status, even 24h after agent administration.

Conclusions: This study demonstrated that a noninvasive imaging strategy enables rapid quantification of receptor availability in response to therapy, a capability that could be leveraged in preclinical drug development, patient stratification, and treatment monitoring.

Estimating paired-agent uptake in altered tumor vasculature using MRI-coupled fluorescence tomography

Angiogenesis inhibiting cancer therapy has become a standard treatment for many cancer types. The ability to examine the effects of these drugs in tumors noninvasively could help assess efficacy early in the treatment course or identify optimal times to introduce other combinatorial treatments. Herein, we examine whether a paired agent MRI-coupled fluorescence tomography approach can be used to monitor the effects of anti-angiogenesis therapy. Using small animal models bearing orthotopic glioma xenografts, we demonstrate noninvasive quantification of paired-agent uptake in response to anti-angiogenesis therapy in vivo. The result provides insights on receptor targeted drug delivery in altered vasculature, a potential important development for treatment monitoring and combinatorial strategies.

Small-animal 360-deg fluorescence diffuse optical tomography using structural prior information from ultrasound imaging

We demonstrate dual modality of free-space fluorescence diffuse optical tomography (FDOT) and handheld ultrasound (US) imaging to reveal both functional and structural information in small animals. FDOT is a noninvasive method for examining the fluorophore inside an object from the light distribution of the surface. In FDOT, a 660-nm continuous wave diode laser was used as an excitation source and an electron-multiplying charge-coupled device (EMCCD) was used for fluorescence data acquisition. Both the laser and EMCCD were mounted on a 360-deg rotation gantry for the transmission optical data collection. The structural information is obtained from a 6- to 17-MHz handheld US linear transducer by single-side access and conducts in the reconstruction as soft priors. The rotation ranges from 0 deg to 360 deg; different rotation degrees, object positions, and parameters were determined for comparison. Both phantom and tissue phantom results demonstrate that fluorophore distribution can be recovered accurately and quantitatively using this imaging system. Finally, an animal study confirms that the system can extract a dual-modality image, validating its feasibility for further in vivo experiments. In all experiments, the error and standard deviation decrease as the rotation degree is increased and the error was reduced to 10% when the rotation degree was increased over 135 deg.

Estimating drug delivery using hybrid system for simultaneous dynamic MRI and fluorescence tomography

Optical tomography is often coupled with high resolution imaging modality like MRI to provide functional information associated with specific anatomical structure noninvasively. MRI-coupled paired agent fluorescence molecular tomography (MRI-PAFT) is a hybrid imaging modality capable of noninvasively quantifying drug-target engagement in vivo utilizing a targeted and an untargeted fluorescence agent. This study compares the uptake kinetics of MRI contrast agent and fluorescence agents in tumor and normal tissue, and demonstrates the potential of utilizing MRI contrast agent kinetic and targeted fluorescence agent kinetics to quantify targeted tumor receptor concentration in glioma tumor model.

Increased Nose-to-Brain Delivery of Melatonin Mediated by Polycaprolactone Nanoparticles for the Treatment of Glioblastoma

PURPOSE

Intranasal administration has been extensively applied to deliver drugs to the brain. In spite of its unfavorable biopharmaceutic properties, melatonin (MLT) has demonstrated anticancer effects against glioblastoma. This study describes the nose-to-brain delivery of MLT-loaded polycaprolactone nanoparticles (MLT-NP) for the treatment of glioblastoma.

METHODS

MLT-NP were prepared by nanoprecipitation. Following intranasal administration in rats, brain targeting of the formulation was demonstrated by fluorescence tomography. Brain and plasma pharmacokinetic profiles were analyzed. Cytotoxicity against U87MG glioblastoma cells and MRC-5 non-tumor cells was evaluated.

RESULTS

MLT-NP increased the drug apparent water solubility ~35 fold. The formulation demonstrated strong activity against U87MG cells, resulting in IC50 ~2500 fold lower than that of the free drug. No cytotoxic effect was observed against non-tumor cells. Fluorescence tomography images evidenced the direct translocation of nanoparticles from nasal cavity to the brain. Intranasal administration of MLT-NP resulted in higher AUC_brain and drug targeting index compared to the free drug by either intranasal or oral route.

CONCLUSIONS

Nanoencapsulation of MLT was crucial for the selective antitumoral activity against U87MG. In vivo evaluation confirmed nose-to-brain delivery of MLT mediated by nanoparticles, highlighting the formulation as a suitable approach to improve glioblastoma therapy.

Whole-mount in situ hybridization of mouse brain to precisely locate mRNAs via fluorescence tomography

Nucleic acids of intact biological tissues are rich in biological information. Whole-mount in situ hybridization is a powerful technique to mine the wealth of data contained in DNAs or RNAs, especially mRNAs. However, there are no simple, rapid approaches to precisely locate mRNAs in whole-mount tissues such as intact brains. By combining the penetration procedures of iDISCO with the signal amplification approach termed hybridization chain reaction, we herein developed a method for whole-brain in situ hybridization at cellular resolution. Based on fluorescence tomography instead of tissue clearing, this method provides a simple, rapid way to precisely locate mRNAs in the whole brain with cytoarchitectonic landmarks. As a proof of principle, we investigated the exact distribution of Cre mRNA in a Thy1-Cre mouse brain. We found high levels of Cre mRNA in most regions of the subcortical nuclei and the brain stem but comparatively low levels in major areas of the cerebral cortex. This method may have broad applications in studies of RNA function and its relations with diseases.

Noninvasive imaging of dual-agent uptake in glioma and normal tissue using MRI-coupled fluorescence tomography

As the role of immuno-oncological therapeutics expands, the capacity to noninvasively quantify molecular targets and drug-target engagement is increasingly critical to drug development efforts and treatment monitoring. Previously, we showed that MRI-coupled dual-agent fluorescence tomography (FMT) is capable of estimating the concentration of epidermal growth factor receptor (EGFR) in orthotopic glioma models noninvasively. This approach uses the dynamic information of two fluorescent agents (a targeted agent and untargeted isotype) to estimate tumor receptor concentration in vivo. This approach generally relies on the two tracers having similar kinetics in normal tissues, which may not always be the case. Herein, we describe an additional channel added to the MRI-FMT system which measures the uptake of both agents in the normal muscle, data which can be used to compensate for differing kinetic behavior.

In Vivo Imaging of Prostate Cancer Tumors and Metastasis Using Non-Specific Fluorescent Nanoparticles in Mice

With the growing interest in the use of nanoparticles (NPs) in nanomedicine, there is a crucial need for imaging and targeted therapies to determine NP distribution in the body after systemic administration, and to achieve strong accumulation in tumors with low background in other tissues. Accumulation of NPs in tumors results from different mechanisms, and appears extremely heterogeneous in mice models and rather limited in humans. Developing new tumor models in mice, with their low spontaneous NP accumulation, is thus necessary for screening imaging probes and for testing new targeting strategies. In the present work, accumulation of LipImage^TM 815, a non-specific nanosized fluorescent imaging agent, was compared in subcutaneous, orthotopic and metastatic tumors of RM1 cells (murine prostate cancer cell line) by in vivo and ex vivo fluorescence imaging techniques. LipImage^TM 815 mainly accumulated in liver at 24 h but also in orthotopic tumors. Limited accumulation occurred in subcutaneous tumors, and very low fluorescence was detected in metastasis. Altogether, these different tumor models in mice offered a wide range of NP accumulation levels, and a panel of in vivo models that may be useful to further challenge NP targeting properties.

In Vivo Follow-up of Brain Tumor Growth via Bioluminescence Imaging and Fluorescence Tomography

Reporter gene-based strategies are widely used in experimental oncology. Bioluminescence imaging (BLI) using the firefly luciferase (Fluc) as a reporter gene and d-luciferin as a substrate is currently the most widely employed technique. The present paper compares the performances of BLI imaging with fluorescence imaging using the near infrared fluorescent protein (iRFP) to monitor brain tumor growth in mice. Fluorescence imaging includes fluorescence reflectance imaging (FRI), fluorescence diffuse ^optical tomography (fDOT), and fluorescence molecular Imaging (FMT^®). A U87 cell line was genetically modified for constitutive expression of both the encoding Fluc and iRFP reporter genes and assayed for cell, subcutaneous tumor and brain tumor imaging. On cultured cells, BLI was more sensitive than FRI; in vivo, tumors were first detected by BLI. Fluorescence of iRFP provided convenient tools such as flux cytometry, direct detection of the fluorescent protein on histological slices, and fluorescent tomography that allowed for 3D localization and absolute quantification of the fluorescent signal in brain tumors.

Near-Infrared Squaraine Dye Encapsulated Micelles for in Vivo Fluorescence and Photoacoustic Bimodal Imaging

Combined near-infrared (NIR) fluorescence and photoacoustic imaging techniques present promising capabilities for noninvasive visualization of biological structures. Development of bimodal noninvasive optical imaging approaches by combining NIR fluorescence and photoacoustic tomography demands suitable NIR-active exogenous contrast agents. If the aggregation and photobleaching are prevented, squaraine dyes are ideal candidates for fluorescence and photoacoustic imaging. Herein, we report rational selection, preparation, and micelle encapsulation of an NIR-absorbing squaraine dye (D1) for in vivo fluorescence and photoacoustic bimodal imaging. D1 was encapsulated inside micelles constructed from a biocompatible nonionic surfactant (Pluoronic F-127) to obtain D1-encapsulated micelles (D1(micelle)) in aqueous conditions. The micelle encapsulation retains both the photophysical features and chemical stability of D1. D1(micelle) exhibits high photostability and low cytotoxicity in biological conditions. Unique properties of D1(micelle) in the NIR window of 800-900 nm enable the development of a squaraine-based exogenous contrast agent for fluorescence and photoacoustic bimodal imaging above 820 nm. In vivo imaging using D1(micelle), as demonstrated by fluorescence and photoacoustic tomography experiments in live mice, shows contrast-enhanced deep tissue imaging capability. The usage of D1(micelle) proven by preclinical experiments in rodents reveals its excellent applicability for NIR fluorescence and photoacoustic bimodal imaging.

X-ray nanoprobes and diffraction-limited storage rings: opportunities and challenges of fluorescence tomography of biological specimens

X-ray nanoprobes require coherent illumination to achieve optic-limited resolution, and so will benefit directly from diffraction-limited storage rings. Here, the example of high-resolution X-ray fluorescence tomography is focused on as one of the most voracious demanders of coherent photons, since the detected signal is only a small fraction of the incident flux. Alternative schemes are considered for beam delivery, sample scanning and detectors. One must consider as well the steps before and after the X-ray experiment: sample preparation and examination conditions, and analysis complexity due to minimum dose requirements and self-absorption. By understanding the requirements and opportunities for nanoscale fluorescence tomography, one gains insight into the R&D challenges in optics and instrumentation needed to fully exploit the source advances that diffraction-limited storage rings offer.

Mesoscopic fluorescence tomography of a photosensitizer (HPPH) 3D biodistribution in skin cancer

RATIONALE AND OBJECTIVES

Photodynamic therapy (PDT) is a promising strategy for treating cancer. PDT involves three components: a photosensitizer (PS) drug, a specific wavelength of drug-activating light, and oxygen. A challenge in PDT is the unknown biodistribution of the PS in the target tissue. In this preliminary study, we report the development of a new approach to image in three dimensions the PS biodistribution in a noninvasive and fast manner.

MATERIALS AND METHODS

A mesoscopic fluorescence tomography imaging platform was used to image noninvasively the biodistribution of 2-[1-hexyloxyethyl]-2 devinyl pyropheophorbide-a (HPPH) in preclinical skin cancer models. Seven tumors were imaged and optical reconstructions were compared to nonconcurrent ultrasound data.

RESULTS

Successful imaging of the HPPH biodistribution was achieved on seven skin cancer tumors in preclinical models with a typical acquisition time of 1 minute. Two-dimensional fluorescence signals and estimated three-dimensional PS distributions were located within the lesions. However, HPPH distribution was highly heterogeneous with the tumors. Moreover, HPPH distribution volume and tumor volume as estimated by ultrasound did not match.

CONCLUSIONS

The results of this proof-of-concept study demonstrate the potential of MFMT to image rapidly the HPPH three-dimensional biodistribution in skin cancers. In addition, these preliminary data indicate that the PS biodistribution in skin cancer tumors is heterogeneous and does not match anatomical data. Mesoscopic fluorescence molecular tomography, by imaging fluorescence signals over large areas with high spatial sampling and at fast acquisition speeds, may be a new imaging modality of choice for planning and optimizing of PDT treatment.

Targeting tumor hypoxia with 2-nitroimidazole-indocyanine green dye conjugates

Tumor hypoxia is a major indicator of treatment resistance to chemotherapeutic drugs, and fluorescence optical tomography has tremendous potential to provide clinically useful, functional information by identifying tumor hypoxia. The synthesis of a 2-nitroimidazole-indocyanine green conjugate using a piperazine linker (piperazine-2-nitroimidazole-ICG) capable of robust fluorescent imaging of tumor hypoxia is described. In vivo mouse tumor imaging studies were completed and demonstrate an improved imaging capability of the new dye relative to an earlier version of the dye that was synthesized with an ethanolamine linker (ethanolamine-2-nitroimidazole-ICG). Mouse tumors located at imaging depths of 1.5 and 2.0 cm in a turbid medium were imaged at various time points after intravenous injection of the dyes. On average, the reconstructed maximum fluorescence concentration of the tumors injected with piperazine-2-nitroimidazole-ICG was twofold higher than that injected with ethanolamine-2-nitroimidazole-ICG within 3 h postinjection period and 1.6 to 1.7 times higher beyond 3 h postinjection. The untargeted bis-carboxylic acid ICG completely washed out after 3 h postinjection. Thus, the optimal window to assess tumor hypoxia is beyond 3 h postinjection. These findings were supported with fluorescence images of histological sections of tumor samples and an immunohistochemistry technique for identifying tumor hypoxia.

Adaptive wide-field optical tomography

We describe a wide-field optical tomography technique, which allows the measurement-guided optimization of illumination patterns for enhanced reconstruction performances. The iterative optimization of the excitation pattern aims at reducing the dynamic range in photons transmitted through biological tissue. It increases the number of measurements collected with high photon counts resulting in a dataset with improved tomographic information. Herein, this imaging technique is applied to time-resolved fluorescence molecular tomography for preclinical studies. First, the merit of this approach is tested by in silico studies in a synthetic small animal model for typical illumination patterns. Second, the applicability of this approach in tomographic imaging is validated in vitro using a small animal phantom with two fluorescent capillaries occluded by a highly absorbing inclusion. The simulation study demonstrates an improvement of signal transmitted (∼2 orders of magnitude) through the central portion of the small animal model for all patterns considered. A corresponding improvement in the signal at the emission wavelength by 1.6 orders of magnitude demonstrates the applicability of this technique for fluorescence molecular tomography. The successful discrimination and localization (∼1  mm error) of the two objects with higher resolution using the optimized patterns compared with nonoptimized illumination establishes the improvement in reconstruction performance when using this technique.

In vivo validation of quantitative frequency domain fluorescence tomography

We have developed a hybrid frequency domain fluorescence tomography and magnetic resonance imaging system (MRI) for small animal imaging. The main purpose of this system is to obtain quantitatively accurate fluorescence concentration and lifetime images using a multi-modality approach. In vivo experiments are undertaken to evaluate the system. We compare the recovered fluorescence parameters with and without MRI structural a priori information. In addition, we compare two optical background heterogeneity correction methods: Born normalization and utilizing diffuse optical tomography (DOT) functional a priori information. The results show that the concentration and lifetime of a 4.2-mm diameter indocyanine green inclusion located 15 mm deep inside a rat can be recovered with less than a 5% error when functional a priori information from DOT and structural a priori information from MRI are utilized.

Temperature-modulated fluorescence tomography based on both concentration and lifetime contrast

It is challenging to image fluorescence objects with high spatial resolution in a highly scattering medium. Recently reported temperature-sensitive indocyanine green-loaded pluronic nanocapsules can potentially alleviate this problem. Here we demonstrate a frequency-domain temperature-modulated fluorescence tomography system that could acquire images at high intensity-focused ultrasound resolution with use of these nanocapsules. The system is experimentally verified with a phantom study, where a 3-mm fluorescence object embedded 2 cm deep in a turbid medium is successfully recovered based on both intensity and lifetime contrast.

Highly efficient detection in fluorescence tomography of quantum dots using time-gated acquisition and ultrafast pulsed laser

Quantum dots (QDs) are widely used in fluorescence tomography due to its unique advantages. Despite the very high quantum efficiency of the QDs, low fluorescent signal and autofluorescence are the most fundamental limitations in optical data acquisition. These limitations are particularly detrimental to image reconstruction for animal imaging, e.g., free-space in vivo fluorescence tomography. In animals studies, fluorescent emission from exogenous fluorescent probes (e.g. QDs) cannot be effectively differentiated from endogenous broad-spectral substances (mostly proteins) using optical filters. In addition, a barrow-band fluorescent filter blocks the majority of the fluorescent light and thus makes signal acquisition very inefficient. We made use of the long fluorescent lifetime of the QDs to reject the optical signal due to the excitation light pulse, and therefore eliminated the need for a fluorescent filter during acquisition. Fluorescent emission from the QDs was excited with an ultrafast pulsed laser, and was detected using a time-gated image intensifier. A tissue-simulating imaging phantom was used to validate the proposed method. Compared to the standard acquisition method that uses a narrow-band fluorescent filter, the proposed method is significantly more efficient in data acquisition (by a factor of >10 in terms of fluorescent signal intensity) and demonstrated reduction in autofluorescence. No additional imaging artifact was observed in the tomographic reconstruction.

Quantitative fluorescence tomography using a trimodality system: in vivo validation

A fully integrated trimodality fluorescence, diffuse optical, and x-ray computed tomography (FT/DOT/XCT) system for small animal imaging is reported in this work. The main purpose of this system is to obtain quantitatively accurate fluorescence concentration images using a multimodality approach. XCT offers anatomical information, while DOT provides the necessary background optical property map to improve FT image accuracy. The quantitative accuracy of this trimodality system is demonstrated in vivo. In particular, we show that a 2-mm-diam fluorescence inclusion located 8 mm deep in a nude mouse can only be localized when functional a priori information from DOT is available. However, the error in the recovered fluorophore concentration is nearly 87%. On the other hand, the fluorophore concentration can be accurately recovered within 2% error when both DOT functional and XCT structural a priori information are utilized together to guide and constrain the FT reconstruction algorithm.

Fluorescence imaging of vascular endothelial growth factor in tumors for mice embedded in a turbid medium

We demonstrate the feasibility of fluorescence imaging of deeply seated tumors using mice injected with an angiogenesis tracer, a vascular endothelial growth factor conjugated with the infrared dye cyanine 7 (VEGF/Cy7). Our optical-only imaging reconstruction method separately estimates the target depth, and then applies this information to reconstruct functional information such as fluorophore concentration. Fluorescence targets with concentrations as low as sub-25 nM are well reconstructed at depths up to 2 cm in both homogeneous and heterogeneous media with this technique.

Three-dimensional reconstruction in free-space whole-body fluorescence tomography of mice using optically reconstructed surface and atlas anatomy

We present a 3-D image reconstruction method for free-space fluorescence tomography of mice using hybrid anatomical prior information. Specifically, we use an optically reconstructed surface of the experimental animal and a digital mouse atlas to approximate the anatomy of the animal as structural priors to assist image reconstruction. Experiments are carried out on a cadaver of a nude mouse with a fluorescent inclusion (2.4-mm-diam cylinder) implanted in the chest cavity. Tomographic fluorescence images are reconstructed using an iterative algorithm based on a finite element method. Coregistration of the fluorescence reconstruction and micro-CT (computed tomography) data acquired afterward show good localization accuracy (localization error 1.2+/-0.6 mm). Using the optically reconstructed surface, but without the atlas anatomy, image reconstruction fails to show the fluorescent inclusion correctly. The method demonstrates the utility of anatomical priors in support of free-space fluorescence tomography.