Fluorescence molecular tomography of DiR-labeled mesenchymal stem cell implants for osteochondral defect repair in rabbit knees.

OBJECTIVES: To assess labelling efficiency of rabbit mesenchymal stem cells (MSCs) using the near-infrared dye 1,1'-dioctadecyl-3,3,3',3'-tetramethylindotricarbocyanine iodide (DiR) and detection of labelled MSCs for osteochondral defect repair in a rabbit model using fluorescence molecular tomography-X-ray computed tomography (FMT-XCT). METHODS: MSCs were isolated from New Zealand White rabbits and labelled with DiR (1.25-20 µg/mL). Viability and induction of apoptosis were assessed by XTT- and Caspase-3/-7-testing. Chondrogenic potential was evaluated by measurement of glycosaminoglycans. Labelled cells and unlabeled controls (n = 3) underwent FMT-XCT imaging before and after chondrogenic differentiation. Osteochondral defects were created surgically in rabbit knees (n = 6). Unlabeled and labelled MSCs were implanted in fibrin-clots and imaged by FMT-XCT. Statistical analyses were performed using multiple regression models. RESULTS: DiR-labelling of MSCs resulted in a dose-dependent fluorescence signal on planar images in trans-illumination mode. No significant reduction in viability or induction of apoptosis was detected at concentrations below 10 µg DiR/mL (p > .05); the chondrogenic potential of MSCs was not affected (p > .05). FMT-XCT of labelled MSCs in osteochondral defects showed a significant signal of the transplant (p < .05) with additional high-resolution anatomical information about its osteochondral integration. CONCLUSIONS: FMT-XCT allows for detection of stem cell implantation within osteochondral regeneration processes. KEY POINTS: • DiR-labelling of MSCs shows no toxic side effects or impairment of chondrogenesis. • Fluorescence molecular tomography allows for detection of MSCs for osteochondral defect repair. • FMT-XCT helps to improve evaluation of cell implantation and osteochondral regeneration processes.

Early recognition of lung cancer by integrin targeted imaging in K-ras mouse model.

Non-small cell lung cancer is characterized by slow progression and high heterogeneity of tumors. Integrins play an important role in lung cancer development and metastasis and were suggested as a tumor marker; however their role in anticancer therapy remains controversial. In this work, we demonstrate the potential of integrin-targeted imaging to recognize early lesions in transgenic mouse model of lung cancer based on spontaneous introduction of mutated human gene bearing K-ras mutation. We conducted ex vivo and fluorescence molecular tomography-X-ray computed tomography (FMT-XCT) in vivo imaging and analysis for specific targeting of early lung lesions and tumors in rodent preclinical model for lung cancer. The lesions and tumors were characterized by histology, immunofluorescence and immunohistochemistry using a panel of cancer markers. Ex vivo, the integrin-targeted fluorescent signal significantly differed between wild type lung tissue and K-ras pulmonary lesions (PL) at all ages studied. The panel of immunofluorescence experiments demonstrated that PL, which only partially show cancer cell features were detected by αvß3-integrin targeted imaging. Human patient material analysis confirmed the specificity of target localization in different lung cancer types. Most importantly, small tumors in the lungs of 4-week-old animals could be noninvasively detected in vivo on the fluorescence channel of FMT-XCT. Our findings demonstrated αvß3-integrin targeted fluorescent imaging to specifically detect premalignant pleural lesions in K-ras mice. Integrin targeted imaging may find application areas in preclinical research and clinical practice, such as early lung cancer diagnostics, intraoperative assistance or therapy monitoring.

Fluorescence-aided tomographic imaging of synovitis in the human finger.

PURPOSE: To propose and evaluate indocyanine green (ICG)-enhanced tomographic optical imaging for detection and characterization of synovitis in affected finger joints of patients with rheumatoid arthritis and differentiation from healthy joints in comparison to 3-T magnetic resonance (MR) imaging. MATERIALS AND METHODS: This prospective pilot study was approved by the institutional ethics committee. Six arthritic proximal interphalangeal (PIP) joints in six patients (five women and one man; mean age ± standard deviation, 62.6 years ± 13.3) with clinically determined rheumatoid arthritis and six healthy PIP joints from six volunteers (four women and two men; mean age, 41.5 years ± 20.2) were examined with an ICG-enhanced fluorescence molecular tomography (FMT) system and 3-T MR imaging as the standard of reference. The degree of inflammation was graded semiquantitatively on a four-point ordinate scale according to the Outcome Measures in Rheumatology Clinical Trials Rheumatoid Arthritis MR Imaging Score, or OMERACT RAMRIS. FMT reconstructions were coregistered with the MR images. Groups were compared by using a two-sided t test, and a weighted κ coefficient was used for comparing FMT and MR imaging semiquantitative scores, as well as assessing intrareader agreement. RESULTS: FMT was used to detect synovitis in all arthritic joints. The reconstructed FMT signal correlated with MR imaging findings in intensity and spatial, transverse profile. Semiquantitative scoring of FMT correlated well with MR imaging findings (weighted κ coefficient = 0.90). The reconstructed quantitative FMT signal, denoting synovial hyperperfusion, was used to differentiate between synovitis and healthy joints (healthy joints, 1.25 ± 0.59; arthritic joints, 3.13 ± 1.03; P < .001). CONCLUSION: FMT enhanced with ICG provided depth-resolved imaging of synovitis in PIP joints. FMT may help detect synovitis in patients with rheumatoid arthritis.

Fast Fourier backprojection for frequency-domain optoacoustic tomography.

We present a time-efficient backprojection image reconstruction approach applied to frequency-domain (FD) optoacoustic tomography based on tissue illumination at multiple, discrete frequencies. The presented method estimates the Fourier transform of a spatial, circular profile of the underlying image using the amplitude and phase data. These data are collected over multiple frequencies using an acoustic transducer positioned at several locations around the sample. Fourier-transform values for absent frequencies are estimated using interpolation based on low-pass filtering in the image domain. Reconstruction results are presented for synthetic measurements using numerical phantoms, and the results are compared with FD model-based reconstructions.

Ex-vivo assessment and non-invasive in vivo imaging of internal hemorrhages in Aga2/+ mutant mice.

Mutations in type I collagen genes (COL1A1/2) typically lead to Osteogenesis imperfecta, the most common heritable cause of skeletal fractures and bone deformation in humans. Heterozygous Col1a1(Aga2/+), animals with a dominant mutation in the terminal C-propeptide domain of type I collagen develop typical skeletal hallmarks and internal hemorrhages starting from 6 day after birth. The disease progression for Aga2/+ mice, however, is not uniform differing between severe phenotype lethal at the 6-11th day of life, and moderate-to-severe one with survival to adulthood. Herein we investigated whether a new modality that combines X-ray computer tomography with fluorescence tomography in one hybrid system can be employed to study internal bleedings in relation to bone fractures and obtain insights into disease progression. The disease phenotype was characterized on Aga2/+ vs. wild type mice between 6 and 9 days postnatal. Anatomical and functional findings obtained in-vivo were contrasted to the ex-vivo appearance of the same tissues under cryo-slicing.

Compression of Born ratio for fluorescence molecular tomography/x-ray computed tomography hybrid imaging: methodology and in vivo validation.

The 360° rotation geometry of the hybrid fluorescence molecular tomography/x-ray computed tomography modality allows for acquisition of very large datasets, which pose numerical limitations on the reconstruction. We propose a compression method that takes advantage of the correlation of the Born-normalized signal among sources in spatially formed clusters to reduce the size of system model. The proposed method has been validated using an ex vivo study and an in vivo study of a nude mouse with a subcutaneous 4T1 tumor, with and without inclusion of a priori anatomical information. Compression rates of up to two orders of magnitude with minimum distortion of reconstruction have been demonstrated, resulting in large reduction in weight matrix size and reconstruction time.

In Vivo Three-Dimensional Raster Scan Optoacoustic Mesoscopy Using Frequency Domain Inversion.

Optoacoustic signals are typically reconstructed into images using inversion algorithms applied in the time-domain. However, time-domain reconstructions can be computationally intensive and therefore slow when large amounts of raw data are collected from an optoacoustic scan. Here we considered a fast weighted ω-k (FWOK) algorithm operating in the frequency domain to accelerate the inversion in raster-scan optoacoustic mesoscopy (RSOM), while seamlessly incorporating impulse response correction with minimum computational burden. We investigated the FWOK performance with RSOM measurements from phantoms and mice in vivo and obtained 360-fold speed improvement over inversions based on the back-projection algorithm in the time-domain. This previously unexplored inversion of in vivo optoacoustic data with impulse response correction in frequency domain reconstructions points to a promising strategy of accelerating optoacoustic imaging computations, toward video-rate tomography.

Skin Surface Detection in 3D Optoacoustic Mesoscopy Based on Dynamic Programming.

Optoacoustic (photoacoustic) mesoscopy offers unique capabilities in skin imaging and resolves skin features associated with detection, diagnosis, and management of disease. A critical first step in the quantitative analysis of clinical optoacoustic images is to identify the skin surface in a rapid, reliable, and automated manner. Nevertheless, most common edge- and surface-detection algorithms cannot reliably detect the skin surface on 3D raster-scan optoacoustic mesoscopy (RSOM) images, due to discontinuities and diffuse interfaces in the image. We present herein a novel dynamic programming approach that extracts the skin boundary as a 2D surface in one single step, as opposed to consecutive extraction of several independent 1D contours. A domain-specific energy function is introduced, taking into account the properties of volumetric optoacoustic mesoscopy images. The accuracy of the proposed method is validated on scans of the volar forearm of 19 volunteers with different skin complexions, for which the skin surface has been traced manually to provide a reference. In addition, the robustness and the limitations of the method are demonstrated on data where the skin boundaries are low-contrast or ill-defined. The automatic skin surface detection method can improve the speed and accuracy in the analysis of quantitative features seen on the RSOM images and accelerate the clinical translation of the technique. Our method can likely be extended to identify other types of surfaces in the RSOM and other imaging modalities.

Deep Learning-Based Spectral Unmixing for Optoacoustic Imaging of Tissue Oxygen Saturation.

Label free imaging of oxygenation distribution in tissues is highly desired in numerous biomedical applications, but is still elusive, in particular in sub-epidermal measurements. Eigenspectra multispectral optoacoustic tomography (eMSOT) and its Bayesian-based implementation have been introduced to offer accurate label-free blood oxygen saturation (sO2) maps in tissues. The method uses the eigenspectra model of light fluence in tissue to account for the spectral changes due to the wavelength dependent attenuation of light with tissue depth. eMSOT relies on the solution of an inverse problem bounded by a number of ad hoc hand-engineered constraints. Despite the quantitative advantage offered by eMSOT, both the non-convex nature of the optimization problem and the possible sub-optimality of the constraints may lead to reduced accuracy. We present herein a neural network architecture that is able to learn how to solve the inverse problem of eMSOT by directly regressing from a set of input spectra to the desired fluence values. The architecture is composed of a combination of recurrent and convolutional layers and uses both spectral and spatial features for inference. We train an ensemble of such networks using solely simulated data and demonstrate how this approach can improve the accuracy of sO2 computation over the original eMSOT, not only in simulations but also in experimental datasets obtained from blood phantoms and small animals (mice) in vivo. The use of a deep-learning approach in optoacoustic sO2 imaging is confirmed herein for the first time on ground truth sO2 values experimentally obtained in vivo and ex vivo.

Compensation of optical heterogeneity-induced artifacts in fluorescence molecular tomography: theory and in vivo validation.

We present a method for reduction of image artifacts induced by the optical heterogeneities of tissue in fluorescence molecular tomography (FMT) through identification and compensation of image regions that evidence propagation of emission light through thin or low-absorption tunnels in tissue. The light tunneled as such contributes to the emission image as spurious components that might substantially overwhelm the desirable fluorescence emanating from the targeted lesions. The proposed method makes use of the strong spatial correlation between the emission and excitation images to estimate the tunneled components and yield a residual image that mainly consists of the signal due to the desirable fluorescence. This residual image is further refined using a coincidence mask constructed for each excitation-emission image pair. The coincidence mask is essentially a map of the "hot spots" that occur in both excitation and emission images, as such areas are often associated with tunneled emission. In vivo studies are performed on a human colon adenocarcinoma xenograft tumor model with subcutaneous tumors and a murine breast adenocarcinoma model with aggressive tumor cell metastasis and growth in the lungs. Results demonstrate significant improvements in the reconstructions achieved by the proposed method.

Fully automated identification of skin morphology in raster-scan optoacoustic mesoscopy using artificial intelligence.

PURPOSE: Identification of morphological characteristics of skin lesions is of vital importance in diagnosing diseases with dermatological manifestations. This task is often performed manually or in an automated way based on intensity level. Recently, ultra-broadband raster-scan optoacoustic mesoscopy (UWB-RSOM) was developed to offer unique cross-sectional optical imaging of the skin. A machine learning (ML) approach is proposed here to enable, for the first time, automated identification of skin layers in UWB-RSOM data. MATERIALS AND METHODS: The proposed method, termed SkinSeg, was applied to coronal UWB-RSOM images obtained from 12 human participants. SkinSeg is a multi-step methodology that integrates data processing and transformation, feature extraction, feature selection, and classification. Various image features and learning models were tested for their suitability at discriminating skin layers including traditional machine learning along with more advanced deep learning algorithms. An support vector machines-based postprocessing approach was finally applied to further improve the classification outputs. RESULTS: Random forest proved to be the most effective technique, achieving mean classification accuracy of 86.89% evaluated based on a repeated leave-one-out strategy. Insights about the features extracted and their effect on classification accuracy are provided. The highest accuracy was achieved using a small group of four features and remained at the same level or was even slightly decreased when more features were included. Convolutional neural networks provided also promising results at a level of approximately 85%. The application of the proposed postprocessing technique was proved to be effective in terms of both testing accuracy and three-dimensional visualization of classification maps. CONCLUSIONS: SkinSeg demonstrated unique potential in identifying skin layers. The proposed method may facilitate clinical evaluation, monitoring, and diagnosis of diseases linked to skin inflammation, diabetes, and skin cancer.

Detection of intramyocardially injected DiR-labeled mesenchymal stem cells by optical and optoacoustic tomography.

The distribution of intramyocardially injected rabbit MSCs, labeled with the near-infrared dye 1,1'-dioctadecyl-3,3,3',3'-tetramethylindotricarbo-cyanine-iodide (DiR) using hybrid Fluorescence Molecular Tomography-X-ray Computed Tomography (FMT-XCT) and Multispectral Optoacoustic Tomography (MSOT) imaging technologies, was investigated. Viability and induction of apoptosis of DiR labeled MSCs were assessed by XTT- and Caspase-3/-7-testing in vitro. 2 × 106, 2 × 105 and 2 × 104 MSCs labeled with 5 and 10 µg DiR/ml were injected into fresh frozen rabbit hearts. FMT-XCT, MSOT and fluorescence cryosection imaging were performed. Concentrations up to 10 µg DiR/ml did not cause apoptosis in vitro (p > 0.05). FMT and MSOT imaging of labeled MSCs led to a strong signal. The imaging modalities highlighted a difference in cell distribution and concentration correlated to the number of injected cells. Ex-vivo cryosectioning confirmed the molecular fluorescence signal. FMT and MSOT are sensitive imaging techniques offering high-anatomic resolution in terms of detection and distribution of intramyocardially injected stem cells in a rabbit model.

An Inversion Scheme for Hybrid Fluorescence Molecular Tomography Using a Fuzzy Inference System.

The imaging performance of fluorescence molecular tomography (FMT) improves when information from the underlying anatomy is incorporated into the inversion scheme, in the form of priors. The requirement for incorporation of priors has recently driven the development of hybrid FMT systems coupled to other modalities, such as X-ray CT and MRI. A critical methodological aspect in this modality relates to the particular method selected to incorporate prior information obtained from the anatomical imaging modality into the FMT inversion. We propose herein a new approach for utilizing prior information, which preferentially minimizes residual errors associated with measurements that better describe the anatomical segments considered. This preferential minimization was realized using a weighted least square (WLS) approach, where the weights were optimized using a Mamdani-type fuzzy inference system. The method of priors introduced herein was deployed as a two-step structured regularization approach and was verified with experimental measurements from phantoms as well as ex vivo and in vivo animal studies. The results demonstrate accurate performance and minimization of reconstruction bias, without requiring user input for setting the regularization parameters. As such, the proposed method offers significant progress in incorporation of anatomical priors in FMT and, as a result, in realization of the full potential of hybrid FMT.

Frequency domain optoacoustic tomography using amplitude and phase.

We introduce optoacoustic tomographic imaging using intensity modulated light sources and collecting amplitude and phase information in the frequency domain. Imaging is performed at multiple modulation frequencies. The forward modeling uses the Green's function solution to the pressure wave equation in frequency domain and the resulting inverse problem is solved using regularized least squares minimization. We study the effect of the number of frequencies and of the bandwidth employed on the image quality achieved. The possibility of employing an all-frequency domain optoacoustic imaging for experimental measurements is studied as a function of noise. We conclude that frequency domain optoacoustic tomography may evolve to a practical experimental method using light intensity modulated sources, with advantages over time-domain optoacoustics.

Limited-projection-angle hybrid fluorescence molecular tomography of multiple molecules.

An advantage of fluorescence methods over other imaging modalities is the ability to concurrently resolve multiple moieties using fluorochromes emitting at different spectral regions. Simultaneous imaging of spectrally separated agents is helpful in interrogating multiple functions or establishing internal controls for accurate measurements. Herein, we investigated multimoiety imaging in the context of a limited-projection-angle hybrid fluorescence molecular tomography (FMT), and x-ray computed tomography implementation and the further registration with positron emission tomography (PET) data. Multichannel FMT systems may image fluorescent probes of varying distribution patterns. Therefore, it is possible that different channels may require different use of priors and regularization parameters. We examined the performance of automatically estimating regularization factors implementing priors, using data-driven regularization specific for limited-projection-angle schemes. We were particularly interested in identifying the implementation variations between hybrid-FMT channels due to probe distribution variation. For this reason, initial validation of the data-driven algorithm on a phantom was followed by imaging different agent distributions in animals, assuming superficial and deep seated activity. We further demonstrate the benefits of combining hybrid FMT with PET to gain multiple readings on the molecular composition of disease.

Multimodal molecular imaging of integrin αvß3 for in vivo detection of pancreatic cancer.

UNLABELLED: Pancreatic ductal adenocarcinoma (PDAC) is a lethal disease. Late detection of then nonresectable or metastasized tumors emphasizes the need for novel imaging approaches. Here, we report on so far nonexploited potentials of αvß3 integrin-targeted molecular imaging technologies for detection of PDAC using genetically engineered mouse models. METHODS: Immunohistochemistry and Western blot were used for characterization of αvß3 expression in murine and human PDAC. We applied IntegriSense 680 fluorescence molecular tomography, intraoperative fluorescence imaging, and (68)Ga-NODAGA-RGD PET for αvß3 integrin molecular in vivo imaging of spontaneous PDAC occurring in Ptf1a(+/Cre);Kras(+/LSL-G12D);p53(LoxP/LoxP) mice. (NODAGA is 1,4,7-triazacyclononane-1,4-bis[acetic acid]-7-[2-glutaric acid] and RGD is arginine-glycine-aspartic acid.) RESULTS: αvß3 integrin is expressed in tumor cells of human and murine PDAC. IntegriSense fluorescence molecular tomography and (68)Ga-NODAGA-RGD PET enabled faithful visualization of PDAC. Furthermore, intraoperative optical imaging with IntegriSense 680 allowed good delineation of tumor borders. CONCLUSION: Imaging approaches targeting αvß3 integrin expand the potential of molecular imaging for identification of αvß3-positive PDAC with potential implications in early detection, fluorescence-guided surgery, and therapy monitoring.

FMT-PCCT: hybrid fluorescence molecular tomography-x-ray phase-contrast CT imaging of mouse models.

The implementation of hybrid fluorescence molecular tomography (FMT) and X-ray computed tomography (CT) has been shown to be a necessary development, not only for combining anatomical with functional and molecular contrast, but also for generating optical images of high accuracy. FMT affords highly sensitive 3-D imaging of fluorescence bio-distribution, but in stand-alone form it offers images of low resolution. It was shown that FMT accuracy significantly improves by considering anatomical priors from CT. Conversely, CT generally suffers from low soft tissue contrast. Therefore utilization of CT data as prior information in FMT inversion is challenging when different internal organs are not clearly differentiated. Instead, we combined herein FMT with emerging X-ray phase-contrast CT (PCCT). PCCT relies on phase shift differences in tissue to achieve soft tissue contrast superior to conventional CT. We demonstrate for the first time FMT-PCCT imaging of different animal models, where FMT and PCCT scans were performed in vivo and ex vivo, respectively. The results show that FMT-PCCT expands the potential of FMT in imaging lesions with otherwise low or no CT contrast, while retaining the cost benefits of CT and simplicity of hybrid device realizations. The results point to the most accurate FMT performance to date.

Spatiotemporal analysis for indocyanine green-aided imaging of rheumatoid arthritis in hand joints.

Rheumatoid arthritis (RA) is the most common chronic inflammatory joint disease, with a prevalence of 0.5 to 1% in the general population. Imaging can possibly aid in early diagnosis, crucial to effective personalized therapeutic strategies and treatment follow-up. The intravenous administration of indocyanine green (ICG) has been considered for identifying synovial hyperperfusion as an RA physiological biomarker. However, while the distribution of ICG in the human hand is a time-dependent process, the particular biodistribution dynamic patterns established following intravenous administration have not yet been studied. For this reason, the dynamic relationships of ICG distribution in the human hand in RA patients using a method based on principal component analysis are analyzed. In vivo analyses were corroborated by simulations of clinical scenarios using a finite element method. Observations of spatiotemporal characteristics are contrasted to fluorescence intensity images and magnetic resonance images of the hand joints, employed as the anatomical and diagnostic reference. Processing results for 450 joints from 5 healthy volunteers and 10 patients show that image features obtained from the spatiotemporal analysis offer good congruence with synovitis and reveal better detection performance compared to observations of raw fluorescence intensity images.

Object localization in the presence of a strong heterogeneous background in fluorescent tomography.

We propose a method for object localization in fluorescent tomography (FT) in the presence of a highly heterogeneous background. Existing approaches typically assume a homogeneous background distribution; thus, they are incapable of accurately accounting for the more general case of an unconstrained, possibly heterogeneous, background. The proposed method iteratively solves the inverse problem over a solution space partitioned into a background subspace and an object subspace to simultaneously estimate the background and localize the target fluorescent objects. Simulation results of this algorithm applied to continuous-wave FT demonstrate effective localization of target objects in the presence of highly heterogeneous background distributions.

Optimal sparse solution for fluorescent diffuse optical tomography: theory and phantom experimental results.

We present a method to accurately localize small fluorescent objects within the tissue using fluorescent diffuse optical tomography (FDOT). The proposed method exploits the localized or sparse nature of the fluorophores in the tissue as a priori information to considerably improve the accuracy of the reconstruction of fluorophore distribution. This is accomplished by minimizing a cost function that includes the L1 norm of the fluorophore distribution vector. Experimental results for a milk-based phantom using a fiber-based cw FDOT system demonstrate the capability of this method in accurately localizing small fluorescent objects deep in the phantom.