Multispectral Opto-acoustic Tomography (MSOT) of the Brain and Glioblastoma Characterization

Abstract

Brain research depends strongly on imaging for assessing function and disease in vivo. We examine herein multispectral opto-acoustic tomography (MSOT), a novel technology for high-resolution molecular imaging deep inside tissues. MSOT illuminates tissue with light pulses at multiple wavelengths and detects the acoustic waves generated by the thermoelastic expansion of the environment surrounding absorbing molecules. Using spectral unmixing analysis of the data collected, MSOT can then differentiate the spectral signatures of oxygenated and deoxygenated hemoglobin and of photo-absorbing agents and quantify their concentration. By being able to detect absorbing molecules up to centimeters deep in the tissue it represents an ideal modality for small animal brain imaging, simultaneously providing anatomical, hemodynamic, functional, and molecular information. In this work we examine the capacity of MSOT in cross-sectional brain imaging of mice. We find unprecedented optical imaging performance in cross-sectional visualization of anatomical and physiological parameters of the mouse brain. For example, the potential of MSOT to characterize ischemic brain areas was demonstrated through the use of a carbon dioxide challenge. In addition, indocyanine green (ICG) was injected intravenously, and the kinetics of uptake and clearance in the vasculature of the brain was visualized in real-time. We further found that multiparameter, multispectral imaging of the growth of U87 tumor cells injected into the brain could be visualized through the intact mouse head, for example through visualization of deoxygenated hemoglobin in the growing tumor. We also demonstrate how MSOT offers several compelling features for brain research and allows time-dependent detection and quantification of brain parameters that are not available using other imaging methods without invasive procedures.

Introduction

Neuroimaging has transformed brain research and clinical neurology by enabling early diagnosis of disease and critical feedback on the efficacy of treatments (Hargreaves, 2008, Wong et al., 2009). In addition, non-invasive imaging has enabled correlations between brain function and behavior (Logothetis, 2008). However, there is always a need to improve upon current technologies. Innovation, driven by assessing and offering alternatives to the limitations of current technologies, has and will continue to drive new discoveries in neuroimaging. There are currently many modalities available for neuroimaging, each with their own benefits and limitations. X-ray-based computed tomography (CT) and magnetic resonance imaging (MRI) provide anatomical information at high spatial resolution but with limited molecular specificity (Histed et al., 2012). On the other hand, positron emission tomography (PET) and fluorescence molecular tomography (FMT) are molecular imaging modalities (Ntziachristos and Razansky, 2010), but have low resolution. Intravital microscopy has molecular specificity at high resolution, but is limited by shallow tissue penetration and requires invasive procedures (Lichtman and Fraser, 2001).

We studied the performance of multispectral opto-acoustic tomography (MSOT) for neuroimaging in mice. MSOT combines the high resolution of anatomical techniques such as MRI, the molecular specificity of PET or FMT and the contrast of optical imaging to offer a highly potent modality that has been implemented recently for small animal imaging in real-time (Buehler et al., 2010, Ntziachristos and Razansky, 2010). It therefore offers several attractive characteristics for studying brain function and disease. Opto-acoustic imaging is based on the generation of acoustic waves following the absorption of light pulses of short (10–100 ns) duration by photo-absorbing molecules or nanoparticles. Transient absorption of light leads to a transient temperature increase in the mK range (Xiang et al., 2007) which in turn gives rise to a thermoelastic tissue expansion that produces ultrasonic waves. By detecting the acoustic waves using ultrasonic detectors and subsequent mathematical inversion, the distribution of optical absorption can be reconstructed in tissues with ultrasound imaging resolution, i.e. 20–200 μm. Single-wavelength images can yield anatomical information on the absorption contrast of different tissue structures. Illumination at multiple wavelengths and the spectral processing of the data captured can further lead to resolving spectral signatures of tissue molecules, extrinsically administered absorbing agents and nanoparticles with certain spectral signatures. Deoxygenated and oxygenated hemoglobin have substantial, unique absorptions in the NIR spectrum and therefore can be resolved with multispectral decomposition of opto-acoustic signals (Brecht et al., 2007, Esenaliev et al., 2002, Wang et al., 2006). Interestingly, injected agents can be distinguished from intrinsic absorbers. For example, perfusion of ICG through the kidney vasculature has previously been documented by MSOT (Buehler et al., 2010, Taruttis et al., 2012). Spectral separation (decomposition) also makes it possible to differentiate vascular MSOT signals derived from injected gold nanoparticles versus those derived from oxygenated and deoxygenated blood (Herzog et al., 2012, Taruttis et al., 2010). By illuminating at different wavelengths, spectrally tunable gold nanorods with different absorption maxima can simultaneously be detected in vivo (Li et al., 2008). Spectral shifts associated with molecular modifications can also be exploited for contrast generation. It was recently demonstrated that MSOT could visualize spectral modifications in response to NIR-probe activation by matrix metalloproteinases, detected ex vivo in human carotid arteries bearing atherosclerotic plaques (Razansky et al., 2012). In addition, sequential 2D slices can be compiled to produce volumetric quantitative molecular imaging in entire organs, small animals, or human tissues. The objective of the experiments herein is to evaluate the performance of MSOT and the potential utility of this technology in the investigation of molecular imaging biomarkers in diseases of the central nervous system.