This review explains the fundamental principles of Multispectral Optoacoustic Tomography (MSOT) and potential applications addressing unmet clinical needs
Vasilis Ntziachristos PhD
Barbara Schröder PhD
Chair for Biological Imaging (CBI)
Technische Universität München, Munich, Germany
Institute for Biological and Medical
Imaging (IBMI), Helmholtz Zentrum München, Neuherberg, Germany
E-mail: [email protected]
Optical visualisation is probably the longest and most established imaging practice in medicine. Optical inspection of a patient is routinely used worldwide during physical examination, surgical intervention, endoscopy, dermoscopy, dentistry or ophthalmology. However, most optical observations in the clinics today are based on the same method used since the beginning of medicine: human vision.
Regardless of directly looking onto tissue or through an optical device or video scope, the human eye is unable to visualise tissue contrast under the surface and offers only limited spectroscopic ability. For centuries, the major impediment of optical methods has been photon scattering in tissues, which makes them opaque and restricts histological imaging to slices only a few microns across. However, recent technical innovations and advancements finally bring an eagerly anticipated change. Optoacoustic (photoacoustic) imaging offers a long needed fundamental solution to the physical limitations of optical imaging by enabling scatter-insensitive visualisation of optical contrast deep inside tissues.
This revolutionary technology is based on ‘listening’ to the absorbed light in tissue instead of ‘looking’ at it. Cross-sectional or three-dimensional optical images can therefore be reconstructed with resolutions equal to those of ultrasound and depths of a few centimetres. This new ability transforms photoacoustic imaging into a powerful imaging modality especially for clinical applications.
Vision plays a fundamental role in how humans perceive and operate. The ability to quickly process and understand complex visual patterns is powerful, but may be misleading when it comes to medical procedures. Part of the medical routine is to have a physician look at tissue, for example when inspecting for oral cancer, in colposcopy, in colonoscopy or in dermatology, using a magnifying camera. It appears similarly common to have a surgeon look at tissue through a microscope or a video laparoscope and make decisions on tissue excision based on morphological features seen on the tissue surface In addition, human vision lacks the ability to obtain information on sub-surface lesions and infiltrating cancer borders or tissue structures such as nerves and blood vessels.
New technological trends improve the resolution and image quality of medical scopes but none of these methods achieves penetration deep under the tissue surface. In addition, important pathophysiological parameters, such as tissue oxygenation or the microvascular bed in host tissues receiving engraftments during plastic surgery, are not assessed today.
Photo credit: Fotolia, Sebastian Kaulitzki, und ediundsepp.
All these limitations come from a single parameter: photon scatter. Photons that enter tissue scatter multiple times in random angles (~10 times per mm) losing directionality and creating a diffusive image, restricting vision in ways similar to the effects seen on a heavy fog day. Although light in the near-infrared range can penetrate several centimetres deep into the tissue, photon scatter does not allow retrieving high resolution images from these depths when using an optical detector. In fact, in ambient light conditions, scatter reflects a portion of the incident light toward our eyes, blocking the ability to visualise deeper than the tissue surface. Several optical microscopy methods have been developed to account for the effects of scatter, but they cannot offer three-dimensional tissue images beyond ~1mm of depth.
An emerging novel technology termed optoacoustic imaging (also called photoacoustic imaging) solves the penetration limitations of optical imaging by using a fundamentally different detection mechanism of optical contrast: sound. Optoacoustic imaging is insensitive to photon scattering thus allowing three-dimensional optical imaging at high resolution: the murky image clears. Improved resolution in turn is the basis for better quantification compared to conventional optical imaging.
In addition, an important feature for clinical applications is the implementation of optoacoustic imaging at different illuminating wavelengths, thereby enabling the simultaneous detection of multiple absorption spectra from different tissue absorbers, including oxygenated or deoxygenated haemoglobin, melanin and other intrinsic or extrinsically administered chromophores. Consequently, Multispectral Optoacoustic Tomography (MSOT) combines the spectral dimension from optical imaging with the accurate image reconstruction and real-time imaging performance of ultrasound (Figure 1) that leads to an unprecedented potential in biological and clinical sensing.
Fig 1: Handheld Multispectral Optoacoustic Tomography (MSOT) for clinical use. (a) Schematic of a real-time MSOT scanner developed for tissue sensing and imaging (modified from [1]). (b) Principle of MSOT operation. Photon pulses at different wavelengths illuminate tissue. The absorption of light by tissue generates ultrasound waves within the tissue, which are detected by ultrasound sensors placed at different positions around the object. Utilisation of image reconstruction techniques can then resolve the ultrasound sources in tissue, that is, resolve optical absorption. (c) By using spectral unmixing techniques, measurements at different wavelengths are employed to resolve the spectra of different tissue chromophores, including oxygenated (HbO2) or deoxygenated haemoglobin (Hb), melanin, potentially other tissues naturally occurring chromophores or extrinsically administered chromophoric agents such as dyes, fluorescent agents and photo-absorbing nanoparticles (‘probe’). Images c and d: courtesy of Dr Nicolas Bezière and Dr Andreas Buehler; Institute of Biological and Medical Imaging, Helmholtz Zentrum Munich, Germany.
Optoacoustic methods have been considered for biomedical applications since the 1970s and several aspects of the technology have been reviewed in the literature.1–10 The emergence of clinical MSOT over the past five years is due to significant progress in the technology of MSOT components and MSOT integration into portable handheld systems that allow real-time multispectral operation for the first time. Notable recent progress includes advances in laser technology, novel detectors as well as image reconstruction and spectral unmixing methods that offer unprecedented accuracy and image quality. In the following, we review the principle of operation and the clinical abilities offered by MSOT.
‘Listening to light’ – the MSOT principle
MSOT operates on the photoacoustic phenomenon, whereby absorption of nanosecond light pulses in tissue leads to minute thermoelastic tissue expansion, which generates weak acoustic (ultrasound) waves. The ultrasound waves are detected at the boundary of the tissue and tomographically reconstructed to give optoacoustic imaging its unique characteristics: three-dimensional (3D) high resolution optical images much beyond the depths achieved by optical microscopy.1 Optoacoustic imaging can generate morphological (anatomical) images of tissue vasculature (see Figure 2) and of absorbing structures and organs by resolving tissue absorption at a particular wavelength. However, by using illumination at multiple wavelengths and employing spectral unmixing techniques, MSOT can also resolve physiological and molecular tissue images.
Fig. 2: Advances in optoacoustic imaging. Microvascular imaging of (a) a mouse ear (courtesy of Dr Murad Omar) and (b) a mouse melanoma vascular bed using high resolution optoacoustic imaging at <10 micron axial resolution through >4mm of depth (adapted from [2], courtesy of Dr Murad Omar). (c) Cross-sectional optoacoustic image of forearm vessels visualised in a human volunteer using the (d) a portable handheld MSOT scanner (e) Non-invasive MSOT image of the human carotid in a healthy volunteer (from [3]). (f) Cross-sectional image through the lower abdomen of an entire mouse (grey scale) and superimposed colour image of tissue oxygenation (red 100% HbO2; yellow/green: hypoxia) within and around a subcutaneous tumour grown subcutaneously (courtesy of Stratis Tzoumas). The image shows the MSOT ability to resolve oxygenation (hypoxia) images of tissues and tumours in label-free mode.
Using identification of the absorption spectra of different tissue moieties, MSOT provides not only images of tissue and vascular morphology, but also molecular images of chromophores such as melanin or oxy- and deoxy- haemoglobin distribution. MSOT therefore offers functional and molecular imaging capabilities in label free mode, that is, without the need of contrast agents, although the use of chromophoric contrast agents can further extend the MSOT applications. For accurate results it is important to utilise multiple wavelengths, typically five or six lines for every 100nm of spectrum covered.
MSOT features over other imaging modalities
Radiological modalities such as X-ray, Computed Tomography (CT), Magnetic Resonance Imaging (MRI), Positron Emission Tomography (PET), Single Photon Emission Computed Tomography (SPECT) or Ultrasonography (US) are playing a major role in clinical diagnostics, theranostics and therapy follow-up. While these methods enable morphological, functional and molecular readings from tissues, several limitations exist. With the exception of ultrasound imaging, radiology typically uses large systems that challenge highly disseminated use (that is, in surgery or ambulatory suites).
Portable gamma cameras are employed for lymph node identification, but radioactivity limits their broad use to bedside or ambulatory care. Ultrasound is well suited for low-cost, portable use and has found many applications outside of radiology, for example in cardiology or neonatal imaging. Nevertheless, several basic diagnostic and clinical care activities today, for example, in surgery, cardiology and beyond require an extended list of morphological and functional measurements, such as real-time imaging of tissue oxygenation, which cannot be provided by the currently available modalities.
MSOT has several advantages compared to established radiological modalities. These advantages relate to the unique contrast visualised by MSOT, that is, optical (photon) absorption versus for example, sound reflections in ultrasound imaging, nuclear magnetic resonance in MRI or the radioactivity of an agent in SPECT and PET. The major difference compared to other optical methods that may also be sensitive to optical absorption (for example, white light imaging) is that MSOT resolves optical absorption not only superficially but also deep within tissue at high resolution, without being significantly affected by photon scatter. Optical absorption relates to important morphological and physiological tissue features that may not be assessed by other imaging modalities, as summarised below:
Characterisation of tissue oxygenation, hypoxia and ischaemia
MSOT is a modality that offers cross-sectional high resolution images of oxygenated and deoxygenated haemoglobin. Based on these images we can then calculate oxygen saturation maps, a measure of tissue oxygenation/hypoxia in vivo and haemoglobin (blood) concentration. The acquired MSOT images can be employed to directly assess tissue ischaemia and infer tissue perfusion for each pixel of the multi-wavelength images collected. The assessment of oxygenated and deoxygenated haemoglobin is done in a label-free mode, that is, without the need to administer contrast agents. No other imaging method offers such performance. MSOT can therefore bring tissue oxygenation images (not only single-point pulse oximeter measured arterial saturation) into ambulatory, vascular and interventional care.
Imaging vascularisation, micro-vascularisation and small vessel flow
Optoacoustic offers highly scalable label-free in vivo imaging of blood vessels ranging from centimetres to microns in diameter, much deeper than what is achieved by conventional optical microscopy and other optical imaging methods. Ultrasound can see flow/perfusion and large vessels, whereas MSOT is better at visualising small vessels in label-free mode (see Figure 2). In addition, recent advances in MSOT methods enable flow measurements even in micron sized vessels,4 non-invasively revealing micro-circulation parameters.
Novel inferred contrast: metabolism and inflammation
Besides the applications listed above, MSOT can capitalise on its unique contrast mechanisms and provide insights into a much larger number of different physiological processes. Inflammation, for example, is associated with vascular and blood volume changes, which can be assessed by MSOT in label-free mode, potentially providing quantitative characterisation of its extent and progression. Likewise, metabolic processes can be assessed by quantifying oxygen utilisation as well as vascular and flow changes.
Dynamic studies
MSOT has already been implemented in video capture mode, offering 5–20 frames per second,9 but it is also possible to implement at a much higher sampling rate, depending on the repetition rate of the light source employed. Therefore, the technology is well suited to study dynamic phenomena,6 as shown in studies assessing kidney clearance of fluorescent agents (Figure 3).
Fig. 3: Dynamic MSOT studies. (a) Time series of images visualising the biodistribution of the fluorescent agent IRDye800CW in green on logarithmic scale overlaid on the vasculature. Both channels are the result of spectral unmixing. (b) Cryoslice image after approximately 15 minutes with overlaid fluorescence as a verification of the MSOT results. (c) A comparison of fluorescence distribution in the kidneys of mice sacrificed after approximately 2, 5 minutes (left) and 15 minutes after injection (right). Note the changes in distribution similar to the time series shown in (a). (d) Temporal evolution of signal (each normalised to their smoothed maxima) in the regions of interest highlighted in the rightmost image, orange showing a region in the renal cortex that displays early and steep signal pickup (ROI1) and black indicating a region in the renal pelvis where probe accumulation is delayed and has a smoother profile (ROI2). Time points of the images in (a) are marked using vertical lines. Reproduced from [6]. ROI = region of interest.
Imaging of melanin
As melanin presents with a strong optoacoustic signal, MSOT has recently been used to assess the sentinel lymph node status and to determine the metastatic status in 214 melanoma patients.10 Importantly, MSOT markedly improved the detection rate of metastasised melanoma cells in the excised node compared to the current standard methods.10 In addition, the node status was assessed non-invasively in 20 patients showing 100% sensitivity and 48–62% specificity. This result fuels excitement for MSOT-based applications in dermatology.
Imaging of extrinsically administered contrast agents and nanoparticles
As mentioned before, MSOT applications can be significantly enhanced by the use of external probes such as fluorescent agents, dyes and photo-absorbing nanoparticles. The dynamic studies in Figure 3 were enabled by the use of a fluorescent agent clearing through the kidneys. Targeted agents have also been used for molecular imaging applications whereby specific receptors, cells or processes can be labelled in vivo and resolved by MSOT.1 Likewise, MSOT has been proven very potent in visualising the bio-distribution of gold nanoparticles or nanoparticles labelled with fluorescent agents or absorbing dyes. With the recent propagation of fluorescent agents for clinical use and the overall growth of nanomedicine, it is expected that MSOT can be employed in agent characterisation as well as diagnostic and theranostic applications.
Quantification and standardisation
Quantification is important in precision and personalised medicine and can help standardise diagnosis and the quality of care within and across clinical centres worldwide. MSOT uses tomography1 based on physical models of photon and sound propagation.5 Model-based inversion mathematics allows quantification of the physical problem. Therefore MSOT can be employed to resolve the contrast listed above in a quantitative manner.
Conclusion
Many radiology modalities require ionising radiation, contrast agents or are inappropriate for ambulatory or bedside measurements due to cost/size. Since MSOT uses safe light energy, portable instrumentation, fast (video-rate) scanning and label-free pathophysiology metrics it can be employed for long-term applications in ambulatory, intraoperative or bedside settings, possibly bringing new clinical ability and novel insights into disease development, basic/clinical discovery and disease or treatment monitoring. It is expected that the unique contrast and unprecedented performance of MSOT1 will generate a new clinical modality serving unmet clinical need in a portable and cost effective manner.
MSOT is now making its first steps into clinical measurements. We anticipate that it will engage in several applications and measurements associated with vascular, microcirculation, tissue oxygenation and melanin-related clinical needs. In addition, by using external agents, we expect the emergence of a much larger application range. Taken together, the implementation of optoacoustic methods in clinical workflow can greatly improve a broad spectrum of clinical applications where visual inspection is central to current diagnosis, treatment and monitoring/follow up.
It is important that the integration of handheld MSOT scanners not only provides benefit to the patient (due to improved diagnostics, treatment, monitoring etc.), but it is performed within scenarios that reduce costs through faster and quantitative disease assessment, reducing the number of unnecessary secondary procedures and improving treatment planning.
References
- Taruttis A, Ntziachristos V. Advances in real-time multispectral optoacoustic imaging and its applications. Nat Photon 2015;9:219–27.
- Omar M et al. Pushing the Optical Imaging Limits of Cancer with Multi-Frequency-Band Raster-Scan Optoacoustic Mesoscopy (RSOM). Neoplasia 2015;17(2):208–14.
- Dima A, Ntziachristos V. Non-invasive carotid imaging using optoacoustic tomography. Optics Express 2012;20(22):25044–57.
- Yao J et al. In vivo photoacoustic imaging of transverse blood flow using Doppler broadening of bandwidth. Opt Lett 2010;35(9):1419–21.
- Mohajerani P et al. Biometric Template Protection: Bridging the performance gap between theory and practice. IEEE Signal Proc Mag 2015;32:88–100.
- Taruttis A et al. Fast Multispectral Optoacoustic Tomography (MSOT) for Dynamic Imaging of Pharmacokinetics and Biodistribution in Multiple Organs. PLoS ONE 2012;7(1):e30491.
- Razansky D, Buehler A, Ntziachristos V. Volumetric real-time multispectral optoacoustic tomography of biomarkers. Nat Protocols 2011;6:1121–9.
- Tzoumas S et al. Unmixing Molecular Agents From Absorbing Tissue in Multispectral Optoacoustic Tomography. IEEE T Med Imaging 2014;33:48–60.
- Buehler A et al. Real-time handheld multispectral optoacoustic imaging. Opt Lett 2013;38(9):1404–6.
- Stoffels I et al. Metastatic status of sentinel lymph nodes in melanoma determined noninvasively with multispectral optoacoustic imaging. Sci Trans Med 2015;7(317):317ra199.
