Apparatus and method for optoacoustic sensing of tissue features in biological tissue and method for processing optoacoustic signals

Abstract

The disclosure relates to an apparatus for optoacoustic sensing of tissue features, such as tissue architecture, morphological information, functional, physiological and / or molecular information regarding biomolecules, in particular metabolites, in biological tissue (1) and to according methods. The apparatus comprises: at least one irradiation device (2) configured to irradiate the tissue (1) with a first electromagnetic radiation (3) exhibiting one or more wavelengths or wavelength ranges in a first spectral range; at least one detection device (5) configured to detect at least one first optoacoustic signal corresponding to ultrasound waves (6) emanating from a region of interest (ROI) in the tissue (1) in response to irradiating the tissue (1) with the first electromagnetic radiation (3); and a processor (7) configured to determine, based on the first optoacoustic signal, at least one first parameter characterizing a first tissue feature of interest in the region of interest (ROI).

Description

[0001] APPARATUS AND METHOD FOR OPTOACOUSTIC SENSING OF TISSUE FEATURES IN BIOLOGICAL TISSUE AND METHOD FOR PROCESSING OPTOACOUSTIC SIGNALS

[0002] Description

[0003] The present disclosure relates to an apparatus and a method for optoacoustic sensing of tissue features in biological tissue and a method for processing optoacoustic signals.

[0004] Currently, glucose monitoring can be based on minimally invasive transdermal sensing, wherein a small enzyme-based electrode is inserted under the skin for sensing glucose in the interstitial fluid (ISF). However, the glucose concentrations in ISF (or in sweat, tears, etc.) do not directly represent the concentration of glucose in the blood, as glucose enters these fluids by diffusion from surrounding capillaries, which lead to 10x glucose dilution in ISF compared to blood. Moreover, different ISF volumes and variations in the skin hydration level result in different dilution volumes that limit the accuracy of the measurement. Other limitations include the signal fluctuations with pH, skin irritation and the need to replace the sensor after a period of use. Glucose concentrations in saliva and tears correlate even more poorly (>1000x dilution) with blood glucose. Therefore, monitoring glucose in bodily fluids other than blood limits the accuracy and reproducibility of such methods and requires frequent calibration with invasive finger-prick measurements. These limitations apply not only to glucose sensing but to other biomolecules that can be similarly detected with electrochemical methods. Moreover, electrode-based sensors are not appropriate for multiple metabolite monitoring so that only one metabolite can be monitored at a time, and they require micro-needles or bio-fluid collection devices (e.g. sweat sampling) that may lead to skin irritation and may require frequent replacement of the micro-needle or collection device. Stacking multiple micro-electrodes can result in sensing more than one metabolite, but generally the number of metabolites that can be sensed is limited by the number of microelectrodes that can be assembled together. Moreover, some micro-electrode- based sensors are not appropriate for longitudinal measurements but for single or timelimited use only. Optical sensors such as the ones included in wearable devices, and typically measuring heartbeat or arterial oxygenation, also belong to the category of sensing tissue parameters, but usually do not afford biochemical information.

[0005] To enable truly non-invasive measurements, non-invasive optical or Raman spectroscopy have been considered which, however, are subject to several physical limitations, such as photon scattering, limited penetration depth, operation only in trans-illumination geometry and weak signals, contrast and sensitivity in detecting metabolites in-vivo.

[0006] It is an object of present disclosure to provide an improved apparatus and method for sensing of tissue features, such as morphological features and / or biomolecules, in particular metabolites, in biological tissue and an according method for signal processing, wherein, in particular, the accuracy and reliability of tissue feature sensing is improved.

[0007] The object is achieved by an apparatus and methods according to the independent claims and a computer program product comprising instructions causing a computer, computer system and / or distributed computing environment and / or a processor of the apparatus to execute the steps of the methods.

[0008] According to a first aspect, an apparatus for optoacoustic sensing of tissue features in biological tissue comprises: at least one irradiation device configured to irradiate the tissue with a first electromagnetic radiation exhibiting one or more wavelengths or wavelength ranges in a first spectral range; at least one detection device configured to detect at least one first optoacoustic signal corresponding to ultrasound waves emanating from a region of interest (ROI) in the tissue in response to irradiating the tissue with the first electromagnetic radiation; and a processor configured to determine, based on the first optoacoustic signal, at least one first parameter characterizing at least one first tissue feature of interest in the region of interest (ROI).

[0009] According to a second aspect, a method for optoacoustic sensing of tissue features in biological tissue comprises: irradiating the tissue with a first electromagnetic radiation exhibiting one or more wavelengths or wavelength ranges in a first spectral range; detecting at least one first optoacoustic signal corresponding to ultrasound waves emanating from a region of interest (ROI) in the tissue in response to irradiating the tissue with the first electromagnetic radiation; and determining, based on the first optoacoustic signal, at least one first parameter characterizing at least one first tissue feature of interest in the region of interest (ROI).

[0010] According to a third aspect, a method for processing optoacoustic signals comprises: providing and / or receiving at least one first optoacoustic signal which has been obtained by detecting ultrasound waves emanating from a region of interest (ROI) in a biological tissue in response to irradiating the tissue with a first electromagnetic radiation exhibiting one or more wavelengths or wavelength ranges in a first spectral range; and determining, based on the first optoacoustic signal, at least one first parameter characterizing at least one first tissue feature of interest in the region of interest (ROI).

[0011] According to a fourth aspect, an apparatus for optoacoustic sensing of biomolecules, in particular metabolites, in biological tissue, comprises: at least one irradiation device configured to irradiate the tissue with a first electromagnetic radiation exhibiting one or more wavelengths or wavelength ranges in the mid-infrared spectral range; at least one detection device configured to detect at least one first optoacoustic signal corresponding to ultrasound waves emanating from a region of interest in the tissue in response to irradiating the tissue with the first electromagnetic radiation; and a processor configured to determine, based on the first optoacoustic signal, at least one first parameter representing a measure of a density of at least one biomolecule of interest in the region of interest.

[0012] According to a fifth aspect, a method for optoacoustic sensing of biomolecules, in particular metabolites, in biological tissue, comprises: irradiating the tissue with a first electromagnetic radiation exhibiting one or more wavelengths or wavelength ranges in the mid-infra- red spectral range; detecting at least one first optoacoustic signal corresponding to ultrasound waves emanating from a region of interest in the tissue in response to irradiating the tissue with the first electromagnetic radiation; and determining, based on the first optoacoustic signal, at least one first parameter representing a measure of a density of at least one biomolecule of interest in the region of interest.

[0013] According to sixth aspect, a method for processing optoacoustic signals, comprises: providing and / or receiving at least one first optoacoustic signal which has been obtained by detecting ultrasound waves emanating from a region of interest in a biological tissue in response to irradiating the tissue with a first electromagnetic radiation exhibiting one or more wavelengths or wavelength ranges in the mid-infrared spectral range; and determining, based on the first optoacoustic signal, at least one first parameter representing a measure of a density of at least one biomolecule of interest in the region of interest.

[0014] According to a seventh aspect, a computer program product comprising instructions causing a computer, computer system and / or distributed computing environment to execute the steps of the method according to the second and / or third and / or fifth and / or sixth aspect, and / or the processor of the apparatus according to the first and / or fourth aspect to execute the steps of the method according to the second and / or third and / or fifth and / or sixth aspect.

[0015] Aspects of present disclosure are preferably based on the approach of determining at least one first tissue feature of interest or at least one first parameter characterizing at least one first tissue feature of interest representing a measure of at least one tissue characteristic, for example an anatomical feature such as size distribution of blood capillaries, a density of microvasculature, thickness of the epithelial layer or a skin structure like a tumor, a density of at least one tissue feature of interest, in particular a metabolite such as glucose, or any other tissue feature in a region of interest in a biological tissue based on at least one first optoacoustic signal, which is and / or has been obtained by detecting ultrasound waves emanating from a region of interest in a biological tissue in response to irradiating the tissue with a first electromagnetic radiation exhibiting one or more wavelengths or wavelength ranges in a first spectral range. This way of sensing tissue features, which is also referred to as “optoacoustic sensing”, solves the physical limitations of any other optical sensor since it addresses photon scattering by illuminating the tissue with light pulses generating ultrasound (and thermal) waves with intensities that depend on the extinction coefficient and concentration of tissue features. Preferably, as will be explained in more detail below, different tissue features can (but do not necessarily have to) be targeted by tuning the illumination to their peak absorption wavelength. Advantageously, ultrasound is not sensitive to photon scatter and, therefore, optoacoustic sensing can (but does not necessarily have to) sample specific volumes or layers in tissue, preferably by so-called time gating, as will be discussed further below, further improving accuracy and sensitivity over optical methods.

[0016] Within the context of present disclosure, the terms “irradiation”, “illumination” and “excitation” and their variants such as “irradiate” and “illuminate” are used synonymously. Further, the terms “detection” and “collection” and their variants such as “detect” and “collect” are used synonymously. Further, the terms “electromagnetic radiation” and “light” are used synonymously. Different forms of electromagnetic radiation can be used, from any part of the electromagnetic spectrum, to excite ultrasound waves, based on the interaction (e.g. absorption) of the particular frequency or frequency range selected by tissue structures or molecules.

[0017] Preferably, the term “tissue feature(s)” herein describes any property that relates to, e.g., a mass and / or density of a parameter (e.g. blood vessels), a number of molecules, a concentration (e.g. a number, mass or volume of the tissue feature(s) relative to a unit volume of the region of interest, e.g. mg / dl or mmol / l), or any other descriptor of a property present in the interrogated volume (region of interest).

[0018] Preferably, a tissue feature can be expressed in relative terms to a standardized measurement, for example a measurement from a phantom with known amount of the feature, in arbitrary units, in percentage, or in absolute units of mass, number of molecules or concentration. The term “standardized measurements” also implies measurements that are used as standards or as reference measurements. Such references can also be collected from the tissue measured by the apparatus disclosed herein, or another tissue employed to provide this reference measurements.

[0019] Accordingly, the term “tissue feature(s)” preferably implies any information collected by an optoacoustic sensor, and may include aspects of tissue and skin architecture, i.e. morphological information, functional and physiological information or molecular information, i.e. the concentration of certain tissue features in certain volumes of tissue.

[0020] Alternatively or additionally, aspects of present disclosure are preferably based on the approach of determining at least one first parameter representing a measure of a density or concentration of at least one tissue feature, for example a biomolecule of interest, in particular a metabolite such as glucose, in a region of interest in a biological tissue based on at least one first optoacoustic signal, which is and / or has been obtained by detecting ultrasound waves emanating from a region of interest in a biological tissue in response to irradiating the tissue with a first electromagnetic radiation exhibiting one or more wavelengths or wavelength ranges in the mid-infrared spectral range. This way of sensing biomolecules, which is also referred to as “optoacoustic sensing”, solves the physical limitations of any other optical sensor as follows: First, it addresses photon scattering by illuminating the tissue with light pulses generating ultrasound (and thermal) waves with intensities that depend on the extinction coefficient and concentration of biomolecules. Different biomolecules can be targeted by tuning the illumination to one or more preferred wavelengths, for example one or more peak absorption wavelengths. Advantageously, ultrasound is not sensitive to photon scatter and, therefore, optoacoustic sensing can sample specific volumes or layers in tissue, preferably by a so-called time gating as further explained below, improving accuracy and sensitivity over optical methods. Second, compared to conventional mid-1 R methods that are limited due to the penetration depth and operation only in trans-illumination geometries, optoacoustic sensing can reach deeper into the tissue and operate in reflectance mode since ultrasound propagates isotopically and the read path is attenuated many orders of magnitude less than mid-1 R photons in tissue. Third, compared to Raman methods, including SRS or CARS, optoacoustic sensing using mid-1 R irradiation offers up to three orders of magnitude stronger signal and contrast to make it significantly more sensitive and specific in detecting metabolites in-vivo.

[0021] Preferably, the term “density” herein describes a property that relates to a mass, a number of molecules, a concentration (e.g. number, mass or volume of the biomolecule(s) relative to a unit volume of the region of interest, e.g. mg / dl or mmol / l) or any other descriptor of the amount of biomolecule(s) present in the interrogated volume (region of interest), and can be expressed in relative terms to a standardized measurement, for example a measurement from a phantom, or other reference medium including tissues, with known amount of a molecule, in arbitrary units, in percentage, or in absolute units of mass, number of molecules or concentration. The term “density” can also relate to measuring a tissue feature that relates to a tissue structure, i.e. an anatomical feature, for example blood vessels. An example molecule detected can be a form of oxygenated or deoxygenated hemoglobin, ora combination thereof. However, since hemoglobin is concentrated within blood vessels, it then allows the characterization of “density” of vasculature and microvasculature, not only of the biomolecule concentration. Many other molecules can serve in a similar role, for example melanin to delineate aspects of the skin or externally administered molecules, such as contrast agents that, by interacting with the radiated electromagnetic energy, produce ultrasound waves.

[0022] Preferably, the first spectral range may imply or be any spectral range in the UV, visible, near infrared (NIR), short wavelength infrared (SWIR) or mid-infrared (mid-1 R) spectral range. Further, the first spectral range can be a series of wavelengths of finite spectral width emitted in parallel or in series, for example in a time-shared fashion, i.e. in an interleaved fashion or one after the other.

[0023] Preferably, within the context of present disclosure, the mid-infrared (mid-1 R) spectral range preferably covers wavelengths from approx. 2.5 pm to approx. 1 mm, preferably wavelengths from 3 pm to 16 pm, in particular from 3 pm to 12 pm or from 3 pm to 8 pm.

[0024] Preferably, the irradiation device is configured to irradiate the tissue with a second electromagnetic radiation exhibiting one or more wavelengths or wavelength ranges in a second spectral range, the at least one detection device is configured to detect at least one second optoacoustic signal corresponding to ultrasound waves emanating from the region of interest (ROI) in the tissue in response to irradiating the tissue with the second electromagnetic radiation, and the processor is configured to determine, based on the second optoacoustic signal, at least one second parameter characterizing at least one second tissue feature of interest in the region of interest, and to determine at least one third parameter based on the at least one first parameter and the at least one second parameter, in particular by normalizing and / or correcting the at least one first parameter by the at least one second parameter.

[0025] Preferably, the first spectral range is in the mid-infrared range and / or the second spectral range is outside the first spectral range, in particular in the visible spectral range, and / or the at least one first parameter represents a measure of a density of at least one biomolecule of interest, in particular at least one metabolite, in the region of interest (ROI) and / or the at least one second parameter represents a measure of at least one of i) a blood density characterizing an amount of hemoglobin molecules present in the tissue and / or ii) a vascular density characterizing an amount of blood vessels containing hemoglobin molecules and / or iii) a lipid density characterizing an amount of lipid molecules present in the tissue and / or the presence of adipose tissue in the tissue.

[0026] Preferably, the irradiation device is configured to irradiate the tissue with a second electromagnetic radiation exhibiting one or more wavelengths or wavelength ranges outside the mid-infrared spectral range, in particular in the visible and / or near-infrared spectral range, the at least one detection device is configured to detect at least one second optoacoustic signal corresponding to ultrasound waves emanating from the region of interest in the tissue in response to irradiating the tissue with the second electromagnetic radiation, and the processor is configured to determine, based on the second optoacoustic signal, at least one second parameter representing a measure of i) a density of an optical absorber, for example a chromophore or light absorbing nanoparticle, in particular a blood density in the tissue, due to an amount of hemoglobin molecules present in the volume, or of ii) a vascular density, typically due to an amount of blood (micro-) vessels that contain said hemoglobin molecules, whereby said volume is preferably the same volume that contains the at least one biomolecule of interest, and to determine at least one third parameter based on the at least one first parameter and the at least one second parameter, in particular by normalizing the at least one first parameter by the at least one second parameter. Depending on the wavelength(s) of the second electromagnetic radiation, the normalization can, alternatively or additionally, be performed based on molecules that are different than hemoglobin, i.e. other tissue molecules that absorb light in the visible or near-infrared part of the spectrum, for example lipid molecules or adipose tissue at 930nm. Accordingly, alternatively or additionally to i) blood density and / or ii) vascular density, the at least one second parameter can represent iii) a lipid density as a measure characterizing an amount and / or the presence of lipid molecules and / or adipose tissue in said volume.

[0027] Preferably, the irradiation device is configured to irradiate the tissue simultaneously with the first electromagnetic radiation and second electromagnetic radiation, preferably by means of frequency-wavelength multiplexing (FWM) which will be explained in more detail below. Alternatively, the irradiation device is configured to irradiate the tissue with the first electromagnetic radiation and second electromagnetic radiation subsequently and / or in a time interleaved manner (time multiplexing). Time interleaved irradiation can occur using trains of pulses of the first and second electromagnetic radiations, whereby the pulses do not overlap but are shifted to each other. The shift can occur in any combination of mixing the two radiations, from interleaving a pulse from the first and second radiation on a per pulse level, to irradiating with the entire pulse train of the first radiation first and then irradiating with the entire pulse train of the second irradiation or any combination thereon. Each pulse train may contain a single pulse or preferably a multitude of pulses, for example 100 pulses or 1000 pulses so that better signal statistics and signal to noise ratio can be achieved, as described below. Preferably, within the context of present disclosure, the visible spectral range covers wavelengths from approx. 380 nm to 750 nm, and / or the near-infrared spectral range covers wavelengths from approx. 750 nm to approx. 990 nm.

[0028] Advantageously, the above-mentioned preferred embodiment, which is also referred to as “vascular density referencing” (VDR) herein, allows for enhancing measurement accuracy and quantification. While the first optoacoustic signal obtained in response to illuminating the tissue with mid-1 R radiation gives a measure of metabolite density in a specific layer and / or volume (typically 60x60x60 pm3), co-illuminating the tissue with a preferably visible wavelength which is sensitive to vasculature (e.g. 488 nm, 532 nm or 630nm) yields a second optoacoustic signal which is indicative of and / or a measure of a vascular density at the, preferably exactly, same layer and / or volume of the tissue. Likewise, co-illumination at a different wavelength, e.g. 930 nm, can yield a second optoacoustic signal that is indicative of and / or a measure of lipid and / or adipose tissue density. A particularly advantageous preferred aspect of present disclosure is the combination of such referencing schemes, such as the VDR, with the time-gating of the optoacoustic signals, so that reference signal can be separated spatially and spectrally, as an alternative to being separated only spatially or only spectrally, as explained in the following.

[0029] Preferably, the first optoacoustic signal and the second optoacoustic signal are detected and / or processed by so-called time gating using the same time gate, so that the layer and / or volume, from which the first optoacoustic signal is obtained, is exactly the same as the layer and / or volume, from which the second optoacoustic signal is obtained. Then density determination relates to a value extracted from the optoacoustic signal contained within the time-gate employed. For example the density of hemoglobin will be any quantity that relates to the mean, median, integration or other operation of the raw or processed signal, for example the signal after application of an absolute operator, a squaring operator, an envelope computing operator, the Hilbert transform or any other processing operator that reveals at least one or more quantities of the signal measured and any combination of these quantities, for the signal measurement in the visible and for a gate that corresponds to a depth that is under the epidermis-layer that contains melanin. In contrast, a gate that will take the signal from the epidermal layer that contains melanin will report on melanin density, whereas a measurement at 930 nm will report on lipid measurement. Alternati vely , as discussed in more detail below, the first optoacoustic signal and the second optoacoustic signal are detected and / or processed by time gating using different time gates, so that the layer and / or volume, from which the first optoacoustic signal is obtained, is different from the layer and / or volume, from which the second optoacoustic signal is obtained.

[0030] The concept of VDR enables referencing of mid-IR measurements of a metabolite concentration (also referred to as “first parameter”) to the amount of blood density, vascular density or lipid density (also referred to as “second parameter”) in the same volume and / or layer. Generally, the target metabolite concentration in blood is much higher (>10x) than in ISF. Therefore, considering the blood density and / or vascular density (measured / ob- tained in response to e.g. visual and / or near-infrared illumination) in the exact same layer from which the metabolite density is measured / obtained in response to mid-IR illumination allows for considerably improving the accuracy of metabolite concentration, preferably by normalizing the first parameter (concentration) obtained in the mid-IR with an actual value of vascular density (second parameter) at the point of measurement, rather than by an assumed uniform vascular density value. Likewise, use of lipid density for referencing and / or normalization can yield measures associated with metabolite concentrations in adipose tissue, enabling applications in precision nutrition, food metabolism and progression monitoring of various conditions including cardiometabolic disease or drug and other treatment efficacy. We further note, that while a normalization operation may increase the accuracy of computing metabolite concentration, measurements in the mid-infrared, visible and / or near-infrared can be also used as measurements (biomarkers) in determining and reporting the presence of metabolites and biomolecules in tissue (i.e. glucose, lipids, hemoglobin) independently of each-other.

[0031] Normalization herein preferably infers classical and / or Artificial Intelligence (Al) operations by which the at least one mid-l R signal (first optoacoustic signal) relating to a concentration of a target metabolite is corrected with the second optoacoustic signal obtained by measuring outside the mid-IR spectral region to account for dilution parameters of the target metabolite detected in the corresponding volume it is detected at. While the dilution volume could be in-principle also detected in the mid-IR, since blood / hemoglobin absorbs in the mid-IR, accurate determination of this parameter is advantageously performed outside the mid-IR spectrum to advantageously minimize contributions from water and other molecules. Likewise, the signal employed for normalization could be collected from the first or second optoacoustic signals at a time gate that is different to the time-gate employed for computing the tissue feature that is being normalized. This means that normalization may not be due to spectral separation but spatial separation, i.e. separation as a function of depth.

[0032] Classical normalization may preferably refer to a division operation. Such an operation may preferably calculate at least one value (referred to as “third parameter” herein) from the first optoacoustic signal (or the first parameter derived therefrom) and the second optoacoustic signal (or the second parameter derived therefrom) measured within a certain time gate in the mid-IR and the visible, respectively, and then divides the mid-IR value (first optoacoustic signal or first parameter) with the visible value (second optoacoustic signal or second parameter, respectively) to obtain the at least one third parameter. The at least one value per spectral range (i.e. the first optoacoustic signal or first parameter and the second optoacoustic signal or second parameter) can be a mean value, a maximum value, a maximum minus a minimum value, and / or an integration of the first or second, respectively, optoacoustic signal over at least a part of the time-gate selected. Preferably, the respective value per spectral range can be any of the raw signal (i.e. the first or second, respectively, optoacoustic signal) or processed forms (i.e. the first or second, respectively, parameter) of the signal. Processed forms of the signal preferably include the absolute value of the signal, the signal at some power, the Hilbert transform of the signal, the envelope of the signal, any differential of the signal, the Fourier T ransform and / or power spectrum of the signal etc. or any combination of these processed and raw signals.

[0033] Preferably, a normalization based on deep-learning and / or machine learning and / or Al operations comprises training of models (typically multi-layer neural networks) to predict the correct value (predicted third parameter) of the concentration of a metabolite based on the signals obtained in the mid-IR (first optoacoustic signal or first parameter) and visible (second optoacoustic signal or second parameter) and the correct value (third parameter) as it is added in a sample measured, computed or measured from the sample measured. The input may be the first and / or second optoacoustic signals at multiple wavelengths and / or different time-gates. Other normalization methods, as described in the literature, can also be applied, for example using look-up tables or functions which map any combination measurements in the mid-IR, visible and / or near-infrared to a new value that yields a more accurate or a preferred representation of the quantity of interest. Such normalizations, as described in this paragraph, present a generalization of the classical division operation and may allow normalization with more than a single measurement, for example normalization of one measurement in the mid-IR with a multitude of other measurements obtained in the mid-IR, visible and / or near-infrared, or any other spectral range of the electromagnetic radiation employed (for example short-wavelength infrared SWIR - 990nm - 2.5 micrometers).

[0034] In a preferred embodiment, VDR further implements the concept of overdriving laser diodes and / or Frequency-Wavelength Multiplexing (FWM), which is explained in further detail below, for realization of cost-effective sensors and / or true simultaneous illumination with mid-IR and visible lines respectively.

[0035] Employing over-driven laser diodes, preferably as described by Stylogiannis et al., Photoacoustics 9:31 ; 2018, allows to generate up to 27-times stronger optoacoustic signal generation over conventional driving modes (see Stylogiannis et al., Sensors 21 :1379; 2021) based on inexpensive <3 mm laser diode modules. A preferred embodiment of present disclosure employs a laser diode in the visible for VDR.

[0036] Preferably, high photonic integration can be achieved by utilizing non-packaged laser diodes, a practice that besides integration also reduces capacitance and inductance loads offering the ability to drive the laser diodes with higher energies and / or shorter pulses. Few nanosecond-width pulsing or even narrower can be achieved this way while retaining high currents >50 Amperes. The laser diodes in this case can be combined with mid-IR lines from QCL-lasers on chip.

[0037] Preferably, the at least one irradiation device and the at least one detection device are provided in the same half-space above and / or adjacent to the irradiated tissue. Advantageously, with this embodiment, which is also referred to as “reflectance mode”, an increased penetration depth can be achieved compared to, in particular, conventional mid- IR spectroscopy measurements, since the use of ultrasound for detection of optical contrast enables penetration twice as deep as conventional mid-IR detection. Despite the strong water absorption in the mid-IR, it was surprisingly found that optoacoustic sensing penetrates deep enough (-170 pm) and can reach deep enough to measure biomolecules at the epidermal-dermal junction, a skin layer rich in micro-vasculature (capillary loops and the superficial vascular plexus). Moreover, it allows detection at the same side of tissue as illumination, i.e. in reflectance mode, since ultrasound generation is isotropic. Preferably, the processor is configured to perform a time-gating operation so as to process only a part of the first and / or second optoacoustic signal, which is obtained only from a certain volume or layer in the tissue. This embodiment allows for a so-called skin rejection and in-blood measurement (SRBM). In addition to enhanced depth and reflectance operation as described above, optoacoustic sensing technology can enable in-blood metabolite detection, i.e. it solves the key limitation of other sensor technologies, which measure glucose “in bulk” from interstitial fluids. In particular, by using time-gating of the mid-IR optoacoustic signal, optoacoustic detection offers two particular advantages: rejecting signal contamination from the skin surface by focusing on later arriving signals from deeper skin layers, and focusing the measurement only to layers rich in microvasculature, i.e. preferably at the epidermal-dermal junction, ~80 to 100 pm deep, thus sensing metabolites in blood capillary-rich volumes as opposed to superficial interstitial fluid achieved by conventional transdermal (and mid-IR) sensors. This embodiment is particularly advantageous when implemented in a portable and miniaturized format of increased manufacturability and multiple wavelength operation to enable unprecedented sensitivity preferably through advanced referencing techniques, in particular VDR, as described above.

[0038] Preferably, the first electromagnetic radiation exhibits a first wavelength, which is coincident with a peak wavelength at which the biomolecule of interest exhibits a (local) maximum absorption or stronger absorption than in other wavelengths in the mid-infrared spectral range, and a second wavelength, which is offset from the first wavelength and at which the biomolecule of interest exhibits an absorption which is less than the (local) maximum absorption and / or which is in the range of or less than a background absorption caused by one or more molecules, in particular water, other than the biomolecule of interest. Preferably, the term “first wavelength” also implies an illumination containing a particular spectral shape, for example a Gaussian or Lorentzian distribution around a wavelength or several wavelength peaks that effectively target (local) absorption maxima of the target biomolecule. Preferably, the at least one detection device is configured to detect a first optoacoustic signal of the first wavelength, the first optoacoustic signal of the first wavelength corresponding to ultrasound waves emanating from the region of interest in response to irradiating the tissue with the first electromagnetic radiation exhibiting the first wavelength, and to detect a first optoacoustic signal of the second wavelength, the first optoacoustic signal of the second wavelength corresponding to ultrasound waves emanating from the region of interest in response to irradiating the tissue with the first electromagnetic radiation exhibiting the second wavelength. Preferably, the term “second wavelength” also implies an illumination containing a particular spectral shape, for example a Gaussian or Lorentzian distribution around a wavelength or several wavelength peaks that effectively target (local) non-absorption maxima of the target biomolecule. Preferably, the processor is preferably configured to determine the at least one first parameter based on the first optoacoustic signal of the first wavelength and the first optoacoustic signal of the second wavelength. Further preferably, the processor is configured to determine the at least one first parameter based on a background-corrected optoacoustic signal which is obtained by subtracting the first optoacoustic signal of the second wavelength from the first optoacoustic signal of the first wavelength.

[0039] These embodiments, which are also referred to as “differential referencing” (DRC) herein, allow for increasing the specificity of detection, in particular compared to conventional mid- IR sensing of which a major limitation is the presence of background signals, including the strong absorption by water. Therefore, a measurement at any single peak does not reveal information of a single metabolite but a multitude of contributions from water and other biomolecules. To address this issue, the preferred embodiments include using wavelength pairs, for example Quantum Cascade Laser (QCL) lines, in order to offer a measurement at an absorption peak or wavelength of strong absorption of a target metabolite and a slightly, preferably minimally, shifted illumination to capture an off-peak signal, that contains background contributions, serving as a reference to subtract said background contributions.

[0040] Preferably, built on recent progress with QCL manufacturing (see, for example, Jung et al., Nature Commun 5:4267, 2014; Jiang et al. AIP Advances 8:085021 , 2018), DRC can utilize and / or produce distributed feedback (DFB) QCLs to target glucose, lipids and other metabolites. One or multiple pairs can be used per metabolite with QCL pairs targeting specific on and off metabolite peaks (i.e. first and second wavelength, respectively). As disclosed in the aforementioned publications, the QCLs can feature two rapidly electrically switchable emission wavelengths (at -100 nanosecond switching time). This enables background-free, or at least background- reduced, optoacoustic sensing that can be seamlessly integrated in FWM modulation due to the fast switching. The two-colour DFB-QCL concept and the tuning method was demonstrated in the aforementioned paper at room temperature. The wavelength pairs can therefore be produced by wavelength switching from a single laser and / or using one or two QCL pairs and / or technology that combines below-threshold continuous-wave biasing and above-threshold pulsed biasing to produce micro-spectra around the peak. While differentiation using one wavelength pair could improve detection over using a single wavelength, using two or more QCL pairs for two or more different peaks of the same metabolite can further increase the specificity of detection over targeting a single peak. This can be achieved by processing the two or more differences obtained from each wavelength pair. Processing in this case could be any statistical measure of extracting one value of higher accuracy from multiple measurements, including summing all the differences, averaging them, taking the mean or median and so on.

[0041] Multiple wavelengths, as applied in one or more wavelength pairs, can be also modulated by a FWM (frequency-wavelength multiplexing) scheme for concurrent illumination as described in Stylogiannis A., et.al. Nature Comm 13:4448 2022. In this case, pulse trains of slightly different periodicity are employed to carry the different wavelengths. The Fourier Transform of these pulse trains reveals multiple discrete modulation frequencies that are slightly different per wavelength carried. Then a demodulator can be employed to separate the frequencies and retrieve the optoacoustic signal strengths as a function of depth for each wavelength. FWM leads to SNR improvement over time-sharing methods that emit pulses at different time points to separate optoacoustic signals in time.

[0042] Preferably, the wavelength pair approach can be generalized by producing spectral microscans using bias-tree methodology as described in the literature (see, e.g., Nair, Anil & Minesi, Nicolas & Kuenning, Nicholas & Keller, Alex & Spearrin, Raymond (2023) Applied Physics B 129) to combine continuous and pulsed current to drive the QCLs. The biastree approach uses the CW current to scan the laser wavelength within 2-3 cm’1range and produce micro-spectra around an absorption peak of a metabolite while continuously pulsing the laser (above threshold), with or without FWM modulation, to generate maximum optoacoustic signal strength. Preferably, the micro-spectra can be processed with spectral analysis approaches (deep learning, spectral unmixing) to extract the presence of the spectrum of the molecule of interest, in the spectral range scanned, over the background absorption spectrum.

[0043] A particular embodiment using micro-scanning techniques, utilizes micro-scanning of any light source, not necessarily pulsed light sources. By means of fast scanning, ideally within less than a micro-second, or preferentially less than 100 ns, not only pulsed but also continuous wave sources can be employed for the direct generation of a DRC signal. This concept is based on the change of absorption within the scanned region, as shown in Fig. 4 for the S1 molecule, as a function of scan time. This approach can lead to optoacoustic signal generation based on lower cost and higher energy light sources, as opposed to pulsing light sources. For example, a fast can would scan two or more wavelengths, in a discrete or continuous fashion within the range Xi - X2. The scan could be mono-directional (i.e. from Xi to X2), bi-directional or continuously alternating between directions or discrete wavelengths. Advantageously, for sufficient fast scanning in the > 1 kHz range, the optoacoustic signal generated depends on the absorption change within the scan range, not the absolute absorption. As mentioned above, this absorption change is the target of the DRC approach. This is the reason that implementation of the method with a fast scanning approach yields an optoacoustic signal directly relating to the absorption change of interest, i.e. an absorption change that is mostly or entirely attributed to the spectrum of the biomolecule of interest. The scanning characteristics can vary in this scheme, from continuous scanning over a selected wavelength (wavenumber) range, to two or more discrete wavelengths placed within this wavelength range. Such scan could be implemented with the bias-tree techniques discussed in the previous paragraphs or with fast switching between two or more wavelengths, using a separate or common light source, for example two or more QCL’s, each one triggered by a pulse train shifted in time from the others. In a preferred embodiment, this time shift is shorter than the thermal relaxation of the molecules of interest. In this latter case, the pulse trigger scheme causes a modulation of the optoacoustic signals between the absorption states sampled by the two or more wavelengths. Such signal relates only to the low and high absorption states of the biomolecule of interest, which when assuming that the background (S2 in Fig. 4) does not vary significantly, is the direct difference signal. In this case subtraction or spectral unmixing may be avoided.

[0044] In all DRC implementations, the selection of wavelength range scanned or wavelengths and / or wavelength pairs employed depends preferably not only on the maximum (peak) of the metabolite of interest, but also on the relative change of absorption of the metabolite of interest within the scan range as well as the relative change of absorption of all other background absorbers. Preferably, the range of absorption is set at a spectral window whereby the molecule / metabolite of interest demonstrates a high absorption peak and a high absorption change (steep drop of absorption around the high absorption peak - see Fig. 4, S1), whereby the background absorbers (Fig. 4, S2) offer low and / or insignificant change. Optimization of these three parameters (peak, steepness and background change) selects the optimal scan window, i.e. the wavelength range for scanning purposes or for placing the at least one pair of peaks.

[0045] Finally, another preferred approach for establishing DRC borrows concepts from frequency comb methods as they are applied to dual-comb spectroscopy (Coddington I., et. al., Optica 3 414 2016) but advantageously adapted herein for optoacoustic DRC. Illumination with frequency combs using pulse trains can give rise to optoacoustic signals and bears the advantage of generating multiple discrete light frequencies (colors) that can be employed to sample a broad spectrum containing multiple peaks of a target biomolecule. While demodulation of single frequency comb is a time consuming process using a Fourier Transform Spectrometer (FTS) (see Lloyd LB., et. al. Rev .Sci. Instrum. 51 1488 1980) or other elaborate demodulation techniques that require frequency scanning, dual frequency comb spectroscopy allows for an implicit demodulation using detection at a beat frequency generated for each comb tooth pair using the two frequency combs at slightly different color spacing (Friedlein JT., et. al. Nature Comm.11 :3152 2020). Although this approach typically requires normalization with the incident light for quantitative measurements, it is useful to provide multiple lines that can form DRC pairs around multiple peak lines of a metabolite of interest. Nevertheless, frequency comb may provide up to tens of thousands of lines of evenly spaced narrow color lines, that may not be necessary for common metabolite sensing or implementation in portable inexpensive sensors due to the complexity of the optics and of the data processing required. Moreover, frequency comb implementation as provided for dual-comb spectroscopy may yield sub-optimal energy in the wavelengths of interest to DRC, that could require long acquisition times due to the low signal to noise ration achieved for each of the wavelengths available. Inventively herein, the DRC can be implemented using the one-, or a few pair approach described above using the same demodulation technique, assuming coherent light sources in par with the characteristics used by frequency comb sources. In this case, similar wavelengths simultaneously illuminating the sample, will give rise to an optoacoustic signal resonating at the beat frequency of the two wavelengths, which can be tuned to the bandwidth of any ultrasound detector.

[0046] Preferably, the at least one detection device is configured to detect the first optoacoustic signal and / or second optoacoustic signal many times within a period of time to obtain a series of first optoacoustic signals and / or second optoacoustic signals, respectively, and to store the series of first optoacoustic signals and / or second optoacoustic signals in a data memory. Preferably, the processor is configured to apply a moving average and / or median filtering operation on the series of first optoacoustic signals and / or second optoacoustic signals to obtain an averaged first optoacoustic signal and / or an averaged second optoacoustic signal, respectively, and to determine the at least one first parameter based on the averaged first optoacoustic signal and / or to determine the at least one second parameter based on the averaged second optoacoustic signal.

[0047] This approach, which is also referred to as “single-point noise analytics” (SPNA) herein, advantageously allows for increasing sensitivity and data quality control. Mid-IR optoacoustic detection operates on strong absorption contrast compared to other optical methods such as Raman (CARS, SRS). Nevertheless, the measurements may contain noise due to shot noise, laser fluctuations, sensor contact / movement and possibly ambient interference. Increasing the sensitivity implies identifying signal that may be hidden in the noise. For this reason, measurements can be averaged over a period of time, e.g. averaging 10,000 measurements 1 MHz (~10ms). For shot noise limited systems, averaging improves the SNR by the square root of the number of averages, i.e. 100x in the given example. However, averaging does not describe the actual noise characteristics of non-shot- limited systems. Therefore, according to present approach, all individual measurements (first and / or second optoacoustic signals detected over a period of time) are collected and stored for post-processing, and a moving average is used to analyze the noise statistics to exact accurate SNR values, preferably filter out at least 10% of outliner values (preferably by median filtering operations) to improve SNR and preferably also issue warnings for very low SNR’s, as a quality control operation.

[0048] Preferably, all individual measurements collected are stored for post processing, and a moving average operator in time is used to process a number of points N in time as follows: Assume a sequence of first and / or second optoacoustic signals detected at points (in time) of total number T, said time-points collected over a period of time. For example, 1 second or 10 seconds acquisition at 1 MHz yields a sequence of signals at in total T=106or 107points in time, respectively. Preferably, N is smaller than or equal to T, e.g. N=10,000 points. For the first point (in time) of the T sequence (which can be defined anywhere in the sequence), 10,000 points are collected, typically adjacent points, which contain that first point. For the sequence of N points, statistical analysis is executed, i.e. statistical and signal processing metrics such as the mean, the standard deviation, the median, the root mean square, the signal to noise ratio, the frequency content etc). These metrics are stored, together with the value of the first point. Then the process repeats for a second point in the T sequence, and so on. Typically, for each point Ta analyzed, N is an odd number and the sequence collects N / 2-0.5 points earlier than Ta and N / 2-0.5 points later than Ta. This process can be performed in real-time or post-processing.

[0049] Advantageously, this process provides metrics of the quality of the signal measured (by means of SNR and other statistical measurements) for quality control purposes. It can identity parts of the sequence T that were prone to noise or movement and overall extract the SNR and quality of the entire T sequence, when N=T or for each point of the sequence T. Another very advantageous function of this process is to identify outliers (shot noise, non-regular values due to transient interference or movement) and remove them from the sequence T. Different filter operators can be applied for this operation, a preferred one being the median filter. However other filters, including filters trained under machine learning / deep learning models (e.g. see Dehner C., et.al. IEEE TMI 41 (11) 3182 2022), can be applied.

[0050] Preferably, irradiating the tissue with the first electromagnetic radiation exhibiting one or more wavelengths or wavelength ranges in the mid-infrared spectral and / or irradiating the tissue with the second electromagnetic radiation exhibiting one or more wavelengths or wavelength ranges outside the mid-infrared spectral range, in particular in the visible spectral range, comprises a frequency-wavelength multiplexing (FWM). In conventional approaches, a multi-wavelength operation is typically achieved by time sharing of wavelengths (i.e. different wavelengths are applied successively), which prolongs the measurements by a multitude of the number of wavelengths employed. For an illumination module of e.g. five wavelengths, this means that measurements will take five times longer than a measurement at a single wavelength. This can be avoided by using a FWM scheme which provides an illumination of the target (tissue) with light pulses in the form of multiple parallel pulse trains, that act as the equivalent of a frequency comb to produce a multitude of discrete frequencies in the frequency domain. Each pulse train carries one wavelength (in the mid-IR or outside the mid-IR, preferably VIS), and it has a unique periodicity that is slightly different than all other pulse trains concurrently employed. As a result, each wavelength is carried at multiple discrete frequencies (F). These discrete frequencies can be seen by taking the Fourier Transform of the optoacoustic signal that has been collected. FWM allows for a simultaneous illumination of a target with all colors. Further, it results in a signal to noise enhancement by square root of N, offering the best known modulation scheme in optoacoustic sensing for multi-wavelength operation. Demodulation is then digital preferably by using the inverse discrete Fourier transform.

[0051] In summary, mid-1 R optoacoustic sensing of metabolites (such as glucose) in cells and tissues according to aspects of present disclosure is particularly advantageous over conventional approaches, where sensors are a) inherently sensitive to the metabolically inactive skin surface (stratum corneum), changes in skin humidity, superficial lipids and other molecules contaminating the glucose measurement, and are subject to challenges in market application, b) detect metabolic content from subcutaneous interstitial fluids, in the form of “bulk” measurements, which, similarly to biochemical and optical sensors, is not the clinically relevant property of glucose concentration in blood, leading to inaccuracy, and c) are sensitive to background signals coming from other metabolites that challenge the accuracy of the measurement and limit specificity and reproducibility.

[0052] Preferably, the at least one detection device is configured to detect the at least one first and / or second optoacoustic signal over time, and wherein the processor is configured to convert the first and / or second optoacoustic signal into a first and / or second converted optoacoustic signal (also referred to as “energy signal”), respectively, which comprises, in particular only positive, signal components characterizing an amplitude and / or intensity and / or energy of the first and / or second optoacoustic signal as a function of time, and to determine the at least one first and / or second parameter based on the first and / or second converted optoacoustic signal, respectively.

[0053] In this embodiment, the optoacoustic signal collected is preferably first converted into an energy signal. The optoacoustic signal typically contains positive and negative values indicative of the propagation direction of the optoacoustic signal collected. While processing can be done directly on this positive-negative signal, it is advantageous for representation and computational purposes to obtain the energy signal that does not contain direction and / or positive or negative signals. This can be achieved by different operations done on the optoacoustic signal collected, for example by the Hilbert transform. Alternatively, one can obtain the Fourier transform of the optoacoustic signal and operate directly in the Frequency Domain as will be explained in more detail below. Another option is to convert the Fourier Transformed signal back to the time-domain, using the inverse Fourier Transform, after the Fourier Transform representation of the signal is processed so that it loses the phase (i.e. the imaginary part) and is converted to an amplitude only signal in the Fourier Transform domain (i.e. the frequency domain). Operations in the Frequency domain preferably include taking the absolute of the signal or the squared of the signal and then applying the inverse Fourier Transform. Other transforms, including wavelet transforms, envelope computations and other such conventional signal processing schemes can be applied to the optoacoustic signal in either time- or frequency-domains.

[0054] Preferably, the processor is configured to select a portion of the first and / or second, in particular converted, optoacoustic signal, which has been detected during a specified time gate corresponding to a volume or layer in the tissue, and to determine the at least one first and / or second parameter based on the selected portion of the first and / or second optoacoustic signal, respectively.

[0055] In this embodiment, the processor is configured to perform a so-called time-gating operation to process the at least one first and / or second optoacoustic signal, which is obtained only from a certain volume in the tissue. Preferably, the volume is defined by the intersection of the distribution of the illuminating light in tissue and the acceptance characteristics of the detector element, i.e. its aperture and the spatial sensitivity within the volume that the detected optoacoustic signals (i.e. the ultrasound waves detected) propagate into. Accordingly, “time gating” then refers to selecting certain time period(s) along the detected optoacoustic signal, i.e. a certain time gate, which is detected over time, so that only a part or parts of the optoacoustic signal is or are selected. The time of travel of the optoacoustic signal detected corresponds to the depth that this optoacoustic signal was generated, a conversion that is achieved through multiplying the time with the speed of sound in the tissue. Therefore, selection of a time gate corresponds to selection of a certain volume within the tissue. Then, any computation performed on the optoacoustic signal that corresponds to that time gate selectively measures at least one tissue feature within that volume. This is a very advantageous feature of optoacoustic sensing, over optical sensing, since it can reveal properties in certain volumes within tissue, as opposed to bulk tissue measurements that are allowed by other optical sensors.

[0056] Preferably, the processor is further configured to select a first portion of the first optoacoustic signal, which has been detected during a specified first time gate, to select a second portion of the second optoacoustic signal, which has been detected during a specified second time gate, the specified first time gate and specified second time gate corresponding to different volumes or layers (GO, G1 , G2) in the tissue, to determine the at least one first parameter based on the selected first portion of the first optoacoustic signal, to determine the at least one second parameter based on the selected second portion of the second optoacoustic signal, and to determine the at least one third parameter based on the at least one first parameter and the at least one second parameter, in particular by normalizing and / or correcting the at least one first parameter by the at least one second parameter. In this embodiment, the signals, in particular energy signals, from one gate are employed to normalize or to correct signals, in particular energy signals, from a different gate. Alternatively, as already discussed above, the signals, in particular energy signals, obtained from one gate (e.g. in response to irradiating the tissue in a second spectral range) may also be employed to normalize or to correct signals, in particular energy signals, from the same gate (e.g. in response to irradiating the tissue in a first spectral range which is different from the second spectral range).

[0057] Preferably, the specified first time gate corresponds to at least one first volume or layer (G1 , G2) in the tissue and the specified second time gate corresponds to at least one second volume or layer (GO) in the tissue, wherein the first volume or layer (G1 , G2) is located in a deeper region of the tissue than the second volume or layer (GO), which is in particular located at a superficial region of the tissue.

[0058] Preferably, the specified first time gate corresponds to an epidermal-dermal junction layer in the tissue or a dermal layer containing blood micro vessels, and the specified second time gate corresponds to an epithelial skin layer of the tissue.

[0059] Preferably, the processor is configured to determine and / or specify the time gate, in particular the first time gate and / or the second time gate, based on a pre-definable position and / or dimension, in particular diameter and / or thickness, of a volume or layer of interest in the tissue.

[0060] Preferably, a tissue feature from the most superficial gate (second time gate) that contains optoacoustic signal is used to normalize or correct the optoacoustic signal from a gate (first time gate) that defines a volume deeper in the tissue. This correction can be applied to either a measurement at a single wavelength or multiple wavelengths.

[0061] Preferably, the processor is configured to determine, based on the selected first and / or second portion of the first and / or second optoacoustic signal, at least one of the following first and / or second parameters: a total energy of the selected first and / or second portion of the first and / or second optoacoustic signal, and / or a maximum value and / or minimum value and / or mean value and / or standard deviation value of the selected first and / or second portion of the first and / or second optoacoustic signal, and / or a frequency content in the selected first and / or second portion of the first and / or second optoacoustic signal characterizing a size distribution of components, e.g. blood vessels, of the first and / or second volume or layer in the tissue, and / or one or more derivative values of the selected first and / or second portion of the first and / or second optoacoustic signal indicative of spatial changes in the first and / or second volume or layer in the tissue, and / or one or more spectral features of the selected first and / or second portion of the first and / or second optoacoustic signal, in particular including principle component spectra and / or relative spectra corresponding to different concentrations of molecules derived through spectral unmixing techniques. Preferably, the above first and / or second parameters are determined for one or more wavelengths or wavelength ranges in the first and / or second spectral range and / or optionally include a time characteristic of the respective parameters.

[0062] Preferably, the processor is configured to generate, from at least two of the first and / or second parameters, respectively, for the respectively selected first and / or second portion of the first and / or second optoacoustic signal: a one-dimensional vector, or a multi-dimensional matrix if the first and / or second parameters were determined for two or more wavelengths or wavelength ranges and / or if the first and / or second parameters include a time characteristics of the respective parameters.

[0063] Preferably, the processor is configured to determine the at least one third parameter from the respectively generated one-dimensional vectors or multidimensional matrices by means of at least one vector or matrix operation, respectively.

[0064] Correcting the signals according to one or more of the above-described preferred aspects accounts for the fact that the excitation light has certain strength that should be known or calibrated for providing quantitative measurements. This strength of light is further modified by the characteristics of the tissue surface, i.e. tissue feature contained in the tissue surface. Therefore, the light that illuminates a volume (first volume or layer) deeper in tissue depends on the initial light intensity and its modification through at least one tissue feature on the tissue surface (second volume or layer), for example the skin epithelial layer. A change of the at least one tissue feature at the epithelial layer will modify the light illuminating a deeper layer and, therefore, the optoacoustic signal reported in the time-gate of a deeper layer contains information on tissue features in the more superficial layer. Therefore, an improved, in particular more precise, quantification of a tissue feature in the deeper layer requires a correction of this intensity modification from the superficial layer.

[0065] Further, correcting the signals may also account for a spectral illumination wherein the illuminating light has a certain spectral composition and / or distribution, i.e. a certain pattern of intensities per wavelength. This spectral pattern will be further modified by tissue features on the tissue surface, based on the specific tissue feature contained in the superficial volume (second volume). Therefore, the spectral pattern that illuminates a volume (first volume) deeper in tissue may not be flat but have a pattern that depends on the illumination source and the way this illumination is modified by the superficial volume.

[0066] It is, therefore, a particularly advantageous feature of present disclosure to select a time gate (second time gate) that corresponds to a superficial volume and record the light intensity in all wavelengths employed for illumination. This measurement yields a plurality of (reference) intensity values at the employed wavelengths which is also referred to as “reference spectrum” in the following, even if illumination is performed at a single wavelength. In the latter case, the reference spectrum refers to and / or comprises a single intensity value.

[0067] To calculate the reference spectrum, the surface of the tissue is preferably identified first. This can be performed, e.g., by recording the first instance and / or appearance of a significant optoacoustic signal, i.e. optoacoustic signal above noise, by simply computing the distance where the skin is expected to be found based on the geometrical characteristics of the sensor, by training models using deep learning, machine learning or artificial Intelligence algorithms, an independent measurement of the skin (for example an image or a distance sensor), etc.

[0068] Then, a time gate (second time gate) is defined that preferably corresponds to a surface volume. This time gate can be defined based on the dimension (thickness) of the volume desired and can be set by different methods, for example as duration representing or corresponding to the total duration of the optoacoustic signal, or as an actual dimension based on the known geometrical characteristics of the tissue measured. For example, for skin, illuminated in the visible spectral range, this time gate can be set to 80 micrometers (pm), corresponding to superficial epithelial skin layers. It can be set to 20 micrometers (pm) corresponding to the stratum corneum only.

[0069] The time gate can be set also automatically, preferably by utilizing information from the optoacoustic signal itself or by other measurements such as ultrasound sensor measurements, as discussed in the following, or images, for example images obtained by an optoacoustic or an ultrasound imaging device that is operated independently or in tandem with the sensor.

[0070] Preferably, the optoacoustic signal defined in this time gate can then be analyzed to produce the reference spectrum. The reference spectrum can be a one-dimensional vector, where each cell of the vector contains one measurement computed from the optoacoustic signal in the time gate; whereby each cell corresponds to one wavelength employed for excitation. Alternatively, the reference spectrum can also be a multidimensional vector where instead of one cell per wavelength it contains one vector, or one matrix of two or more dimensions.

[0071] Preferably, one parameter and / or property can be computed per wavelength for the onedimensional reference spectrum. This property can be the energy of the optoacoustic signal, calculated as the integral of the optoacoustic signal or the energy of the optoacoustic signal as a function of time within this time gate.

[0072] Alternatively or additionally, a number of multiple parameters and / or properties can be computed for every gate, corresponding to different tissue features. For example, for every time gate defined, including the time gate (second time gate) corresponding to the surface volume or layer (second volume or layer), one can extract a multitude of parameters including total energy, maximum value of the signal, minimum value of the signal, frequency content representative of size distribution in this volume, derivative values indicative of spatial changes in the volume or a multitude of spectral features including principle component spectra, relative spectra corresponding to different concentrations of molecules derived through spectra unmixing techniques.

[0073] Alternatively or additionally, assuming that multiple sensor measurements are performed as a function of time, one can also include time characteristics in the reference spectrum, including the mean over time, standard deviation, derivatives of change over time and all other processing schemes and feature extraction from time signals. This process generates therefore a multi-dimensional matrix that constitutes the reference spectrum and contains information about tissue feature on the volume (second volume) and implicitly the characteristics of the illumination parameters. Independent characterization of the illumination parameters (e.g. the relative or absolute intensity per wavelength) can be employed to normalize this reference spectrum, to separate the effects of the source from the tissue feature values contained in this volume.

[0074] Preferably, the at least one detection device is further configured to irradiate the tissue with ultrasound waves and to detect at least one ultrasound signal corresponding to ultrasound waves emanating from, in particular reflected by, the tissue in response to irradiating the tissue with the ultrasonic waves, and wherein the at least one processor is configured to identify, based on the at least one ultrasound signal, one or more interfaces between different tissue types in the tissue and / or one or more volumes or layers of different tissue types in the tissue.

[0075] Preferably, the at least one processor is configured to determine and / or specify the time gate, in particular the first time gate and / or the second time gate, based on the identified one or more interfaces between different tissue types in the tissue and / or the identified one or more volumes or layers of different tissue types in the tissue.

[0076] The process defined above for a first and second layer can be also adapted to operate with a third or more layers. In this case signals from a third or more layer(s) can be used to accurately compute tissue features and / or spectral modifications in subsequent layers, or define reference spectra in each of these layers. In this manner, modifications of the reference spectrum from more than one superficial layers are computed for each subsequent layer.

[0077] In these preferred embodiments, the ultrasound element employed for optoacoustic detection is preferably also employed to emit ultrasound waves and detect ultrasound waves that are interacting with tissue and propagate back to the detector. The ultrasound emission can be interleaved with the light emission. Typically, ultrasound and light intensity can be emitted simultaneously, and then optoacoustic signals separated by ultrasounds sig- nals reflected back from tissue in time, since optoacoustic signals appear first and ultrasound signals appear later in time. This is because the speed of light is faster than that of the ultrasound wave used for excitation. A sensor like this typically utilizes a switch that adapts the detector to the particular characteristics of ultrasound emission and ultrasound vs. optoacoustic detection, i.e. it adapts the amplification and sensitivity of the sensor and the ultrasound element to the requirements (signal to noise characteristics) of each of ultrasound or optoacoustic signals and possibly protects amplification units with a trans- mit / receive (T / R) switch.

[0078] Implementation of this hybrid optoacoustic / ultrasound operation provides additional information through the ultrasound contrast that can be employed for different applications. Related to the optoacoustic sensor, this hybrid operation can provide additional information through the ultrasound signal to assign adaptive time gates. This can be done by identifying signals on the ultrasound signal that correspond to different tissue interfaces representative of different volumes (layers) of interest.

[0079] The ultrasound signal can also be used in direct analogy to the optoacoustic signal processing described in the previous paragraphs to extract additional tissue features based on ultrasound contrast.

[0080] Preferably, the reference spectrum is used to correct the measurements from a volume (first volume) deeper in tissue, defined by a time gate (first time gate) at a later time point than the time gate (second time gate) corresponding to a surface volume or layer. The first time gate can be set with the same methods (see above) by which the second time gate is set. It can begin immediately after the second time gate, and its thickness can be set by known geometrical characteristics or adaptively. Likewise, a spectrum of the first volume (also referred to as “first volume spectrum”) can be defined with the same parameters as the reference spectrum of the second volume or layer, all defined with identical or similar methods.

[0081] Preferably, the first volume spectrum can be normalized by the reference spectrum. In a preferred embodiment, the spectrum is a one-dimensional vector of optoacoustic signal energy (signal intensity). Then, the spectrum of the first volume is normalized by the ref- erence spectrum in order to remove the effects of the illumination source and of the superficial layer and reveal more accurately the actual values of the tissue features in the first volume.

[0082] In a preferred embodiment, the second time gate is set to measure from the epithelial skin layer whereby the first time gate is set to measure from the epidermal-dermal junction which is rich in capillary loops. In the visible or near-infrared, the first time gate can be also set to measure from the dermal layer that contains blood microvessels known to be modified by disease. For example, vessel density is modified by progression of diabetes. Vessel endothelial function, which can be measured by sensor measurements over time (for example over 1 , 3 or 5 min) by recording relative time changes in the dermal layer, reveals cardiovascular disease progression.

[0083] Preferably, the time gates can be set in different volumes within skin or tissue to study tissue features associated with different conditions, such as measuring concentrations of tissue features such as glucose (typically in the SWIR or mid-IR) or changes in morphological and function features of the microvasculature (in the visible or NIR) or fat concentrations (e.d. using 930 nm illumination). However, a larger number of time gates (and therefore volumes) can be defined, for example one can use n = 3, 4, 5 or more gates to probe different volumes within tissue. In such cases for each deeper gate, it is possible to select a larger first gate to convert the problem to a two-gate problem (for further details see Fig. 7 and the corresponding description below). Alternatively, for each n gate the reference spectrum can be defined on the n-1 gate.

[0084] Preferably, the irradiation device is configured to irradiate the tissue simultaneously in all wavelengths by means of frequency-wavelength multiplexing (FWM), for realization of cost-effective sensors and / or true simultaneous illumination at multiple wavelengths. Multiple wavelengths, as applied in one or more wavelength pairs, can be also modulated by a FWM (frequency-wavelength multiplexing) scheme for concurrent illumination as described in Stylogiannis A., et.al. Nature Comm 13:4448 2022. In this case, pulse trains of slightly different periodicity are employed to carry the different wavelengths. The Fourier Transform of these pulse trains reveals multiple discrete modulation frequencies that are slightly different per wavelength carried. Then a demodulator can be employed to separate the frequencies and retrieve the optoacoustic signal strengths as a function of depth for each wavelength. FWM leads to SNR improvement over time-sharing methods. Preferably, the at least one irradiation device and the at least one detection device are provided in the same half-space above and / or adjacent to the irradiated tissue. Advantageously, with this embodiment, which is also referred to as “reflectance mode”.

[0085] Preferably, the first and / or second electromagnetic radiation exhibits two or more wavelengths or a wavelength range, and the at least one irradiation device is configured to scan, when irradiating the tissue with the first and / or second electromagnetic radiation, the two or more wavelengths or the wavelength range, respectively, within a predetermined period of time, in particular within less than one microsecond (1 ps), preferably within less than 100 ns.

[0086] Preferably, the concept of a time-gated correction based on a reference spectrum and a least one second spectrum is also well suited for using fast-wavelength-scanning techniques, that can utilize any light source, not necessarily pulsed light sources, for optoacoustic signal generation.

[0087] By means of fast scanning, ideally within less than a micro-second, or preferentially less than 100 ns, not only pulsed but also continuous wave sources can be employed for the direct generation of optoacoustic signals.

[0088] This concept generates optoacoustic signal based on the change of absorption in the absorption spectrum of at least one tissue within the volume scanned. This approach can lead to optoacoustic signal generation based on lower cost and higher energy light sources, as opposed to using pulsing light sources.

[0089] For example, a fast can would scan two or more wavelengths, in a discrete or continuous fashion within the range Xi - X2. The scan could be mono-directional (i.e. from Xi to X2), bidirectional or continuously alternating between directions or discrete wavelengths. Advantageously, for sufficient fast scanning, the optoacoustic signal generated depends on the absorption change within the scan range, not the absolute absorption.

[0090] In this case the reference spectrum and the at least second volume spectrum can be computed in the same way described for classical illumination, however in this case the computed quantities directly relate to the change of the at least one tissue feature. The spectral scanning characteristics can vary in this scheme, from continuous scanning over a selected wavelength (wavenumber) range, to two or more discrete wavelengths placed within this wavelength range. Such a scan can be implemented, e.g., with the bias-tree techniques discussed in the previous paragraphs or with fast switching between two or more wavelengths, using a separate or common light source, for example two or more QCLs, each one triggered by a pulse train shifted in time from the others. In a preferred embodiment, this time shift is shorter than the thermal relaxation of the molecules of interest. In the latter case, the pulse trigger scheme causes a modulation of the optoacoustic signals between the absorption states sampled by the two or more wavelengths.

[0091] Preferably, the at least one detection device is configured to detect the first optoacoustic signal many times within a period of time to obtain a series of first optoacoustic signals, respectively, and to store the series of first optoacoustic signals in a data memory. Preferably, the processor is configured to compute tissue features based on the time series. These features can be also computed for a first time gate, second time gate and so on. Therefore, the correction described above using a reference spectrum can also applied to a time series. Such correction (normalization) can be applied to each time point separately or to tissue features calculated over time, correcting the first volume spectrum with the reference spectrum of the second volume, in this case the spectrum containing tissue features based on time characteristics or spectral-time characteristics.

[0092] Preferably, the at least one first parameter and / or second parameter and / or third parameter characterizes at least one of the following tissue features: morphological and / or anatomical and / or structural features of the tissue, in particular regarding tissue and / or skin architecture, a size distribution of blood capillaries, a density of microvasculature, a thickness of the epithelial layer and / or a skin structure such as a tumor, and / or functional, physiological and / or molecular features of the tissue, in particular regarding the presence and / or concentration of a tissue feature and / or tissue component in the tissue such as biomolecules and / or metabolites such as glucose.

[0093] Further advantages, features and examples of the present disclosure invention will be apparent from the following description of following figures, wherein:

[0094] Fig. 1 shows a schematic representation of an example of an apparatus for optoacoustic sensing of biomolecules; Fig. 2 shows (a) an example of a cross-sectional optoacoustic image of skin tissue, and (b) an example of a corresponding raw optoacoustic signal as a function of depth within the skin tissue;

[0095] Fig. 3 shows a schematic regarding the difference between “in-blood” vs. interstitial fluid detection (left panel), and measurements with and without the influence of the skin surface (right panel);

[0096] Fig. 4 shows examples of absorption spectra of glucose and water;

[0097] Fir. 5 shows a schematic of an example to illustrate frequency-wavelength multiplexing in time and frequency domain;

[0098] Fig. 6 shows (right) an example of a cross-section of skin tissue showing different volumes or layers in a schematic representation, and (left) an example of a corresponding raw optoacoustic signal as a function of time corresponding to depth within the skin tissue;

[0099] Fig. 7 shows (left) an example of a cross-section of skin tissue showing in total three different volumes or layers, and (right) an example of a cross-section of skin tissue showing only two different volumes or layers as a result of selecting a larger first gate; and

[0100] Fig. 8 shows (a) an example of a raw optoacoustic signal and its conversion into an energy signal, and (b) an example of a raw optoacoustic signal and its processing in the Fourier domain to obtain an energy signal.

[0101] Figure 1 shows a schematic representation of an example of an apparatus for optoacoustic sensing of biomolecules, in particular metabolites such as glucose, in biological tissue 1. The apparatus comprises at least one irradiation device 2 configured to irradiate the tissue 1 with a first electromagnetic radiation 3 exhibiting one or more wavelengths or wavelength ranges in the mid-infrared (mid-IR) spectral range and at least one detection device 5 configured to detect at least one first optoacoustic signal corresponding to ultrasound waves 6 emanating from a region of interest ROI in the tissue 1 in response to irra- diating the tissue 1 with the first electromagnetic radiation 3. In present example, the irradiation device 2 and the at least one detection device 5 are provided in the same halfspace above and / or adjacent to the irradiated tissue 1 .

[0102] A processor 7 is configured to determine, based on the first optoacoustic signal, at least one first parameter representing a measure of a density of at least one biomolecule of interest in the region of interest ROI. In present example, the processor 7 is integrated, together with the irradiation device 2 and the detection device 5, in a miniaturized sensor device 8, such as a so-called wearable, which can be worn on a part of the animal or human body, for example on an arm. Alternatively, the processor 7 or a part thereof may be separate from and connected to the irradiation device 2 and the detection device 5 for signal and / or data communication via cables and / or a wireless data connection. The processor 7 preferably comprises a microcomputer and one or more data storage devices in communication therewith.

[0103] Preferably, the processor 7 is configured to perform a time-gating operation so as to process and / or consider only a part of the optoacoustic signal, which is obtained only from a certain volume, depth region and / or layer in the tissue 1. In this way, only optoacoustic signals corresponding to optoacoustic waves 6 which were detected within a predefined time gate G are considered for determining the first parameter representing a density of the biomolecule of interest in the ROI. The time gate G corresponds to a particular depth region and / or layer within the tissue 1.

[0104] This is further illustrated by Figure 2 showing (a) an example of a cross-sectional optoacoustic image of skin tissue, and (b) an example of a corresponding raw optoacoustic signal as a function of depth within the skin tissue. Different time gates GO, G1 , G2 contain information on different layers of the tissue. Preferably, by predefining and / or applying an appropriate time gate G, optoacoustic signals from skin layer(s) rich in vasculature, preferably from the epidermis-dermis junction (which is approx. 80 pm deep and contains capillary loops and the superficial vascular plexus), reached by mid-IR can be selected and / or considered for determining the density of the biomolecule of interest in the ROI, in particular tissue metabolites distributed in blood rather than in interstitial fluid. As apparent from Fig. 2, the optoacoustic image of the skin in the visible (for further details regarding visible irradiation of the tissue see below) shows strong vasculature at the epidermal-dermal junction corresponding to time gate G1 , while the signal shows how different time gates G contain information on different tissue layers.

[0105] This embodiment allows for an “in-blood” metabolite detection and, therefore, solves a key limitation of other sensor technologies, which measure glucose “in bulk” from interstitial fluids. In particular, using time gating of the mid-IR optoacoustic signal (i.e. of the optoacoustic signal obtained in response to irradiating the tissue with mid-IR radiation) has the following advantages: 1 . rejecting signal contamination from the skin surface (cf. time gate GO in Fig. 2) by focusing on later arriving signals from deeper skin layers (cf. time gate G1 in Fig. 2), and 2. focusing the measurement and / or signal processing only to layers (cf. time gate G1 in Fig. 2) rich in microvasculature, i.e. at the epidermal-dermal junction, approx. 80 to 100 pm deep), as opposed to superficial interstitial fluid achieved by conventional transdermal (and mid-IR) sensors.

[0106] This is further illustrated by Figure 3, which shows a schematic regarding the difference between “in-blood” vs. interstitial fluid detection (left panel), and measurements with and without the influence of the skin surface (stratum corneum) (right panel).

[0107] Figure 3 (a) shows a bright-field image of a mouse ear in vivo, with P1 marking a location at blood vessels, and P2 marking a location without blood vessels, (b) Same mouse ear area at 532 nm and (c) at 2850 cm-1(lipids), with an overlay of (b). (d-e) Clarke’s error grid for P1 (d) and P2 (e) showing the correlation between reference glucose and measured glucose, (f) Distribution of results per region and root mean square error of cross-validation (RMSECV). (g) Bright-field image of another mouse ear in vivo, darker structures: vasculature, brighter structures: lipids. Comparison to (c) shows that different locations may present different surface conditions, (i, j) Clarke’s error grid for 10 experiments when measuring at P1 with (j) and without (i) skin rejection - based on the layer selection method, (h) Distribution of results per region and root mean square error of cross-validation (RMSECV) with skin-rejection (SR) or without SR.

[0108] Preferably, the irradiation device 2 (see Figure 1) is configured to irradiate the tissue 1 with a second electromagnetic radiation 4 exhibiting one or more wavelengths or wavelength ranges outside the mid-infrared spectral range, in particular in the visible spectral range. Preferably, the ROI of the tissue 1 is simultaneously irradiated with the first electromagnetic radiation 3 and second electromagnetic radiation 4. The at least one detection device 5 is configured to detect at least one second optoacoustic signal corresponding to ultrasound waves 6 emanating from the region of interest ROI in the tissue 1 in response to irradiating the tissue 1 with the second electromagnetic radiation 4. The processor 7 is configured to determine, based on the second optoacoustic signal, at least one second parameter representing a measure of another density, which can be considered as a reference density or normalizing density, in particular a blood density or vascular density of blood vessels 9 and / or a lipid density, in the tissue 1 containing the at least one biomolecule of interest, and to determine at least one third parameter based on the at least one first parameter and the at least one second parameter, in particular by normalizing the at least one first parameter by the at least one second parameter.

[0109] While using time gating in the mid-IR as described above gives a measure of metabolite density in a desired layer and / or volume (typically 60x60x60 pm), co-illuminating with a preferably visible wavelength (e.g. 488 nm, 532 nm or 630nm), which is sensitive to vasculature, and using the same gate (e.g. G1 , see Fig. 2), one can also retrieve and / or determine a measure of the vascular density at the same exact volume (lateral dimension been given by the common ultrasound detector). In other words, this embodiment includes a concept of referencing mid-IR measurements of metabolite concentration to the amount of vascular density in the same volume (layer). As metabolite concentration in blood is about 10x higher than in ISF, the measurement of vascular density in the same, preferably in the exact same, layer measured by mid-IR improves the accuracy of the determined metabolite concentration considerably. Preferably, referencing the mid-IR value (i.e. the first parameter relating to the metabolite concentration or density) to the amount of vascular density (i.e. second parameter) can be done by dividing the mid-IR value (first parameter) with an actual value of vascular density (second parameter) at the same point and / or volume and / or layer of measurement, rather than just assuming a uniform vascular density value.

[0110] Preferably, the first electromagnetic radiation 3 exhibits a first wavelength Xi, which is coincident with a peak wavelength at which the biomolecule of interest exhibits a (local) maximum absorption in the mid-infrared spectral range, and a second wavelength X2, which is offset from the first wavelength Xi and at which the biomolecule of interest exhibits an absorption which is less than the (local) maximum absorption and / or which is in the range of or less than a background absorption caused by one or more molecules, in particular water, other than the biomolecule of interest. The at least one detection device 5 is configured to detect a first optoacoustic signal of the first wavelength, the first optoacoustic signal of the first wavelength corresponding to ultrasound waves 6 emanating from the region of interest ROI in response to irradiating the tissue 1 with the first electromagnetic radiation 3 exhibiting the first wavelength Xi, and to detect a first optoacoustic signal of the second wavelength, the first optoacoustic signal of the second wavelength corresponding to ultrasound waves 6 emanating from the region of interest ROI in response to irradiating the tissue 1 with the first electromagnetic radiation 3 exhibiting the second wavelength X2. The processor 7 is configured to determine the at least one first parameter (representing a measure of a density of the at least one biomolecule of interest in the ROI) based on the first optoacoustic signal of the first wavelength and the first optoacoustic signal of the second wavelength and / or based on a background-corrected optoacoustic signal which is obtained by subtracting the first optoacoustic signal of the second wavelength from the first optoacoustic signal of the first wavelength.

[0111] This is illustrated by Figure 4 showing examples of absorption spectra S1 , S2 of glucose and water. The absorption spectrum S1 of glucose exhibits a maximum (peak) at a first wavelength Xi, whereas the spectrum S2 shows that water provides a relatively homogeneous, yet high absorption background. The tissue 1 is irradiated with electromagnetic radiation 3 exhibiting a wavelength pair Xi, X2, wherein the first wavelength Xi coincides with the wavelength of the maximum absorption of glucose (see S1) and a second wavelength X2which is slightly offset from the wavelength of the maximum absorption of glucose. As a result, the optoacoustic signal obtained in response to the first wavelength Xi contains information on the absorption of both glucose and water background, whereas the optoacoustic signal obtained in response to the second wavelength X2contains information of the absorption of water background only. Subtracting the latter from the former optoacoustic signal reveals a background-corrected optoacoustic signal of the glucose absorption representing a measure of the density of at least one biomolecule of interest in the ROI.

[0112] Preferably, the apparatus (see Fig. 1) is configured to irradiate the tissue 1 and detect ultrasound waves 6 emanating from the tissue 1 in response thereto by means of frequency-wavelength multiplexing (FWM). Preferably, irradiating the tissue 1 with first electromagnetic radiation 3 in the mid-IR and second electromagnetic radiation 4 in the visual spectral range and / or with first electromagnetic radiation 3 exhibiting a first wavelength Xi and a second wavelength A2and detecting the respectively generated ultrasound waves 6 as described above, is realized by applying FWM.

[0113] Figure 5 shows a schematic of an example to illustrate FWM in time and frequency domain with three wavelengths. The tissue 1 is irradiated with light pulses of three parallel pulse trains (indicated by three arrows of different brightness) acting as the equivalent of a frequency comb to produce a multitude of discrete frequencies in the frequency domain. Each pulse train carries one irradiation wavelength, and it has a unique periodicity that is slightly different than all other pulse trains concurrently employed. As a result, each wavelength is carried at multiple discrete frequencies (F) of the Fourier Transform of the respectively collected optoacoustic signal. In this way, FWM allows for a simultaneous illumination of the tissue 1 with all colors. Further, FWM results in a considerable signal-to-noise enhancement.

[0114] Figure 6 (right) shows a schematic representation of an example of a cross-section of skin tissue showing different volumes or layers. In the given example, the skin tissue shows a tissue surface, which can be regarded as a first volume or layer of the tissue, a second volume located beneath the tissue surface in a deeper region of the tissue, and a third volume located beneath the second volume in an even deeper region in the tissue. A sensor, also referred to as “detection device 5” (see Fig. 1), is placed on the tissue surface.

[0115] Figure 6 (left) shows an example of a corresponding raw optoacoustic signal generated by the sensor as a function of time and corresponding to different depths within the skin tissue. By defining and / or applying different gates Gate 1 , Gate 2 and Gate 3, the optoacoustic signal obtained during these time periods contains information on different volumes at different depths in the tissue, i.e. , in the given example the tissue surface (corresponding to the first gate), the second volume (corresponding to the second gate) and the third volume corresponding to the third gate), respectively.

[0116] According to a preferred aspect of present disclosure, the optoacoustic signal (or a parameter derived therefrom) obtained from a certain volume or layer of the tissue is corrected, in particular normalized, by the optoacoustic signal (or a parameter derived therefrom) obtained from a different volume or layer of the tissue. For example, the optoacoustic signal (or the energy signal and / or a parameter derived therefrom) obtained within the second gate (Gate 2) and corresponding to the second volume is corrected, in particular normalized, by the optoacoustic signal (or the energy signal and / or a parameter derived therefrom) obtained within the first gate (Gate 1) corresponding to the tissue surface (first volume or layer). Likewise, the optoacoustic signal (or the energy signal and / or a parameter derived therefrom) obtained within the third gate (Gate 3) and corresponding to the third volume can be corrected, in particular normalized, by the optoacoustic signal(s) (or the energy signal(s) and / or parameter(s) derived therefrom) obtained during the first gate (Gate 1) and / or the second gate (Gate 2) corresponding to the tissue surface (first volume or layer) or second volume of the tissue, respectively.

[0117] Figure 7 (left) shows an example of a cross-section of skin tissue showing in total three different volumes or layers, wherein the above explanations regarding Figure 6 (right) apply accordingly. In general, more than two time gates (and therefore volumes) can be defined, for example one can use n = 3, 4, 5 or more gates to probe different volumes within the tissue.

[0118] Figure 7 (right) shows an example of a cross-section of skin tissue showing, as a result of selecting a larger first gate, only two different volumes or layers to illustrate that, where a larger number of time gates (and therefore volumes) is defined as exemplarily shown in Figure 7 (left), for each deeper gate, it is possible to select a larger first gate to convert the problem to a two-gate problem. Alternatively, for each n gate a reference spectrum (as discussed in detail above) can be defined on the n-1 gate.

[0119] It is important to note that the terms "first gate" and "second gate" used in Figures 6 and 7and the corresponding description have been chosen according to their order in the skin (for the sake of clarity) and cannot necessarily be equated with the terms "first time gate" and "second time gate", respectively, used in the rest of the description and in the claims, since the "first time gate" always refers to the first optoacoustic signal to be corrected, and the "second time gate" always refers to the second optoacoustic signal with which the first optoacoustic signal is corrected. Therefore, in cases where the optoacoustic signal obtained from a volume in a deeper region of the tissue (e.g. from the second or third volume shown in Figures 6 and 7) is corrected by the optoacoustic signal obtained from a surface volume or layer of the tissue, the "first gate" shown in figures 6 and 7 corresponds to the “second time gate”, and the "second gate" and / or “third gate” shown in figures 6 and 7 corresponds to the “first time gate”. Figure 8 (a) shows an example of a raw optoacoustic signal and its conversion into an energy signal, and Figure 8 (b) shows an example of a raw optoacoustic signal and its processing in the Fourier domain to obtain an energy signal.

[0120] As illustrated in Figure 8 (a), the energy signal, which is also referred to as “converted optoacoustic signal” herein, comprises only positive signal components characterizing an amplitude, intensity or energy of the raw optoacoustic signal as a function of time. As indicated by gates G1 , G1 and G3 along the time axis, the time-gating operations as described above are preferably performed on the energy signal, i.e. after converting the detected raw optoacoustic signal into the energy signal.

[0121] As illustrated in Figure 8 (b), converting the raw optoacoustic signal into the energy signal can be achieved by different operations on the collected optoacoustic signal, for example by the Hilbert transform. Alternatively, one can obtain the Fourier transform of the collected optoacoustic signal and process the Fourier-transformed signal directly in the frequency domain. Another option is to convert the Fourier-transformed signal back to the time-do- main, using the inverse Fourier transform, after the Fourier transform representation of the signal has been processed so that it loses the phase (i.e. the imaginary part) and is converted to an amplitude-only signal in the Fourier transform domain (i.e. the frequency domain). Operations in the frequency domain preferably include taking the absolute of the signal or the squared of the signal and then applying the inverse Fourier transform. Other transforms, including wavelet transforms, envelope computations and other such conventional signal processing schemes can be applied to the optoacoustic signal in either time- or frequency-domain.

Claims

Patent Claims1. An apparatus for optoacoustic sensing of tissue features in biological tissue (1 ), the apparatus comprising:- at least one irradiation device (2) configured to irradiate the tissue (1) with a first electromagnetic radiation (3) exhibiting one or more wavelengths or wavelength ranges in a first spectral range,- at least one detection device (5) configured to detect at least one first optoacoustic signal corresponding to ultrasound waves (6) emanating from a region of interest (ROI) in the tissue (1) in response to irradiating the tissue (1) with the first electromagnetic radiation (3), and- a processor (7) configured to determine, based on the first optoacoustic signal, at least one first parameter characterizing at least one first tissue feature of interest in the region of interest (ROI).

2. The apparatus according to claim 1 , wherein- the irradiation device (2) is configured to irradiate the tissue (1) with a second electromagnetic radiation (4) exhibiting one or more wavelengths or wavelength ranges in a second spectral range,- the at least one detection device (5) is configured to detect at least one second optoacoustic signal corresponding to ultrasound waves (6) emanating from the region of interest (ROI) in the tissue (1) in response to irradiating the tissue (1) with the second electromagnetic radiation (4),- the processor (7) is configured to determine, based on the second optoacoustic signal, at least one second parameter characterizing at least one second tissue feature of interest in the region of interest, and to determine at least one third parameter based on the at least one first parameter and the at least one second parameter, in particular by normalizing and / or correcting the at least one first parameter by the at least one second parameter.

3. The apparatus according to claim 1 or 2, wherein- the first spectral range is in the mid-infrared range and / or- the second spectral range is outside the first spectral range, in particular in the visible spectral range, and / or- the at least one first parameter represents a measure of a density of at least one biomolecule of interest, in particular at least one metabolite, in the region of interest (ROI) and / or- the at least one second parameter represents a measure of at least one of i) a blood density characterizing an amount of hemoglobin molecules present in the tissue (1) and / or ii) a vascular density characterizing an amount of blood vessels (9) containing hemoglobin molecules and / or iii) a lipid density characterizing an amount of lipid molecules present in the tissue (1 ) and / or the presence of adipose tissue in the tissue (1).

4. The apparatus according to claim 3, wherein the density of at least one biomolecule of interest in the region of interest characterizes an absolute or a relative amount, in particular a mass, number of molecules, volume or concentration, of the at least one biomolecule present in the region of interest.

5. The apparatus according to any of the preceding claims, wherein the at least one irradiation device (2) and the at least one detection device (5) are provided in the same half-space above and / or adjacent to the irradiated tissue (1).

6. The apparatus according to any of the preceding claims, wherein- the first electromagnetic radiation (3) exhibits a first wavelength (Xi), which is coincident with a peak wavelength at which a biomolecule of interest exhibits a (local) maximum absorption in the first spectral range, in particular in the mid-infra- red spectral range,- the first electromagnetic radiation (3) exhibits a second wavelength (X2), which is offset from the first wavelength (Xi) and at which the biomolecule of interest exhibits an absorption which is less than the (local) maximum absorption and / or which is in the range of or less than a background absorption caused by one or more molecules, in particular water, other than the biomolecule of interest,- the at least one detection device (5) is configured to detect a first optoacoustic signal of the first wavelength (Xi), the first optoacoustic signal of the first wavelength (Xi) corresponding to ultrasound waves (6) emanating from the region of interest (ROI) in response to irradiating the tissue (1) with the first electromagnetic radiation (3) exhibiting the first wavelength (Xi), and to detect a first optoacoustic signal of the second wavelength (X2), the first optoacoustic signal of the secondwavelength (X2) corresponding to ultrasound waves (6) emanating from the region of interest (ROI) in response to irradiating the tissue (1) with the first electromagnetic radiation (3) exhibiting the second wavelength (X2), and- the processor (7) is configured to determine the at least one first parameter based on the first optoacoustic signal of the first wavelength (Xi) and the first optoacoustic signal of the second wavelength (X2).

7. The apparatus according to claim 6, wherein the processor (7) is configured to determine the at least one first parameter based on a background-corrected optoacoustic signal which is obtained by subtracting the first optoacoustic signal of the second wavelength (X2) from the first optoacoustic signal of the first wavelength (Xi).

8. The apparatus according to any of the preceding claims, wherein- the at least one detection device (5) is configured to detect the first optoacoustic signal and / or second optoacoustic signal many times within a period of time to obtain a series of first optoacoustic signals and / or second optoacoustic signals, and to store the series of first optoacoustic signals and / or second optoacoustic signals in a data memory, and- the processor (7) is configured to apply a moving average and / or median filtering operation on the series of first optoacoustic signals and / or second optoacoustic signals to obtain an averaged first optoacoustic signal and / or an averaged second optoacoustic signal, and to determine the at least one first parameter based on the averaged first optoacoustic signal and / or to determine the at least one second parameter based on the averaged second optoacoustic signal.

9. The apparatus according to any of the preceding claims, wherein the at least one detection device (5) is configured to detect the at least one first and / or second optoacoustic signal over time, and wherein the processor is configured to convert the first and / or second optoacoustic signal into a first and / or second converted optoacoustic signal (energy signal), respectively, which comprises, in particular only positive, signal components characterizing an amplitude and / or intensity and / or energy of the first and / or second optoacoustic signal as a function of time, and to determine the at least one first and / or second parameter based on the first and / or second converted optoacoustic signal, respectively.

10. The apparatus according to any of the preceding claims, wherein the processor (7) is configured to select a portion of the first and / or second, in particular converted, optoacoustic signal, which has been detected during a specified time gate corresponding to a volume or layer (GO, G1 , G2) in the tissue (1), and to determine the at least one first and / or second parameter based on the selected portion of the first and / or second optoacoustic signal, respectively.

11. The apparatus according to claim 10, wherein the processor (7) is configured to select a first portion of the first optoacoustic signal, which has been detected during a specified first time gate, to select a second portion of the second optoacoustic signal, which has been detected during a specified second time gate, the specified first time gate and specified second time gate corresponding to different volumes or layers (GO, G1 , G2) in the tissue (1), to determine the at least one first parameter based on the selected first portion of the first optoacoustic signal, to determine the at least one second parameter based on the selected second portion of the second optoacoustic signal, and to determine the at least one third parameter based on the at least one first parameter and the at least one second parameter, in particular by normalizing and / or correcting the at least one first parameter by the at least one second parameter.

12. The apparatus according to claim 11 , wherein the specified first time gate corresponds to a first volume or layer (G1 , G2) in the tissue and the specified second time gate corresponds to a second volume or layer (GO) in the tissue (1), wherein the first volume or layer (G1 , G2) is located in a deeper region of the tissue (1) than the second volume or layer (GO), which is in particular located at a superficial region of the tissue (1).

13. The apparatus according to claim 11 or 12, wherein the specified first time gate corresponds to an epidermal-dermal junction layer in the tissue or a dermal layer containing blood micro vessels, and the specified second time gate corresponds to an epithelial skin layer of the tissue (1).

14. The apparatus according to any of claims 10 to 13, wherein the processor is configured to determine and / or specify the time gate, in particular the first time gate and / or the second time gate, based on a pre-definable position and / or dimension,in particular diameter and / or thickness, of a volume or layer of interest in the tissue (1).

15. The apparatus according to any of the claims 11 to 14, wherein the processor is configured to determine, based on the selected first and / or second portion of the first and / or second optoacoustic signal, at least one of the following first and / or second parameters:- a total energy of the selected first and / or second portion of the first and / or second optoacoustic signal,- a maximum value and / or minimum value and / or mean value and / or standard deviation value of the selected first and / or second portion of the first and / or second optoacoustic signal,- a frequency content in the selected first and / or second portion of the first and / or second optoacoustic signal characterizing a size distribution of components, e.g. blood vessels, of the first and / or second volume or layer in the tissue,- one or more derivative values of the selected first and / or second portion of the first and / or second optoacoustic signal indicative of spatial changes in the first and / or second volume or layer in the tissue,- one or more spectral features of the selected first and / or second portion of the first and / or second optoacoustic signal, in particular including principle component spectra and / or relative spectra corresponding to different concentrations of molecules derived through spectral unmixing techniques, wherein the above first and / or second parameters are determined for one or more wavelengths or wavelength ranges in the first and / or second spectral range and / or optionally include a time characteristics of the respective parameters.

16. The apparatus according to claim 15, wherein the processor is configured to generate, from at least two of the first and / or second parameters, respectively, for the respectively selected first and / or second portion of the first and / or second optoacoustic signal:- a one-dimensional vector, or- a multi-dimensional matrix if the first and / or second parameters were determined for two or more wavelengths or wavelength ranges and / or if the first and / or second parameters include a time characteristics of the respective parameters.

17. The apparatus according to claim 16, wherein the processor is configured to determine the at least one third parameter from the respectively generated one-dimensional vectors or multidimensional matrices of the by means of at least one vector or matrix operation, respectively.

18. The apparatus according to any of the preceding claims, wherein the at least one detection device (5) is further configured to irradiate the tissue with ultrasound waves and to detect at least one ultrasound signal corresponding to ultrasound waves emanating from, in particular reflected by, the tissue (1) in response to irradiating the tissue (1) with the ultrasonic waves, and wherein the at least one processor is configured to identify, based on the at least one ultrasound signal, one or more interfaces between different tissue types in the tissue (1) and / or one or more volumes or layers of different tissue types in the tissue (1).

19. The apparatus according to any of the claims 10 to 17, wherein the at least one processor is configured to determine and / or specify the time gate, in particular the first time gate and / or the second time gate, based on the identified one or more interfaces between different tissue types in the tissue (1) and / or the identified one or more volumes or layers of different tissue types in the tissue (1).

20. The apparatus according to any of the preceding claims, wherein the first and / or second electromagnetic radiation (3) exhibits two or more wavelengths or a wavelength range, and the at least one irradiation device (2) is configured to scan, when irradiating the tissue with the first and / or second electromagnetic radiation, the two or more wavelengths or the wavelength range, respectively, within a predetermined period of time, in particular within less than one microsecond (1 ps), preferably within less than 100 ns.21 . The apparatus according to any of the preceding claims, wherein the at least one first parameter and / or second parameter and / or third parameter characterizes at least one of the following tissue features:- morphological and / or anatomical and / or structural features of the tissue, in particular regarding tissue and / or skin architecture, a size distribution of blood capillaries, a density of microvasculature, a thickness of the epithelial layer and / or a skin structure such as a tumor,- functional, physiological and / or molecular features of the tissue, in particular regarding the presence and / or concentration of a tissue feature and / or tissue component in the tissue such as biomolecules and / or metabolites such as glucose.

22. A method for optoacoustic sensing of tissue features in biological tissue (1), the method comprising:- irradiating the tissue (1) with a first electromagnetic radiation (3) exhibiting one or more wavelengths or wavelength ranges in a first spectral range,- detecting at least one first optoacoustic signal corresponding to ultrasound waves (6) emanating from a region of interest (ROI) in the tissue (1) in response to irradiating the tissue (1) with the first electromagnetic radiation (3), and- determining, based on the first optoacoustic signal, at least one first parameter characterizing at least one first tissue feature of interest in the region of interest (ROI).

23. A method for processing optoacoustic signals, the method comprising:- providing and / or receiving at least one first optoacoustic signal which has been obtained by detecting ultrasound waves (6) emanating from a region of interest (ROI) in a biological tissue (1) in response to irradiating the tissue (1) with a first electromagnetic radiation (3) exhibiting one or more wavelengths or wavelength ranges in a first spectral range, and- determining, based on the first optoacoustic signal, at least one first parameter characterizing at least one first tissue feature of interest in the region of interest (ROI).

24. A computer program product comprising instructions causing- a computer, computer system and / or distributed computing environment to execute the steps of the method according to claim22 or 23, and / or- the processor (7) of the apparatus according to any of the claims 1 to 21 to execute the steps of the method according to claim 22 or 23.

Images