EP3016590A1

EP3016590A1 - Apparatus&method for determining the concentration of a substance in a fluid

Info

Publication number: EP3016590A1
Application number: EP14748133.7A
Authority: EP
Inventors: Panayiotis Anastasios Kyriacou; Victor Olagovich Rybynok
Original assignee: City University of London
Current assignee: City University of London
Priority date: 2013-07-02
Filing date: 2014-07-02
Publication date: 2016-05-11
Also published as: GB201311859D0; WO2015000997A1

Abstract

A method for determining the concentration of a substance in a fluid, the method comprising: (i) operating a light emitter to illuminate a fluid; (ii) operating a light detector to determine the intensity of light that has passed through the fluid to the detector; (iii) calculating an amount of light absorbed by the fluid; (iv) determining a mean light pathway through the fluid from the detector to the emitter; (v) retrieving data concerning the optical absorptivity of the substance; and (vi) determining from the calculated absorbance, the determined mean light pathway and the retrieved absorptivity of the substance, a measure of the concentration of the substance in the fluid. Apparatus for determining the concentration of a substance in a fluid is also disclosed.

Description

Apparatus & Method for Determining

the Concentration of a Substance in a Fluid

Field

This invention relates to apparatus for measuring the concentration of a substance in a fluid. A method is also disclosed.

In one illustrative implementation, the fluid comprises blood, and the apparatus is capable of measuring the concentration of a component of the blood, for example glucose. In a preferred implementation the apparatus is operable to non-invasively measure the concentration of the substance.

The teaching provided below refers in detail to an apparatus that is configured for non-invasively determining the concentration of glucose in a subject's blood, but it will be appreciated and should be remembered that this is merely illustrative of the teachings of the invention and that the teachings of the invention have many other applications. For example, the apparatus disclosed herein could be used to measure the concentration of other blood components, in vitro or in vivo, or indeed components of other fluids.

Background

It is increasingly apparent that diabetes will be a significant health challenge for the 21 st century. Due to an obesity epidemic, increasingly sedentary lifestyles and an ageing population, it is the case that the mean prevalence of this condition is currently doubling every generation.

Those persons who are suffering from diabetes need to periodically determine the concentration of glucose in their blood. The most common way of doing this is the invasive "fingerstick" method that has been in use for at least the last thirty years. This method involves the subject periodically pricking their finger and testing a small sample of their blood with a portable meter (such as the CONTOUR^® USB meter sold by Bayer pic, for example).

Blood glucose monitors that continuously monitor the concentration of glucose in the blood are also available. These continuous blood glucose monitors (CGM) are also invasive, and typically comprise a disposable glucose sensor that is placed just under the skin, and is worn for a few days until it is replaced. The sensor is linked to a non- implanted transmitter that wirelessly communicates with an electronic receiver that is worn like a pager and displays blood glucose levels on a practically continuous manner, as well as monitors rising and falling trends. Such CGMs, whilst providing for continuous measurement of blood glucose levels, have a number of drawbacks. For example, such CGM devices still need to be calibrated with a traditional blood glucose "fingerstick" system. Another issue is that these devices measure glucose concentration in interstitial fluid, and the level of glucose in a subject's interstitial fluid tends to lag behind the level of glucose in the subject's blood, and thus the accuracy of the measurements obtained may be compromised.

Non-invasive blood glucose measurement devices have previously been proposed, but a significant problem is the calibration of such devices. In particular, as the amount of protein, fats, and water in the blood vary widely from person to person, a universal measurement method has not yet been developed. These, and other, vascular tissue components render empirical calibration difficult (if not impossible), and such calibration is necessary for the accurate evaluation of blood glucose concentration using conventional spectroscopic techniques. Previously proposals for addressing this calibration problem have used special empirical algorithms and large sets of experimentally obtained data, and the techniques previously proposed vary from simple curve fitting methods to regression analysis, and artificial neural networks. Despite many prior proposals, decades of experimental research would tend to indicate that direct empirical use of information obtained from an optical sensor to predict blood glucose concentration is highly problematic, and probably impossible.

That said, it is also the case that apparatus that could continuously and noninvasive^ measure the concentration of a blood component (such as glucose) would have significant impact on the quality of care that could be delivered to patients. For example, continuous monitoring of blood glucose would allow the close examination of how the blood glucose level reacts to insulin, exercise, food, and other factors. Such a monitor would be of utility in virtually every department of a hospital. Perhaps more importantly its impact in the community would be even greater. Every individual diabetic could have one, allowing for better trend monitoring and also allowing their physician to see how well their diabetes has been controlled. Diabetics, general practitioners, emergency and community services would all gain from such a device. This device may also have economic advantages as it would not need disposables. It could be incorporated into other clinical devices and also potentially in every-day devices such as a watch or mobile phone.

It is one aim of the present invention to provide apparatus that can be operated to measure the concentration of a substance in a fluid, particularly but not exclusively to provide apparatus that can be used to non-invasively measure the concentration of a blood component (such as glucose) in the blood of a subject. Summary

To this end, a presently preferred embodiment of the present invention provides a method for determining the concentration of a substance in a fluid, the method comprising: operating a light emitter to illuminate a fluid; operating a light detector to determine the intensity of light that has passed through the fluid to the detector; calculating an amount of light absorbed by the fluid; determining a variation in mean light pathways through the fluid from the detector to the emitter; retrieving data concerning the absorptivity of the substance; and determining from the calculated absorbance, the determined mean light pathway variation and the retrieved absorptivity of the substance, a measure of the concentration of the substance in the fluid.

Another embodiment of the present invention provides: apparatus for determining the concentration of a substance in a fluid, the apparatus comprising: a light emitter for illuminating a fluid; a light detector for determining the intensity of light that has passed through the fluid to the detector; means for calculating an amount of light absorbed by the fluid; means for determining a variation in mean light pathways through the fluid from the detector to the emitter; means for retrieving data concerning the absorptivity of the substance; and means for determining from the calculated absorbance, the determined mean light pathway variation and the retrieved absorptivity of the substance, a measure of the concentration of the substance in the fluid.

In envisaged implementations of the teachings of the invention, the fluid comprises blood flowing through vascular tissue, the apparatus and method being operable to provide a non-invasive measure of the concentration of the substance in the blood. In this instance, the variation in mean light pathways through the vascular tissue may result from expansion and contraction of blood vessels within the tissue as the heart of the subject pumps blood around their body.

The variation in mean light pathway may be determined by measuring a Doppler frequency shift that is representative of the speed at which the vascular tissue expands and contracts.

The emitter may comprise a multi-wavelength monochromatic light source, such as a plurality of LEDs. The Doppler frequency shift may be observed in a laser that is arranged to illuminate the vascular tissue.

Other features, advantages, aims and aspects of the present invention are set out in the following detailed description and claims. Brief Description of the Drawings

Various aspects of the teachings of the present invention, and arrangements embodying those teachings, will hereafter be described by way of illustrative example with reference to the accompanying drawings, in which:

Fig. 1 a schematically illustrates the light flux envelope between an emitter and a detector through a scattering medium, such as vascular tissue;

Fig. 1 b schematically illustrates the propagation delay of light traversing the scattering medium;

Fig. 2a schematically illustrates the concept of the mean light pathway;

Fig. 2b schematically illustrates a mean light pathway through a non- homogeneous scattering medium;

Fig. 3a schematically illustrates vascular tissue volume changes through a subject's cardiac cycle;

Fig. 3b schematically illustrates a PPG signal;

Fig. 4 is a schematic representation of apparatus embodying the teachings of the present invention; and

Fig. 5 is a functional representation of apparatus embodying the teachings of the present invention.

Detailed Description

Before providing a detailed explanation of one illustrative arrangement that embodies the teachings of the present invention, it is appropriate at this juncture to provide a summary of the physical and mathematical principals on which the teachings of the present invention are based.

Photoplethysmography (PPG) is a measurement technique that uses light to noninvasive^ obtain a volumetric measurement of an organ with each cardiac cycle. Pulse Oximetry (PO) is a well established empirical technique that allows the degree of arterial blood oxygen saturation (Sp0₂) to be evaluated from PPG signals. An advantage of PO is that one can look directly into the arterial blood stream through the skin and bypass other parts of the vascular tissue (e.g. bones or muscles). The apparatus disclosed herein can be used, inter alia, to non-invasively evaluate arterial blood glucose concentration from PPG signals using a technique that we call "Dynamic Pulsatile Spectroscopy" (DPS). In developing our technique, we have devised a rational model that is based on the fundamental laws of physics and enables a mathematical analysis of the PPG signals and arterial blood constituents to be undertaken.

Pulse Oximetry is based on a rational model that utilises the conventional Beer- Lambert law, but neglects scattering within vascular tissue, and assumes that arterial blood chromophores consist of only oxygenated and deoxygenated haemoglobins. Despite these assumptions, the equations produced from that PO model enable Sp0₂ values to be estimated from PPG signals quite accurately with respect to empirical relationships established with invasive blood oxygen saturation measurements. Although the analytical solution obtained from the rational model is not as accurate as empirically obtained results, it theoretically proves that non-invasive Sp0₂ evaluation is generally possible. The analytical solution equations also reveal which parameters should be used for empirical calibration (i.e. the so-called R values).

Unlike the model used for PO, the DPS technique that we have developed takes vascular tissue scattering into consideration and includes additional arterial blood analytes. In order to combine light scattering with conventional absorption spectroscopy, our DPS technique uses a method that we have named Beer-Lambert law along Non- Linear mean Light Pathways (BLNLP).

In BLNLP, light is considered to be a flux of elementary particles - namely, photons. The speed of these photons is equal to the speed of light, and is defined by the electromagnetic properties of the media through which the light propagates. A light photon can be emitted absorbed or scattered by an optical electron of an atom or molecule, collectively called matter particles. Photon energy is proportional to the light wavelengths and the proportionality coefficient is the Plank constant. Additional postulates used in BLNLP were deduced by analysis of the following physical models: the Beer-Lambert law, Monte Carlo light scattering modelling and the light energy transport integral equation.

Although the motion of individual photons is determined statistically by the photon scattering and scattering phase distribution functions, we have determined that the average light energy propagation pathways within the vascular tissues are deterministic and not random. Fig. 1 demonstrates this physical concept and shows a schematic representation of the light flux envelopes 7 between an emitter 1 and detector 3 in a scattering media 5 (for example, through vascular tissue).

Each envelope has a characteristic "banana" shape, and represents a space area through which most photons travel from the light emitter 1 to the light detector 3. Such envelopes can be obtained by light scattering modelling on the homogeneous, plain, and semi-infinite media models utilising the Monte Carlo method to solve the light radiation transfer equation.

The motion of the photons can be explained from two points of view: dynamically by means of light pulse propagation delay experiments (as depicted schematically in Fig. 1 b), and statically by solving light transfer equations using Monte Carlo analysis (as depicted schematically in Fig. 1 a). Referring to Figs. 1 a and 1 b, time interval " during which light pulse energy travels to the light detector corresponds to light energy with the shortest pathlength in the banana envelope. Time interval "II" corresponds to a longer pathlength, and so on up to the end of the light pulse propagation time interval "VI".

The light energy propagation envelopes shown in Fig. 1 a can be split into a series of smaller "canoe" shaped envelopes 9 through each of which a certain fraction of light energy propagates through the matter. As shown in Fig. 2a, each of these canoe shapes equates to the mean light pathway through the envelope median. In this approximation, fractions of the light energy corresponding to each envelope are propagating along the mean light pathways, and each mean light pathway corresponds to the particular light pulse propagation delay time intervals depicted in Fig. 1 b.

Once we have introduced the concept of the light propagation pathway in the scattering media, the light intensity can be introduced for the resulting scattered light. Light intensity can be computed for each point of the mean light pathway curve along its tangent and in the emitter-detector direction. Thus light intensity degrades along the mean pathways according to the Beer-Lambert law, apart from the fact that the Beer- Lambert integral is taken along the mean light pathways rather than a conventional straight line:

where: Α_ρ._λ is the absorption for pathway p, from emitter to detector for light of wavelength λ; I_{D ρ}._λ is the intensity of light of wavelength λ at the detector attributable to pathway ρ,; Ιε_Ρί λ '^s the intensity of light of wavelength λ at the emitted attributable to pathway p,; l_pi is the mean light pathway for pathway p,; and μ_{Άΐ λ} is the attenuation coefficient of the medium through which light of wavelength I travels.

If we assume, as shown schematically in Fig. 2b, that there are ηι_λ matter segments 1 1 along the mean light pathway, each with a different attenuation coefficients atx (for illustrative purposes, three segments are depicted in Fig. 2b), then we can write the following equation:

ι

In a similar manner to the conventional Beer-Lambert law, the attenuation coefficient μ_{3ΐ 5ί} χ can be computed as follows:

(3)

^at Sj ⁼ ^ ^ak ^{' c}sj k

k = l where the absorptivity coefficient a_{k ¾} has the same meaning as in absorption spectroscopy— namely, a function of wavelength λ, unique for each chromophore k present in the segment j of the sample; and c_{s k} is the concentration of chromophore k within the segment j (shown in Fig. 2b).

Equations (1 ) to (3) above form the basis for the DPS theory that we have devised.

Referring now to Fig. 3a, there is depicted a single mean pathway 13 (i.e. one of the pathways shown in Fig. 2a) for light propagating through vascular tissue 15. I_E\ is a light intensity at the emitter side and I_{tD ¾} is a light intensity at the detector side. Fig. 3b is a schematic representation of the PPG signal obtained at the detector in Fig. 3a.

When arterial blood refills arteries and capillaries within the vascular tissue during diastole, the mean light pathway expands causing more light to get absorbed by the incoming layer of blood. Conversely, when blood leaves the capillaries during systole the mean light pathway contracts and less light is absorbed. Thus, Ι_{ο λ} changes with blood inflow and outflow, and if Ι_{Ε λ} does not change between time t₀ and t (a reasonable assumption), then the following equation can be written for the absorbance (1 ) variation as blood moves through the tissue:

(4)

AA t₀, t) = A t) - A {t₀) = - log

If we let μ_3ΐΑΒ λΟ0 be the absorption coefficient for the arterial blood as in (3); B XOQ is a total mean light pathway within all arterial blood segments as in Fig. 2b; and Ατχ(ί) is the total absorbance within non-arterial blood segments as in (2); then:

Α_λ (t) = _3ΐ AB λ (0 · ^AB λ (0 + _λ (t) (5)

Two assumptions can be made about (5): μ_3ΐΑΒ λΟ0 does not change significantly between t₀ and t (i.e. μ_3ΐΑΒ λΟ0 = μ₃ΐΑΒ 5ά₃θ^;ο)); and A_{T ¾}(t) is not altered to a significant extent by the blood pulsation (i.e. A_{T ¾}(t) = A_{T ¾}(t₀)). From these assumptions and equations (4) & (5), the following equation can be written:

ΔΑ_λ(ί₀, t) = μ_{ΛΑΒ λ}(ί) ^■ ΔΙ_{ΑΒ λ}(ΐ₀, t) (6) where Al_{AB ¾}(t₀, t) is a Mean Light Pathway Variation (MLPV), which represents a thin layer of arterial blood responsible for the change in light absorption between systole and diastole heart cycles:

ΔίΑΒ λ(£θΌ = ^ΖΑΒ λ( - *ΑΒ λ(£θ) (7)

The light intensity (/_D, Fig. 3a) used in BLNLP and the light power (P_D, Fig. 3b) sensed via a photodetector are two different light energy characteristics. By definition, detected light power can be computed by integrating light intensity over the photodetector's surface and over all mean light pathways leading to the photodetector (Fig. 2a). In order to be able to use light power instead of light intensity in (4) as is done in PO, one more assumption has to be made: namely that AZ_{AB /l}(t₀, t) (the Mean Light Pathway Variation) is not significantly different from one light pathway to another. With this assumption, (4) can be transformed as follows:

^^(T°^{' T) =} - '°¾ P»

Combining (6) with (3) produces the following set of linear equations, which in our DPS theory we have named "DPS equations":

; = [i -n_AB] _(9c)

Since in this set MLPV Al_¾. (t₀, t) is introduced as a physical quantity with the meaning illustrated in fig: 1 , rather than phenomenologically, it is possibile to solve these equations for the blood constituent concentrations c_{AB c}. (t). Indeed, the absorbance variation ΔΑ_λ. (ί₀, t) can be computed directly from the PPG signals using (8), and the absorptivities α_{ΑΒ ε}. _λ. can be obtained individually for each blood component using conventional spectroscopic techniques. Measurement of the MLPV can be accomplished by a variety of different techniques, some of which will later be described in detail. What is immediately apparent, however, that there is a clear relationship between light pulse propagation delay and MLPV, and hence the peak of the wave PD A( iⁿ Fig. 1 b should move left and right on the time scale with each heartbeat (in effect, the PPG signal is modulated by the heartbeat). The time variation between P_E A( spike and PD A( peak should be proportional to the MLPV with a proportionality coefficient equal to the speed of light in arterial blood. Thus it is possible to measure the MLPV with at least the light pulse propagation delay experiment.

Indexes j in equations (9a) (9b) and (9c) identify the blood components taken into account. The smaller the concentration of a given component and the smaller its absorptivity, the higher the PPG/MLPV system resolution that will be required to evaluate the concentration of that component. The component with the smallest absorption coefficient (3) defines the total number of components which have to be included into equations n_AB, as concentration variations of the components with the higher absorption coefficients will affect the evaluation accuracy. Component absorptivities have to be different by at least one wavelength to avoid the equations being linearly dependent. The number of wavelengths at which PPGs should be monitored must be equal to or greater than the number of blood components, whose concentrations are taken into DPS equations.

In a similar manner to PO, another assumption can be made on MLPV equality between different light wavelengths:

^AB A_j = ^AB j = ^AB (10)

From (8) and (9), the following ratio can be introduced:

_ ΔΑ_λ. (ΐ₀, ΐ)

The proportion taken by a blood component in respect to all the components present in blood is defined as a molar fraction:

A R r^■ 11

F_{AB Ci}(t) = (12)

f_{e = 1} C_{AB Ck} {I)

Substituting (10) and (12) into (8) and rearranging it produces the following equation:

ΑΑ_λ. (ί₀, t) = Al_AB (t₀, t) · c_{AB Ck} (t)

ΠΑΒ

^^AB Aj · F_AB (13)

1= 1

Substituting (13) into (1 1 ) results in the following:

Since the sum of all molar frations MF_{AB c}. as defined in (12) is equal to one, then the following equation can be stated:

TlAB -l

MF A, B c- = 1 MF_{AB c}. (t)

nAB

i = l (15)

Combining (15) and (14) produces the following set of equations: (16a) π-ΑΒ-1

V_CnAB A, (t) = AB_CnA j · ( ¹ - MF_{AB Ci}(t) I _(16b) i=l vc_nAB x_nAB {t) + v_AB-_{l XnAB} {t) (16c)

; = [l ... (n^ - l)] (16d)

This set of equations is a part of our DPS theory and has been named Static Pulsatile Spectroscopy (SPS) equations. The number of unknown molar fractions MF_{AB c}. (t) in these equations is equal to the number of equations and they are solvable. Unlike DPS, SPS equations do not require extra parameters apart from the PPG signals.

Logically, SPS is an expansion of the PO technique, while DPS is an attempt to identify what additional parameters (i.e. MLPV) should be measured is order to extend this technique for the evaluation of arterial blood analytes concentrations. It is easy to see that Sp0₂ is the molar fraction as defined in (12) and the theoretical relation of Sp0₂ and R values in PO is a special case of the more general SPS equations. To prove this, we can assume that blood analytes consist of oxygenated and deoxygenated haemoglobins. Thus, with two components, two light wavelengths are required. Absorptivity coefficients for both types of haemoglobin can be obtained using conventional spectroscopy. Substituting those absorptivities into (15) and solving it in respect to MF(t) produces an equation which matches exactly the theoretical Sp0₂ equation used in PO.

In general terms, our theory has established that it is possible to measure the concentration of components in a fluid if one knows the absorptivities of the individual components, and the amount of fluid. Considering a pulsatile fluid flow, for example through vascular tissue (as one illustrative example), the concentration of fluid constituents can be determined using the above equations if the mean light pathway variation attributable to the pulsed flow can be measured. Our technique is based upon the general concept that the amount the amount of light energy that is absorbed along a given pathway is the sum of the absorption that occurs over each segment of the pathway, and the absorption for each segment is the sum of the absorption due to the light-absorbing components in that segment (where the absorption for a given light- absorbing component equals the absorptivity for that component multiplied by its concentration).

In one illustrative implementation of the teachings of the invention, we have determined the mean light pathway variation by shining laser light through vascular tissue and measuring frequency shifts in that laser light that are attributable to the Doppler effect (a technique that we refer to as "Laser Wave Doppler").

Use of the Doppler effect to measure flow rates has previously been proposed, for example in a technique commonly known as laser Doppler flowmetry (LDF). In systems that embody this technique, a laser beam penetrates the tissue of an individual and photons in the beam are scattered by blood cells flowing through the tissue in a generally transverse direction relative to the laser beam. The scattered photons are detected by a detector, and have a frequency that is Doppler-shifted (relative to the frequency of the incident laser light) proportionally to the speed at which the blood is flowing through the tissue.

The technique that we have developed differs from this LDF technique in that we use the Doppler effect to measure the speed at which vascular tissue (in this particular example) expands and contracts as blood pulses through the tissue. This expansion and contraction of the tissue causes a corresponding expansion and contraction of a laser beam travelling along a pathway (the Mean Light Pathway (MLP)) through the tissue. The expansion and contraction of the MLP gives rise to a Mean Light Pathway Variation (MLPV), and the speed with which the MLP varies (the Mean Light Pathway Variation Speed (MLPVS)) causes the electromagnetic wave of the laser light beam to change its frequency due to the Doppler effect. By determining the Doppler frequency shift of the laser light beam electromagnetic wave we can evaluate the MLPVS, and by integrating the MLPVS over time we can determine the MLPV. The determined MLPV variation can then be substituted in the abovementioned DPS equations, and the concentration of components of the blood can be calculated. As the Laser Wave Doppler frequency resulting from variations in the Mean Light Pathway is much lower than the Doppler Flow frequency, because blood particles move much quicker than the tissue, and hence the laser wave Doppler frequency can be separated from the Doppler flow frequency by a band pass filter.

It will be apparent to persons of ordinary skill in the art that techniques other than the aforementioned Laser Wave Doppler technique could be employed to determine the mean light pathway variation.

For example, in one envisaged implementation we track the "time of flight" of photons in the laser beam to non-invasively measure the Mean Light Pathway Variation.

Referring again to Fig. 1 b, as the vascular tissue expands and contracts, so the length of the Mean Light Pathway will change and the peak Ρ_ολ (Fig. 1 b) will move left and right as the heart of the subject beats. The time variation between the Ρ_Ελ spike and the Ρ_ολ peak is proportional to the MLPV with a proportionality coefficient that is equal to the speed of light in arterial blood, and hence the MLPV can be determined. Whilst this "direct" measurement is possible, in practice the equipment required is bulky and expensive and the measurement itself is complex, and hence the aforementioned Laser Wave Doppler technique is preferred.

In another implementation we can avoid some of the complexities of the Photon

Time of Flight Tracking technique described above by considering the Fourier Spectra of Light Photon Density Waves. If we assume that light-tissue interaction is linear then the spike Ρ_Ελ of light in Fig. 1 b can be split into Fourier spectra. In other words, the light intensity can be modulated harmonically (i.e. the emitter light intensity changes in accordance with a sine function), and the intensity-modulated light beam is known as the light photon density wave. By measuring the response to each sine frequency in a wide range of frequencies, and subsequently combining those responses using reverse Fourier transform we can generate an approximation to the spike response as in the Photon Time of Flight Tracking measurement, and thereby determine the MLPV.

In another implementation, when the light beam intensity is harmonically modulated the resulting wave is known as the Light Photon Density Wave. When such a wave is traversing vascular tissue, the frequency of the intensity modulation changes with the subject's heartbeat due to the Doppler effect. This frequency change can be detected using high radio-frequency technique, and the Mean Light Pathway variation can be calculated.

In another implementation, called Optical Diffusion Tomography modeling, it is possible to model the response of the tissue using a technique similar to that which is described in European Patent No. 1959819.

Referring now to Fig. 4 of the drawings, there is a depicted an illustrative representation of apparatus 17 for determining the concentration of a component in a fluid. This apparatus utilises the aforementioned Laser Wave Doppler technique to determine the MLPV.

The apparatus 17 comprises a light emitter 19 that is coupled by means of an optic fibre to a fibre splitter 21 which splits the incident light from the emitter 19 into a first beam that is directed via an optic fibre to vascular tissue 25 (that comprises, in this illustrative example, a subject's finger), and a second reference beam that is directed to a laser wave Doppler module 23 by an optic fibre. Part of the light of the first beam from the splitter 21 travels through the vascular tissue (the remainder being absorbed) to a fibre splitter 27 that is coupled by respective optic fibres to a light detector 29 and the aforementioned laser wave Doppler module 23. A control and data processing system 31 , for example a computing resource such as a computer, receives data from both the light detector 27 and the laser wave Doppler module 23. In one envisaged implementation the light emitter comprises a multi-wavelength monochromatic source of light, such as a plurality of light emitting diodes. In another implementation the light emitter 19 could comprise, in an illustrative implementation, a single-wavelength laser light source (for example, a 100 mW, 980 nm fibre laser), although it is envisaged that the laser could be tunable to enhance the measurement of the mean light pathway variation. The laser Doppler module comprises a band pass filter (to filter low frequency amplitude modulation and high frequency noise), and logic for determining - in the manner aforementioned - the mean light pathway variation speed. Light detector 29 comprises, in a preferred implementation, a multichannel photodetector.

Light from fibre splitter 21 is modulated by the pulsating vascular tissue and the resulting modulated laser light is split by fibre splitter 27 and passed to the detector 29 and the laser wave Doppler module 23. The light detector outputs data representing the intensity of light received, and from this information the control and data processing system can compute the amount of light absorbed as the light traverses the vascular tissue. The laser wave Doppler module 23 filters out frequency shifts attributable to the speed of the blood flowing through the tissue, leaving the frequency shifts attributable to the speed at which the vascular tissue expands and contracts. The Laser Wave Doppler module computes the mean light pathway variation speed and outputs this data to the control and data processing system for processing. The control and data processing system calculates, from the mean light pathway variation speed, the mean light pathway variation. Given the observed absorbance, the calculated mean light pathway variation and data concerning the absorptivity of a given blood constituent (which data may be preprogrammed into the control and data processing system or retrieved from a data store), the control and data processing system can calculate the concentration of that constituent in the blood using the DPS equations set out above.

Thus the system depicted schematically in Fig. 4 can be operated to noninvasive^ determine the concentration of absorbers (such as glucose, for example) in the blood of a subject.

A possible implementation of the apparatus 17 is apparatus 33 (which can be used to non-invasively determine the concentration of absorbers in a fluid) as shown in Fig. 5. In this instance the laser wave Doppler module 23 of Fig 4 comprises an acousto-optic modulator module 35 that is configured to shift the frequency of the reference laser beam output from the first fibre splitter 21 by a small amount (typically a few kilohertz), a light detector 41 , an attenuator 37 that is configured to attenuate the reference beam so that it does not saturate the detector, and a fibre mixer 39 that is configured to generate interference fringes from the light from the second fibre splitter (i.e. light that has travelled through the sample 25) and the attenuated reference beam. These fringes provide a frequency signal that is proportional to the Doppler shift imparted by the pulsating sample, but frequency shifted (by the acousto-optic module) so that it does not overly the signal representative of intensity variation caused by the pulsatile flow (i.e. the output of fibre splitter 27).

As it is relatively difficult to shift a laser beam by a few kilohertz, in one particularly preferred arrangement the acousto-optic module comprises a first acousto- optic modulator that is configured to shift the frequency (for example, increase) of the reference beam by an amount X, and a second acousto-optic modulator that is configured to shift the frequency of the reference beam in an opposite direction (for example, reduce) by the same amount X. In theory the beam output by the acousto- optic module should have the same frequency as the reference beam, but in practice as the two modulators are not exactly identical, the beam has a slightly different frequency from the reference beam.

Whilst the foregoing arrangements represent a significant achievement and have great utility, the scope of the present invention is not limited to in-vivo concentration measurement. In general terms, the techniques described herein are of use wherever one can introduce a mean light pathway variation between two beams of laser light travelling through a sample. For example, in one envisaged implementation the teachings of the invention could be implemented with a cuvette that includes an internal step so that the cuvette has a region with a first diameter and a region with a second larger diameter. Light beams shone through each of the two regions would have a known mean light pathway variation, and by observing the absorbance of each beam one it is possible to calculate the concentration of given components using the observed absorbance, the known mean light pathway variation and the absorptivity for the component of interest. A conical flask could be employed as an alternative to a stepped cuvette.

It will be appreciated that whilst various aspects and embodiments of the present invention have heretofore been described, the scope of the present invention is not limited to the particular arrangements set out herein and instead extends to encompass all arrangements, and modifications and alterations thereto, which fall within the spirit and scope of the invention.

It should also be noted that whilst particular combinations of features have been set out herein, the scope of the present invention is not limited to these particular combinations and instead extends to encompass any combination of features herein disclosed.

Claims

1 . A method for determining the concentration of a substance in a fluid, the method comprising:

(i) operating a light emitter to illuminate a fluid;

(ii) operating a light detector to determine the intensity of light that has passed through the fluid to the detector;

(iii) calculating an amount of light absorbed by the fluid;

(iv) determining a mean light pathway through the fluid from the detector to the emitter; (v) retrieving data concerning the optical absorptivity of the substance; and

(vi) determining from the calculated absorbance, the determined mean light pathway and the retrieved absorptivity of the substance, a measure of the concentration of the substance in the fluid.

2. A method according to Claim 1 , where step (iv) comprises determining a variation in mean light pathway, and step (vi) comprises determining, from the calculated absorbance, said variation in mean light pathway and the retrieved absorptivity of the substance, a measure of the concentration of the substance in the fluid.

3. A method according to Claim 2, wherein said light emitter is operated to illuminate a pulsatile flow of fluid through a pulsating sample.

4. A method according to Claim 3, wherein the amount of light absorbed by said fluid is calculated directly from the intensity of light detected at said detector as said fluid flows through said pulsating sample.

5. A method according to Claim 4, wherein the calculation of the amount of light absorbed provides a measure of light absorption for each pulse of the pulsating sample.

6. A method according to Claim 5, wherein said step (vi) comprises:

determining the speed at which the mean light pathway through said pulsating sample varies for each pulse of the pulsating sample; and

calculating from said determined speed the variation in mean light pathway for each pulse of said pulsating sample.

7. A method according to Claim 6, wherein the speed of mean light pathway variation for each pulse of said pulsating sample is determined from a Doppler frequency shift induced in the light as it passes through said pulsatile flow of fluid.

8. A method according to Claim 7, wherein said Doppler frequency shift is obtained by interfering light that has travelled through said pulsating sample with a reference beam of light that has not travelled through said pulsating sample.

9. A method according to Claim 1 or 2, wherein step (iv) comprises determining a light propagation delay that occurs as light travels through said fluid, and calculating from said delay a mean light pathway through said fluid.

10. A method according to Claim 3, further comprising amplitude modulating light from said light emitter with a high frequency modulation signal.

1 1 . A method according to Claim 10, wherein the speed of mean light pathway variation for each pulse of said pulsating sample is determined from a Doppler frequency shift induced in the high frequency modulation signal as said light passes through said pulsatile flow of fluid.

12. A method according to any of Claims 3 to 1 1 , wherein said pulsatile flow of liquid is in-vivo and said sample comprises a body part.

13. A method according to Claim 1 or Claim 2, wherein the mean light pathway is determined using Fourier spectra of light photon density waves.

14. A method according to Claim 3, wherein the mean light pathway variation is determined by imaging said pulsating sample, for example by means of optical tomography, and determining, from a plurality of images of said sample as said sample pulsates, a measure of the variation in mean light pathway.

15. A method according to Claim 1 , wherein said fluid is provided in a container configured such that a first beam of light travels through the container along a first mean light pathway, and a second beam of light travels through the container along a second mean light pathway different to said first.

16. A method according to Claim 15, wherein said container comprises a cuvette with an internal step or a conical flask.

17. Apparatus for determining the concentration of a substance in a fluid, the apparatus comprising:

a light emitter for illuminating a fluid;

a light detector for determining the intensity of light that has passed through the fluid to the detector;

means for calculating an amount of light absorbed by the fluid;

means for determining a variation in mean light pathways through the fluid from the detector to the emitter;

means for retrieving data concerning the absorptivity of the substance; and

means for determining from the calculated absorbance, the determined mean light pathway variation and the retrieved absorptivity of the substance, a measure of the concentration of the substance in the fluid.

18. Apparatus according to Claim 18, wherein said light source comprises a laser.

19. Apparatus according to any of Claim 17, wherein said emitter comprises a multi- wavelength monochromatic light source, for example a plurality of LEDs.