Optical Monitoring
This invention is concerned with improvements to methods used for non-invasive monitoring of patients using electromagnetic energy.
It is desirable to obtain indications of the concentrations of certain chemicals in various parts of the body in order to detect abnormal and dangerous conditions.
Examples of chemicals of interest include those which are linked to the transport or utilisation of oxygen, such as oxy and de-oxy haemoglobin in blood, oxy and de-oxy myoglobin in muscle or oxidised and reduced intra-cellular enzymes such as cytochrome aa3. Other chemicals of interest include glucose in blood or tissue or cholesterol in blood.
The purpose of deterrnining such chemical concentrations is to allow correction of abnormal and dangerous conditions and it is therefore desirable to carry out repeated measurements or continuous rel-time measurements. This is particularly desirable when measuring those chemicals linked to transport or utilisation of oxygen because the concentrations of these are known to be liable to change very rapidly; significant changes can occur in 5-10 seconds.
In order for abnormal and dangerous conditions to be determined it is also necessary for the measurement of chemical concentration to be made in a quantitative way so that comparison with accepted normal values might be made.
It is known that electromagnetic energy can pass through tissue and that when wavelengths in the near infra-red region of the spectrum are used then deeper penetration of tissues is possible as compared to that achieved with visible wavelengths.
It is also known that by interrogating tissue with two or more specific wavelengths of
electromagnetic energy, followed by mathematical analysis of the intensity of the energy emerging from passage through the tissue, ftfis possible to deteirnine estimates of the changes in concentration of one or more chemical species in the tissue being interrogated. In order to perform the calculation use is made of the Beer-Lambert law. Here it is given for a single absorber and a single wavelength of measurement.
a
A, = log - = β, Cxdp ζ + Ψ Equation 1
17
in which Aj is the attenuation of the electromagnetic wave measured in optical density units (OD), &0 is the incident wave intensity, & is the intensity of the wave emerging from the tissue, et is the specific extinction coefficient of the absorbing compound in μmolarJcm'1 , Ct is the concentration of the absorbing compound and dp is the distance in cm between the points where the wave enters and emerges from the tissue. Due to scattering of the electromagnetic wave in tissue its path will be longer than the physical spacing by a factor ζ, known as the scattering factor.
Attenuation of the wave also occurs due to scattering and an additive term, -P,is used to describe this. It is useful to consider the product e.C as the absorption coefficient, μτ
Problems arise with this method of monitoring chemical variables in patients because it is not possible to deteraiine absolute chemical concentration. This is partly because the scattering factor, ζ, is not known and partly because the additive term, Ψ, is not known. Instead of calculating absolute chemical concentration, therefore, it is now common practice to measure changes in attenuation from an arbitrary starting point and from such changes in attenuation to calculate changes in chemical concentration.
Although in some circumstances this current method can provide useful information for the monitoring of patients it is a disadvantage that it is not possible to determine
absolute quantitative values of the chemical concentrations. This means that measurements made in one subject can not be compared with those made in another subject and measurements made in one subject on one occasion can not be compared with measurements made in that subject on a second or subsequent occasion.
The absolute chemical concentrations can be determined if certain effects of scattering can be measured and if other certain effects of scattering can be modelled mathematically and if these steps can be performed rapidly and cost-effectively. This approach is the essence of the present invention.
As indicated in equation 1, one effect of scattering in tissues is that the length of the path travelled by the interrogating wave is greater than the physical distance, dp . In fact the so-called optical pathlength, dσ is greater than dp by something like 3 to 5 times depending on the tissue type and the wavelength of light.
This effect of scattering could be taken into account if da could be measured. It is known that there are at least two methods to measure optical path length.
Firstly, the time taken for a short pulse of electromagnetic energy at the appropriate wavelength to travel through the tissue can be measured and distance then calculated as:
d0 = c.t Equation 2)
where c is the velocity of light in the tissue and t is the time taken for passage of the short pulse through the tissue. Due to the high speed of light the value of t is very short for tissue path lengths of a few cms and so there are practical problems in its measurement. It is known that very short pulses of infra-red light, in the order of a
few psecs, can be produced by means of lasers and it is also known that such short pulses can be detected by a streak camera and these devices may be used together to measure optical path length in human tissues. However, this apparatus is large, costly, and can not be used in routine clinical circumstances.
A second method to measure optical pathlength in tissue is based on the modulation, for example with a sinusoidal function, of the intensity of the light beam passed into the tissue. In this case the intensity of the light emerging from the tissue will also vary sinusoidally but the phase of this intensity variation will lag that of the incident beam by an amount that is proportional to the optical pathlength. Thus measurement of the phase shift between the incident and detected beams will allow optical pathlength to be calculated.
Complications arise in practice due to the multiplicity of scattering events in tissue. This situation leads to a multiplicity of optical paths, each producing a corresponding phase-shift. The resulting received optical intensity therefore contains a summation of phase-shifted components representing the group of scattered paths. In order to derive a useful estimate of optical path length, for example the mean, the received wave intensity must be analysed rapidly in order that the estimate may be used to provide continuous real-time calculations of chemical concentration.
The phase-shift method can be implemented easily, and at comparatively low cost, when the object under investigation contains relatively few scattering discontinuities. Under these circumstances there is a small number of scattering paths, and the summated signal at the receiver may be analysed using high frequency phase sensitive circuits. However, tissue of interest such as the brain, limbs, Uver, kidneys, muscle, bone all scatter significantly and in thick sections there will be substantial attenuation of the incident wave as well as multiple-path mixing contained in the received signal.
According to one aspect of the present invention, there is provided a method of
deteirr-ining the concentration of a selected chemical or biological species in vivo, in which electromagnetic radiation is directed through the tissue and emergent radiation is analysed, the frequency of the radiation being selected so as to correspond to a known absorption frequency for the species of interest and the radiation applied to the tissue being intensity modulated in accordance with a predetermined mathematical function, characterised in that:
(i) emergent radiation is received by detectors positioned in at least three locations, each of which has a different linear separation from the radiation input;
(ϋ) emergent radiation detected by said detectors is subjected to signal processing to determine, in real time, (a) the mean optical path lengths for each detector at each input wavelength; and (b) the attenuation of the input radiation at each detector and for each input wavelength; and
(iii) the concentration of the species of interest in the tissue is calculated using known specific extinction coefficients for the species of interest at the wavelengths used and the mean optical path length and attentuation data derived from the input and emergent radiation by the signal processing step.
With preferred embodiments of the present invention there is provided means whereby an electromagnetic wave is modulated before being passed into tissue and the corresponding signals collected at three or more points after passage into and out of tissue segments containing phase information related to group optical pathlength properties are processed quickly in real-time in order to recover optical path length estimates for use in the calculation of absolute chemical concentration. There is further provided means with which phase-shifted modulation components of light propagated through tissue are analysed quickly in real-time in order to allow the attenuation due to scattering phenomena as specified by the additive term, Ψ, in equation 1 to be derived.
Subsequent combination of the two components of the information derived by the two aspects of the invention then allows calculation of absolute concentrations of
specific chemicals within the tissue under interrogation having value for medical diagnosis or therapy.
According to a second aspect of the present invention, there is provided pparatus for use in determining the concentration of a selected chemical or biological species in vivo, which comprises:
(a) a plurality of laser diodes;
(b) a modulation circuit arranged to modulate the outputs of said laser diodes at two or more modulation frequencies;
(c) a frequency selection circuit arranged select said modulation frequencies;
(d) means for inputting the modulated laser radiation into human or animal tissue which is to be examined;
(e) at least three detectors for attachment to said human or animal to receive radiation emergent from said tissue; and
(f) signal processing means for processing data derived from said detectors, said modulation circuit and said frequency selection circuit.
One embodiment of the present invention is now described purely for illustrative purposes with reference to the accompanying drawings, in which:
Figure 1 illustrates the passage of an optical beam through animal tissue; and
Figure 2 is a schematic illustration of apparatus in accordance with this invention.
Referring to the drawings, Figure 1 shows a beam of electromagnetic energy [1] that has been modulated with a specific function. It will be known to those skilled in this branch of science that such energy may be considered either as a wave or as particle-like photons. The energy passes through tissue [2] that has both scattering and absorbing properties. Specific optical absorption may take place by particular
chemical constituents in the tissue and of interest in the care of patients is the optical absorption due to the chemicals oxy-haemoglobin and deoxy-haemoglobin and the concentration of each of these chemicals individually or as some form of scaled ratio is required. There is also interest in the specific optical absorption by the intra-cellular enzyme cytochrome oxidase which may exist in the oxidised or reduced form and is known to have particular optical absorption bands in the near infra-red part of the electromagnetic spectrum for each of these two states.
The energy emerging after propagation through the tissue [3] will have been transformed by the transfer function of the tissues, including absorption by haemoglobin in its two forms and cytochrome oxidase in its oxidised and reduced state and by scattering events. The transfer function of the tissues is given by the output/input ratio, « l_90ι
Extraction and analysis of the transfer function is achieved in the present invention and this is then used in the calculation of the absorption coefficient, μn and so- called reduced scattering coefficient, μs of the tissue.
Figure 2 provides a schematic description of an instrument constructed in accordance with the present invention. In order to calculate the absolute concentrations of haemoglobin in its two forms the electromagnetic energy transmitted into the tissue consists of time-multiplexed beams of two or more wavelengths, λn, between 700 nm and 900 nm. In this particular embodiment of the present invention this is achieved by switching laser diodes [6] on and off in sequence and combining the laser diode outputs and conveying the combined energy by means of an optical fibre bundle [7] and an attachment probe [11].
During the laser ON period the intensity of the laser-generated light is modulated by a function, fm, by means of a modulation circuit [8]. The modulation function may be sinusoidal having a frequency of vm. Two or more modulating frequencies are used, vm„ vm2, etc. The actual values for the present clinical applications are in the range 50 MHz to 500 MHz and are determined by a frequency selection circuit [9].
Electromagnetic energy propagates through the tissue of interest [10] according to the structure and composition of the tissue. The energy will pass through different regions of the tissue and examples of particular paths are shown as 17a, 17b and 17c . Having propagated into the tissue of interest the energy is collected at three or more points by means of attachments [12a, 12b, 12J, which may be connected to the signal processing elements of the system by spatially separate optical fibres; alternatively, the individual optical fibres may be bundled together for convenience.
The spacing between the input point, [11], and the collection points, [12a, 12b, 12c], has some significance. The corresponding physical pathlengths, (dj0 (dj b and (d^ are chosen such that the there is sufficient penetration of the deeper tissue regions by electromagnetic wave paths 17., 17b and 17c. In human tissues this means that (d)a should be greater than 2.8 cms., (d^ should be greater than (d^aand (dp)e should be greater than (dj)t.
In a preferred arrangement the collected energy is transported by individual optical fibres such as [13] to a photomultiplier where it is detected and amplified [14]. An alternative arrangement may use an array of detector sensors fixed directly to the tissue surface.
With the preferred arrangement nml, vm2, etc. With this arrangement the signal output of the photomultiplier consists of the low frequency spectrum of 10kHz to 15kHz rather than the comparatively high frequency modulating frequency of 50 to
500 MHz and is therefore more straightforward to process. The detected optical signal contains phase shifts related to optical pathlength and intensity information in the form of a depth of modulation related to attenuation.
Whilst the heterodyne method is the preferred approach the present invention may also be realised using well-known homodyne detection.
Mean optical path length at each wavelength and for each detector position is calculated within a control and processing unit [15] by deteπnining the phase shift of the modulating signal. Firstly, a single modulating frequency is used and optical path
length and modulation depths are determined for each detector position. Then, for a single detector position path length and modulation depth are determined for each of the two or more frequencies of modulation.
Phase shifts due to multiple optical path lengths are determined by performing a Fast Fourier Transform (FFT) within the control and processing unit [15] on the spectrum of signals available at the output of the detector unit [14]. An important aspect of the particular embodiment of the invention which facilitates accurate phase measurement is the arrangement of the control and processing unit and the specific instrument modules (6, 8, 9 and 14) which includes the use of a data and control bus.
The analysis of the phase-shift information from the FFT provides statistical data relating to the scattering properties of the tissue under interrogation.
In order to calculate the concentrations of oxy-haemoglobin, {HbO2}, and de-oxy haemoglobin, {Hb}, use is made of the known approaches to electromagnetic wave transport based on the so-called diffusion approximation. This is modelled within the instrument and the approximation allows relationships to be derived between, firstly, phase shift and vm, μ,', μ„ and dp at each wavelength, λ„, of the interrogating electromagnetic wave and, secondly, the modulation depth and vm, μ,', μm and dp at each wavelength, λn. The specific measurements made by the instrument designed according to the present invention are used with the diffusion approximation in order to calculate the two unknowns, μ,' and μω for each wavelength, λ„. The wavelength dependent values of the extinction coefficients for HbO2 and Hb are known. Use of these together with the calculated values of ,'and μ„, for each wavelength, λ„, then allows the absolute concentrations of {HbO2} and {Hb} to be calculated.
For clinical convenience the measurements of {HbO2} and {Hb} are used to derive the ratio {HbO2}/[{Hb}+{HbO2}] which is then expressed as a percentage by multiplication by 100 in order to produce a measurement of absolute oxygen saturation. The instrument described here as representing one embodiment of the present invention incorporates this feature.
The continuous real-time monitoring of oxy-haemoglobin, {HbO2}, and de-oxy haemoglobin, {Hb}, with the present invention also allows automatic correction of abnormal and dangerous levels to be achieved. For this purpose, the output of the instrument disclosed in Fig 2 is used as an input to a process control loop, the output of which is used to adjust the concentration of oxygen in the gas supply to the patient.
Description of Operation of the Instrument
In order to perform measurements on a subject using the instrument firstly probes 11 and [12a , 12b , 12c] must be affixed with appropriate spacing between the input point and the collection points. For example, for examination of the brain (dp)a should typically be 3 cm. In this case (d^ will then be 3.8 cm and (d^ will be 4.6 cm.
Attachment probes 11 and [12a, 12b , 12c] should also lie in a straight line.
According to the present invention the instrument will then perform a sequence of operations in order to derive quantitative values for chemical concentrations. If we consider here the measurement of O2Hb, HHb and oxidised cytochrome aa3 three optical wavelengths, typically 760nm, 840nm and 905nm, will be required. The currently preferred sequence of operation will be:
1. Three laser diodes [6] will be switched on and off sequentially at a repetition rate of typically 1 kHz.
2. The laser ON period is divided into two parts. During the first part of the laser ON period a modulating signal of vml is applied to the laser. The frequency of vml may be typically 100 MHz. During the second part of the laser ON period a second modulating signal, vm2, is applied to the laser. The frequency of vm2 can be 140 MHz.
3. Energy collected at [ 12a , 12b , 12c] during the laser ON period is fed either to a multiplexed detector, typically a photomultiplier type R6357; or to three parallel photomultipliers; or to three silicon detectors. The photomultiplier(s) will preferably have dynode drive at a frequency offset firstly from vml and, secondly, from vm2 by an amount which is typically 10 kHz and will thus produce an output by the heterodyne principle. If silicon detectors are used then the outputs will be mixed with a reference signal derived from vml and vm2 in order to produce the desired output.
4. Steps 2 and 3 are repeated for each of the laser diodes in sequence.
5. The output values produced in step 3 are used to calculate the transfer function of the interrogated tissue at each wavelength. This then allows expressions for scattering factor, ζ and additive term, Ψ, to be determined. Application of the Beer-Lambert law then allows μa and μ to be deteπnined..
6. Based on a second aspect of the invention the output values produced in steps 3 and 4 for each collector position are used to calculate the slope values S^,
Sac and Sphase for the DC, AC and phase expressions derived from the Diffusion Approximation of the transport equation. (For description of the Diffusion Approximation see, for example, A Ishimaru "Diffusion of light in turbid material", Applied Optics, vol 28: 2210-2215, 1989; the content of this document is incorporated herein reference thereto). The instrument then calculates values for μa and μs' from the slope expressions.
7. The values for μa and μs ' derived either as in 5 or in 6 then allow the concentrations of O2Hb, HHb and oxidised cytochrome aaj to be calculated.
8. Selection of different wavelengths for the interrogating energy allows deteπnination of the concentrations of other chemical species provided their molar extinction coefficients are known and are spectrally unique.
To achieve "proof of principle" certain practical experiments have been conducted as follows.
The signal processing and analysis strategies were assessed using simulated signals. Signals representing the simulated phase shift signals were generated in order to test the real-time processing arrangement for extracting the tissue transfer function based on Fast Fourier Transform (FFT). A particular interest exists in deteiTri-triing the absorption and reduced scattering coefficients of the tissue ex-trnined due to {HbO2} and {Hb} and also due to the oxidised component of the respiratory enzyme cytochrome aa3. Therefore, for test purposes, three operating wavelengths were required in order to derive three equations that could be solved for the three unknown quantities. These were selected to be 760 nm, 840 nm 905 nm simply for the purpose of a test.
Two alternative implementations of the present invention have been tested.
In the first implementation, the extraction of the scattering factor, ζ. and the so-called additive term, Ψ, from equation 1 was used. Both of these, being related to scattering phenomena, appear as unknowns. For the purpose of this proof of principle embodied with the present invention these parameters were determined by measuring the tissue transfer function at two modulating frequencies, vml and v^ The resulting set of two simultaneous equations was then solved to yield expressions for ζ and Ψ at each interrogating wavelength. These expressions were then used with equation 1 to determine values for the absorption coefficient, μm and the reduced scattering coefficient, μ ', of the tissue at each of the three wavelengths.
In the second implementation, the use of the spatial variation of detected energy was employed. Signals simulating the modulation depth of the interrogating beam and the resulting attenuation seen in the detected beam were also fed into a digital signal processing module (DSP). The demodulation of the intensity modulated interrogating beam was then extracted successfully by an appropriate algorithm. This yielded values for a DC component and an AC component.
The extracted values of phase-shift, DC component and AC component simulated for three wavelengths and three transmit-receive distances were then used to derive values for the absorption coefficient, μ„, and reduced scattering coefficient, μ„, of the tissue.
A Diffusion approximation for photon propagation was used and both infinite and semi-mfinite boundary conditions considered in order to evaluate various configurations of input point [11] and collection points [12a , 12b , 12e ].
As transmit-receive distance is varied so too do DC and AC components and phase shift. A derivation from the diffusion approximation allows these relationships to be considered to be linear with slopes SΛ , S^ and Sphate. In order to extract the two unknowns, μa and .,', any combination of two equations from the three linear relationships may be solved.
With either of the two possible methods tested for deteπ-iining μa and /,', the concentrations of the absorbing species of interest, that is to {HbO2}, {Hb} and the oxidised component of the respiratory enzyme cytochrome aa3, were then calculated from these using known values of absorber specific extinction coefficients at each of the three interrogating wavelengths.
Practical measurements have been evaluated using certain modulated light sources could be used to generate the interrogating beam. It is feasible to use light emitting diodes LEDs (e.g. Hitachi type HLP 40RG) at three wavelengths,, 905 nm, 840 nm and 760 nm modulated at, for example, 48 MHz.. For extended penetration it is possible to use laser diodes e.g. Sony SLD104AU, modulated at, for example, 100 MHz. Peltier cooling of these devices is essential to achieve adequate overall signal to noise ratio.
The off-set frequency, chosen as an example to be 10 kHz., and the test phase-shifted signal were input to the DSP module where they were sampled. A conventional FFT algorithm was used to extract the phase information and this has shown successful recovery of the simulated phase-shift. Optical pathlength, which is equivalent to the product dpC,, is calculated as (phase-shift)c/2πv,--. Timing of analysis was programmed to test the possible use of up to four multiplexed intensity modulated light sources, emitting interrogating energy at wavelengths λl , λ2 , λ3 and λ4 selected from devices emitting at specific wavelengths between 700 nm and 1500 nm. DSP modules are available to enable, for example, 210 kHz sampling allowing more than four wavelengths to be used for determining the concentrations of more than four optically absorbing species.
In order to use the well-known heterodyne detection method the instrument has been tested with photomultipUers. Type R6357 may be used. Performance when modulated at the carrier plus offset, for example 100.001000 MHz, is adequate in terms of stabiUty and linearity. Alternatively it is possible to use semiconductor detectors, for example PIN diodes, and mix the received signal with an offset reference signal to recover the phase shift and intensity information.
The tests were carried out with optical phantoms having appropriate optical properties, of absorption and scattering. This can be produced from mixtures of milk and India ink.
Although the invention has been described with reference to laser sources, in particular laser diodes, it will be appreciated that the invention will operate with and includes within its scope other sources of electromagnetic radiation; one non-limiting example of such other sources is Ught-emitting diodes.