This invention relates to a non-invasive system for measurement of oxygen saturation in arterial blood.
When the oxygen saturation in arterial blood is measured using the principle of pulse oximetry, it is expressed as Sp02.
In all the known methods that have been proposed so tar, a part of the body is exposed to red light (R) of the visible spectrum and also by infra-red light (IR), some of which light (R + IR) enters into the body part. From within the body part such entry light (R + IR) exits the body part, the exiting light being detected to obtain photo plethysmo graph (PPG) signals.
The known methods extract the DC and AC (pulsatile)
4
portions of R and IR PPGs and compute Sp02 from*these values employing approximate empirical relationships based on ratios. The pulse oximeters used for the purpose require extensive calibration, the time and labour involved in which not only render the procedure tedious but also error-prone.

tfis, therefore, an object of this invention to do away with the aboyementioned drawbacks by proposing a method and device for the non-invasive measurement of oxygen saturation in arterial blood requiring little calibration for its use and operation.
The method for the non-invasive measurement of oxygen saturation in arterial blood, according to this invention, comprises the steps of exposing a part of the patient's body to R and 1R light; detection of these lights on exiting the body part by photo diode circuits; normalizing the R and IR PPGs; removing the noise in the R and IR PPGs. by filters, to obtain signals vR and vIR; obtaining the natural
logarithm of vR and vIR to get \n{vR) and ln(v//j), and the peak amplitudes of \n(vR ) and
\n(vIRX that is, VpR and VpIR;, obtaining the square roots of the R and IR PPG signals
v'ln(v^)) and N/ln(vTi?); determining the slopes of the linear portions of $n(yR)) and
,Jtn(vm) as UIR and mIR; and calculating the oxygen saturation

Normalizing R and IR PPG signals is carried out in the following manner. After signal leaving the body either by transmission or reflection, the signal picked up by the detector is lowpass filtered for removal of noise and other interference and then natural logarithm is effected on the signals in order to completely normalize it, that is, to make the signal free from the undesirable influence of the patient and circuit parameters. The device for the non-invasive measurement of oxygen saturation in arterial blood,
according to this invention, comprises means for exposing a part of the patient's body to R
and IR light; a photo diode circuit for detection of these lights on exiting the body part;

means for the normalization of the R and IR PPGs; filters for removing the noise in the R and IR PPGs to obtain signals vR and vIR; means for calculating the natural logarithm of vR and vIR to get ln(v^) and ln(v/^)? for calculating the peak to peak amplitudes of ln(v* ) and 1n(v/j?)fl that is, VpR and VpJR;; obtaining the square roots of the R and IR PPG signals yjk^yR)) and ^ln(vm); determining the slopes of the linear portions of yJbt(vR)) and yjhx(vm)
as mR and rn[R; and calculating the oxygen saturation

This invention will now be described with reference to the following Example and drawings, which illustrate, but which do not limit, the scope of this invention.

In the known art oxygen saturation (Sp02) measurement pulse oximeters from different manufacturers are available in the medical field. All these pulse oximeters work on the principle of photoplethysmography (PPG). A typical PPG signal has.an AC component at heart rate due to arterial blood pulsations and a DC component due to non-pulsatile skin, tissue and bone. The ratio of the pulsatile signals measured at red and IR wavelengths is related to oxygen saturation. In modern pulse oximeters, the method of Sp02 estimation relies on an empirical equation obtained by fitting a curve to the data obtained from a group of healthy volunteers and the
ratio of red and IR PPG signals. However, the PPGs are
■/ influenced by the patient-dependent parameters such as skin
pigmentation, volume of the tissuq and so on. Pulse
oximeters are, therefore, extensively calibrated to remove the
effect of patient- and sensor-dependent parameters on the
computation of Sp02. Each manufacturer has his own
calibrating factors depending upon the characteristics of the
group of volunteers participating in the calibration process.
This invention, however, arrives at the value of Sp02 without requiring extensive calibration. The R and IR PPGs are processed appropriately with a view to removing the source, detector and patient dependent parameters. SpC>2 is then computed utilizing the processed red and IR PPGs and the well-known extinction coefficients of hemoglobin and oxygenated he moglobin at the chosen red and IR wavelengths. Thus the procedures developed for estimation of Sp02 remove the effect of input intensity of the light source, detector sensitivity and patient-dependent parameters. Theoretical and experimental

verifications confirm that the computation of Sp02, employing this invention is indeed unaffected either by the variation in the input intensity of the sources or the sensitivities of the detectors employed. Accordingly, extensive calibration which is normally done for the conventional pulse oximeters is not essential in cases where this invention is used.
A part of the body, say, the finger is exposed to two light sources having different wavelengths, one in the red (R) region of the visible spectrum and the other in the infrared (IR) region. These lights enter the body part and immediately on exiting the body part (either after getting reflected within the body part or passing (getting transmitted) through the body part without
reflection) the "reflected" lights (PPGR*) or transmitted lights
(PPGn) are detected by photo diode circuits.
The noise and interference (created by 50 Hz / 60 Hz supply
frequency) is removed using low pass filters.
Designating the resulting signals as vR and V[Rt the natural
logarithm ofvR and v[R is calculated to get ln(v#) and ln(vIR).
Next, the square roots of the R and IR PPG signals v'ln(v.)) and v'Wv.., \ are then computed, The slopes of the linear portions of vln(vF)) and yituvv, j are determined as mR and mm . Then, the oxygen saturation is computed as


Hemoglobin in arterial blood transports nearly 98 % of oxygen and the remaining 2 % is permanently dissolved in plasma and hence is of no use. One hemoglobin molecule can carry up to four oxygen molecules. Thus one can measure the level of oxygen (Sa02) in arterial blood indirectly by determining the amount of oxygenated hemoglobin as a percentage of total hemoglobin in arterial blood. If we denote
the amount of hemoglobin without oxygen as (Hb) and that of
the oxygenated hemoglobin as(Hb02) in arterial blood, then Sa02 as a percentage can be expressed as

Here the symbol ( } denotes the amount of a particular
quantity. The above equation assumes that the amount of carboxy-hemoglobin and methemoglobin are negligible. When Sa02 is measured using the principle of pulse oximetry, the abbreviation used Sp02. In pulse oximetry, a part of the body is illuminated with two monochromatic light sources, one in the red region and the other in the infrared (IR) region. Either the reflected or transmitted light is then detected. The detected signal is called PhotoPlethysmoGraph (PPG). Existing methods extract the dc and pulsatile components of R and IR PPGs and estimate the ratio of quotients of AC and DC amplitudes of

the R and IR PPG signal This ratio is then used to compute the
saturation level of oxygen in arterial blood employing a linear
approximation to an experimentally derived calibration curve
obtained from healthy young volunteers after induction of
hypoxemia.
The present invention is intended for the measurement of Sp02
from suitably processed R and IR PPGs. The processing results
in the final expression for calculation of Sp02 to become not
only independent of human dependent parameters such as skin
pigmentation, volume of the intervening tissue but also
independent of input source intensity and detector sensitivity.
* Thus the proposed invention estimation of Sp02 does not
require extensive calibration.
As in the existing methods, in the case of this invention also, a part of the body is exposed, to two monochromatic light sources, one in the red region (visible light spectrum) and the other in the infrared region and either the reflected lights (PPGRA.) or transmitted lights (PPGT^,) are detected using photo diode circuits. The sources and the detectors will have to be housed on the same plane to obtain a PPG*R. but PPGTA. can be obtained only with the source and the detector arranged on two

different but parallel planes. Hence PPGn is preferably pobtained from the extremities of the body, such as, the earlobe and fingertip.
Though the invention proposed is applicable equally for both reflectance type and transmittance type PPGs, the description given herein is for a transmission type PPG only for brevity. However, one can obtain relevant expressions and explanations simply by replacing the word "absorption" by "reflection" and using appropriate symbols. To obtain the required PPGs, two light sources (one R and one IR light emitting diodes) having output intensities of IR and IIR are placed on one side of an extremity, say, finger. The transmitted light through the finger is detected by placing appropriate photo detectors (photo diodes) on the side opposite to that of the source (Fig.l (a)). For reflection type the light
travels through the skin and gets reflected (Fig. 1(b) and Fig, 1(c)). The incident R and IR waves travel through the finger and the transmitted light after absorption within the finger is detected by the photo detectors. Since the total volume covered by the blood vessels is quite small, majority of the photons emitted from the sources go through the path made of epidermis-tissue-soft bone-tissue-epidermis and reach the detectors. Only a very small fraction, 1 in 50,000 to 1 in

100,000 of the photons emitted by the source goes through the path that includes blood vessels (epidermis-tissue-blood-tissue-bone-tissue-epidermis or epidermis-tissue-bone-tissue-blood-tissue-epidermis or epidermis-tissue-blood-tissue-bone-tissue-blood-tissue-epidermis).
The output of the photo diode will have three components, one at the DC due to the light path epidermis-tissue-bone-tissue-epidermis, the second at very low frequency due to the light path epidermis-tissue-venous blood-bone-tissue-epidermis and the third at the frequency of heart rate due to the light path epidennis-tissue-arterialblood-bone-tissue-epidennis.
To model the light propagation, the path of light through the finger is assumed to have a uniform cross section of area A having a total length / as shown in Fig. 2. A disc of thickness dx9 with attenuation across the disc dix on its input (intensity of light) ix is chosen in this model. The volume (nearly 99 %) of the disc being considered will be mostly made of cells of dermis, tissue and bones depending on the position of the disc. Each cell of a particular type attenuates certain amount of light passing through that cell depending on its optical property and also scatters some of the incident light. Let ot., ar.} and cc . be the
coefficients of optical attenuation (including the effect of scattering) of the dermis, tissue and bone cells at a wavelength X. Let the cross sectional areas of these cells heAp^, ATI and ABo respectively. Each cell will then attenuate light depending on their optical characteristics. The attenuation due to a typical cell, say tissue, will be as given in Fig.

3(a). This attenuation can be considered as though a traction of the cross sectional area of a particular cell is completely opaque and the rest of the cross sectional area of that cell completely transparent as indicated in Fig. 3(b). Then the total attenuation across the disc of area A and thickness dx due to a particular type of cell, say tissue, will be a_r.^_.
11A, 11
N A dx , where N is the number of tissue cells per unit volume (cells/m ). We can write
an equation for attenuation dix across the disc of length dx as
(2)
Here JV__ and JST are the number of dermis and bone cells per unit volume (cells/m3) respectively. Rearranging and integrating equation (2) with limits x=0 to x=l, we get

Here IiYX is the intensity of the received light at the detector and Isx is the intensity of the source emitting light at a wavelength L The DC output from the photo diode will then be
Here KDx is the sensitivity of the photo detector at X, the wavelength of the source. As
mentioned earlier, some of the photons also travel through arteries carrying blood that contains 55 % plasma, 43 % red blood cells (mostly made of Hemoglobin, Hb and Hb02) 1.5% of white blood cells (leukocytes) and 0.5% platelets. Of these components of arterial blood, attenuation due to red blood cells alone are significant as the amount of white blood cells and platelets are negligible compared to the number of red blood cells and the plasma has nearly zero optical attenuation in the wavelength region of interest. Let Am and AHio be the cross sectional areas and a..and aMM be the coefficients of optical

attenuation of hemoglobin and oxyhemoglobin respectively. We now combine together all the individual arteries that are in the path 3 of light as a single equivalent artery having an equivalent length of interaction with light as H and equivalent radius r as indicated in Fig. 2. As blood is pumped into the arteries, the vessels enlarge and shrink at the rate set by the heart and hence r is a time varying quantity. Let r be the maximum radius of the equivalent artery (radius of the vessel when the blood flowing in it, is a maximum). Then we can write the attenuation due to arterial blood as.
a =-f (a*uA»N» + ^W^^W2^ & (3)
Here i^ is the instantaneous intensity of light that passes through arterial blood vessels after passing through skin, tissue and bone. In rH is the instantaneous cross sectional area and 2TZ rH is the maximum value of the cross sectional area of the equivalent artery. Rearranging equation (3) and integrating with limits 0 to £ we get

Here ibox is the time varying component of light intensity received at the detector and a I0A
is the peak value of that portion of the light intensity attenuated by arteries also, apart from dermis, tissue and soft bones. Hence the pulsatile ac component at the output of the detector would be
(4)

We Claim:
1. A method for the non-invasive measurement of oxygen saturation in arterial blood comprising the steps of exposing a part of the patient's body to R and IR light; detection of these lights on exiting the body part by photo diode circuits; normalizing the R and IR PPGs; removing the noise in the R and IR PPGs, by filters, to obtain signals vR and v^ obtaining the natural logarithm of vR and vIR to get ln(yR) and ln(vja), and the peak amplitudes of ln(v* ) and ln(v/*), that is, VpR and VPJX;; obtaining the square roots of the R and IR PPG signals yjhx(vR)) and ,Jh(v^)'7 determining the slopes of the linear portions of yj\xi(yR)) and yj\n(vIR) as mR and mIR: and calculating the oxygen saturation

2. A method for the non-invasive measurement of oxygen
saturation in arterial blood substantially as herein described
with reference to the Example.
3. A device for the non-invasive measurement of oxygen
saturation in arterial blood, comprising two light emitting
sources for exposing a part of the patient's body to red light (R)
in the visible spectrum and infra red light (IR); a photo diode
circuit for detection of these lights on exiting the body part;
means for the normalization of the exiting red (R) and infrared
(IR) PPGs (PPGTJI or PPGJU); filters for removing the noise in
the R and IR PPGs to obtain signals vR and v1R means for