`
`APL_MAS_ITC_00015678
`
`1
`
`1
`
`,
`
`44
`
`Design of pulse oximeters
`
`4.2.3 He,noglobin absorbance spectra
`
`The chemical binding of the different hemoglobin species changes the physical
`properties of the hemoglobin as well. Figure 4.2 shows the extinction coefficients
`of oxyhemoglobin, reduced hemoglobin, methemoglobin and carboxyhemoglobin
`at wavelengths in the range of interest in pulse oximetry.
`The absorbance of light in the red region of the spectrum is much higher for
`reduced hemoglobin than for oxyhemoglobin. The extinction coefficients of both
`hemoglobin species are equal at the point isosbestic point (805 nm). The reduced
`hemoglobin is more transparent to light from the infrared region than
`oxyhemoglobin.
`The extinction coefficient of carboxyhemoglobin is about the same as that of
`oxyhemoglobin at the wavelength of 660 nm while it is almost transparent in the
`infrared region. Methemoglobin absorbs much light in the red region of the
`spectrum and its extinction coefficient remains higher than that of oxyhemoglobin
`in the infrared region.
`
`mothemoolol,In
`
`087.0.0910'Dil
`
`\,iduced
`\Dern<glothn
`
`to.
`
`C
`11
`
`0 0
`
`0C0
`
`2
`XLU
`
`01,
`600
`
`840
`
`1
`680
`
`720
`
`1
`760
`800
`840
`Wavelength (nm)
`
`C[rbo/'llemoolobin
`1
`Bao
`960
`920
`
`1000
`
`Figure 4.2 Extinction coefficients of the four most common hemoglobin species oxyhemoglobin,
`reduced hemoglobin, carboxyhemoglobin, and methemoglobin at the wavelengths of interest in
`pulse oximetry (courtesy of Susan Manson, Biox/Ohmeda, Boulder, CO).
`
`4.3 BEER'S LAW IN PULSE OXIMETRY
`
`Pulse oximeters determine the oxygen saturation of arterial blood by measuring
`the light absorbance of living tissue at two different wavelengths and using the
`arterial pulsation to differentiate between absorbance of arterial blood and other
`absorbers.
`
`61
`
`MASIMO 2053
`Apple v. Masimo
`IPR2022-01299
`
`
`
`RX-0035.0062
`
`APL_MAS_ITC_00015679
`
`Light absorbance in pidse oximetry
`
`45
`
`4.3.1 Criteria for the choice of wavelengths
`Different reasons lead to the most common choice for wavelengths used in pulse
`oximetry. The red skin pigmentation absorbs a great amount of light at
`wavelengths shorter than 600 nm and therefore it is not desirable to measure
`light absorbance in this range. Large differences in the extinction coefficients of
`reduced hemoglobin and oxygenated hemoglobin change the absorbance of ] ight
`significantly, even when the oxygen saturation changes slightly. A good choice
`for a wavelength in the red region is 660 nm because of a large difference in the
`extinction coefficients.
`Another issue for the wavelength choice is flatness of the absorption spectra
`shown in figure 4.2 around the chosen wavelength. Otherwise shifts in the peak
`wavelength of the LEDs (see section 5.3) will result in a larger error. The
`absorbance spectra of reduced hemoglobin and oxygenated hemoglobin are
`relatively flat at 660 and 940 nm (Moyle 1994).
`Mannheimer et al (1997) have shown that sensors fabricated with 735 and
`890 nm emitters read more accurately at low saturations under a variety of
`conditions, while 660 and 990 nm emitters read more accurately at high
`saturations.
`
`4.3.2 Absorbance in hemoglobin solutions
`
`The different species of hemoglobin are the main light absorbers in arterial and
`venous blood. Most of the hemoglobin in human blood is either oxygenated or
`reduced hemoglobin which determine the functional oxygen saturalion SO2
`(equation (4.5)). The concentrations of oxygenated hemoglobin (cilbO2) and
`reduced hemoglobin (cHb) can be expressed as a function of SO2 as a fraction and
`the sum of the concentrations CHb02 and cHb
`cHb02 = SO2~CHbO2 + CIIb )
`CHb - (1 - SC)2)( CHb02 + CHb).
`
`(4.8)
`(4.9)
`According to Beer's law we derive the total absorbance At of a solution
`containing only reduced and oxygenated hemoglobin as absorbing substances
`from equation (4.4)
`
`At =81.-Ib02(A)CHbo: d}lb02 +EHb(A)cHbdHh-
`
`(4.10)
`
`Assuming that the optical path length d is the same for the oxygenated
`hemoglobin (411)02) and reduced hemoglobin (dilb) and using equations (4.8),
`(4.9), and (4.10), we derive
`
`At = ~EHI:>02 (A)SC)2 + £Ht,(A)(1- S02) ~(CHI:> +CHb02 )d. (4.11)
`
`Thus At can be expressed for known concentrations of hemoglobin in terms
`of functional oxygen saturation as a fraction, the extinction coefficients of
`hemoglobin, and the length of the optical path. Values for the extinction
`coefficients of adult reduced hemoglobin (4[b) and adult oxygenated hemoglobin
`
`62
`
`MASIMO 2053
`Apple v. Masimo
`IPR2022-01299
`
`
`
`RX-0035.0063
`
`APL_MAS_ITC_00015680
`
`46
`
`Design of pulse oximeters
`
`(€HbO,) at the two wavelengths most commonly used in pulse oximetry (660 nm
`and 940 nm) have been measured by ZOIstra et al (1991) (see table 4.1).
`
`Table 4.1 Table of extinction coefficients of reduced and oxygenated hemoglobin in adults at the
`wavelengths of 660 nm and 940 nm (values from Zijlstra etal !99]).
`
`Wavelength, nm
`660
`940
`
`Extinction coefficient. L mmol-' cm-1
`Hb
`Hb02
`0.08
`0.81
`0.18
`0.29
`
`Figure 4.3 shows the characteristics of light absorbance for a sample with a
`fixed concentration of total functional hemoglobin (CHboa + CHb) of 1 mmol L-1,
`a fixed path lenglh d of 1 cm and varying functional oxygen saturations. The two
`lines shown in figure 4.4 represent the properties for the two most commonly
`used wavelengths in pulse oximetry (660 nm and 940 nm). The absorbance of
`light at a wavelength of 940 nm increases with an increased oxygen saturation. At
`660 nm the absorbance of light decreases rapidly with an increasing functional
`oxygen saturation (Pologe 1987).
`It is possible to determine the concentrations of hemoglobins in hemoglobin
`solutions or hemolized blood by using a device such as a spectrophotometer (see
`section 3.3).
`
`~----ai 660 nm~
`-S-al-~ nrnl
`
`1
`1
`80
`60
`40
`20
`Functional oxygen saturation (%)
`
`100
`
`0.9 -
`08
`0.7 -
`06-
`0.5 -
`0.4 -
`0,3 -
`0.2 -
`0.1
`0
`
`0
`
`Normalized absorbance
`
`Figure 4.3 Changes in light absorbance in hemoglobin solutions as a function of functional
`oxygen saturation for the wavelengths used in pulse oximetry. Absorbonce decreases rapidly with
`increasing oxygen saturation at 660 nm (dashed line) but increases slightly with increasing oxygen
`saturation 21 940 nm (solid line).
`
`4.3.3 Pulsation of the blood
`
`Light traveling through biological tissue (e.g. the finger or earlobe) is absorbed
`by different absorbing substances. Primary absorbers of light in the region of
`interest are the skin pigmentation, bones, and the arterial and venous blood.
`Instead of measuring the arterial oxygen saturation of the blood in vitro with a
`sample of arterial blood and a spectrophotometer, or at a wide range of different
`wavelengths as with the Hewlett-Packard ear oximeter, pulse oximeters take
`
`63
`
`11 1
`
`1
`
`MASIMO 2053
`Apple v. Masimo
`IPR2022-01299
`
`
`
`RX-0035.0064
`
`APL_MAS_ITC_00015681
`
`Light absorbance in pulse oximetry
`
`47
`advantage of arterial pulsation. Figure 4.5 shows the amount of absorbed and
`transmitted light in living tissue as a function of time.
`The arteries contain more blood during systole than during diastole, and
`therefore, their diameter increases due to increased pressure. This, effect occurs
`only in the arteries and arterioles but not in the veins. The absorbance of light in
`tissues with arteries increases during systole mainly because of the larger amount
`of absorbing substances (hemoglobin), due to the fact that the optical path length
`d in the arteries increases. This alternating part of the total absorbance allows us
`to differentiate between the absorbance due to venous blood, a constant amount of
`arterial blood, and other nonpulsatile components such as skin pigmentation (dc
`component of the total absorbance) and the absorbance due to the pulsatile
`component of the arterial blood (ac component). The alternating part of the light
`absorbed by the living tissue usually does not exceed 1 % to 2% o f the constant
`absorbance of the dc components. The time varying signal of transmitted light is
`refurred to as the plethysmographic (or photoplethysmographic) signal.
`The intensity of the light passing through the tissue during diastole is high
`(IH)· The absorbers that are present during diastole are the DC components. All
`DC components except the nonpulsating arterial blood are collectively
`represented by ED((,11, cDC, and dDC· The diameter of the arterial vessels is
`minimal (dmin) and therefore the absorbance due to arterial hemoglobin is
`minimal and the amount of transmitted light is high ( /H) and has a peak (see
`figures 4.4 and 4.5)
`
`41 = ID e-£De(A)cDC d[)Ce-I£Hb(A)CHb+£HbO2 (A)CHbOG]dmin (4.12)
`
`ft 1 1-
`10«4~- ''9'777-,i//07 Nonpulsaling arterial blood
`
`t
`
`Pulsatingarterial blood
`
`Venous blood
`
`E E
`
`0* Ii' -**~jor-u
`
`[r,cluent ligi,1 4]
`
`1- 1
`
`One cardiac
`cycle
`
`Other tissue
`
`t
`
`Figure 4.4 Absorbed anci transmilted light in living tissue. The amount of absorbed light
`coirclates witli the pulsation of aiterial blood. A constant amount of- light is absorbed by the skin
`pigmentation. bone, other t issue, venous blood and tlic noripulsating part of the arterial blood.
`More blood is present in the aiteries during systole and therefore more light is absorbed. The
`intensity of the transmitted light varies from 41 (maximum) to /L (minimum) within one cardiac
`cycle
`The optical path length in the arteries increases during the systole to dmax. The
`amount of absorbed light reaches a maximum peak and therefore the transmitted
`light reaches the low peak /L
`
`64
`
`MASIMO 2053
`Apple v. Masimo
`IPR2022-01299
`
`
`
`RX-0035.0065
`
`APL_MAS_ITC_00015682
`
`48
`
`Design of pulse oximeters
`
`4. = /O e-EDC(A)CDC £4)Ce-££Hb(A)6·Hb +EHbO2(A)£.Hboildmax
`
`(4.13)
`
`The light intensity I of the light arriving at the photodetector is a function of the
`diameter d of the arteries and arterioles. During one cardiac cycle this diameter
`changes from dmin to dmax· By substituting d with dmin + Ad we derive the
`fullowing expression from Beer's law, where / is expressed as a function of 41
`and Art, the part of the diameter that changes from 0 to dinax - dmin with time
`(4.14)
`
`I =IHe -[8Ht,(A)c·Hb +€HbO2 (A)CHI,02 ]Ad
`
`Figure 4.5 shows these properties in a simplified model.
`
`253<venous,v.AT:.1 ,
`02< Arteriali-1 Photo-
`4ioAM.&*blood, r,~ov
`LED 'r---*pigment.*%' bjood €
`detector
`~----'*ation,
`3.Metc. *455*«
`1
`
`1
`1
`1
`1
`1
`1
`1
`1
`
`Ught
`intensity /
`
`/0
`
`dmin
`
`1-11 1
`1
`lH-- - 4
`1 Ill
`1 1
`.1. 1 .1
`
`dDC
`
`dmax
`
`Figure 4.5 Beer's law in pulse oximetry. The DC components of the tissue (e.g. skin
`pigmen[ation. bone, venous blood and the nonpulsating part of the arterial blood) absorb a constant
`amount of the inciden, light lo. The effective optical path length in the DC components without the
`constant level of arterial blood is represented by dD£· During diastole [hc optical path length
`[hrough the arteries has a minimum length of dmin and the ligh[ in[ensity at [hc photodetector is
`maximal (41)· 'Ilhe optical path length reaches a maximum d~ax during systolc and the hemoglobin
`in the artenes absorbs a maximum amount, causing / to decrease to a minimum level of /L
`
`4.3.4 Measurement of pulse oximeters
`
`The reading of the pulse oximeter SpO2 is an estimation of the arterial oxygen
`saturation Sa02· Measuring at two wavelengths allows us to distinguish the
`concentrations of only two different absorbers (Hb and HbO2). But in humans
`more species of hemoglobin, such as carboxyhemoglobin and methemoglobin, are
`present. These other hemoglobins absorb light as the functional hemoglobiris do
`and therefore influence our measurements. As long as we do not measure at as
`many wavelengths as absorbers are present in the blood, we can not determine the
`concentrations of Hb and HbO2 and therefore the arterial oxygen saturation
`correctly (Barker and Tremper 1987).
`
`65
`
`MASIMO 2053
`Apple v. Masimo
`IPR2022-01299
`
`
`
`RX-0035.0066
`
`APL_MAS_ITC_00015683
`
`Light absorbance in pulse oximetr¥
`
`49
`
`Due to the fact that Hb and HbO2 are the main absorbers, the error may be
`small. Nevertheless, the results of determining either the actual functional o r
`fractional oxygen saturation (see equations (4.5) and (4.7)) of the arterial blood
`are not exact. This problem is also discussed in sections 10.1.1 and 11.1.1. The
`oximeter reading becomes less accurate if the concentrations of dyshemoglobins
`are larger than in normal humans. Section 11.7 deals with the presence of high
`concentrations of dysfunctional hemoglobins.
`
`4.4 SATURATION VERSUS NORMALIZED RATIO
`
`The arterial oxygen saturation can be derived based on Beer' s law as a function
`of the ratio of absorbances at two wavelengths. Due to nonlinearities in the LEDs,
`the photodetector, and light absorbance in the tissue, the absorbances have to be
`normalized in the ratio. This model results in a theoretical calibration curve, but
`it is not used in practice as will be described in the following sections.
`
`4.4.1 Normaliz.ation
`
`The measured light intensities at the different wavelengths have to be normalized
`before they can be compared with each other due to the fact that the light-
`emitting diodes (LEDs) may emit light with different intensities. The absorbing
`characteristics of the DC components and the sensitivity of the photodetector
`differ for the two different wavelengths and the tissue absorption and path length
`varys widely from patient to patient and with the probe site (de Kock and
`Tarassenko 1991). The normalized signal In is calculated by dividing the
`transmitted light intensities (the raw signals) by their individual maximum peaks
`(41,R for the red wavelength and IH,IR for the infrared wavelength). From
`equation (4.14) we derive
`
`In - J_ = e-[€Hb(A)CHb +EHb02 (A)CHI:,02 ]Ad
`IH
`
`(4.15)
`
`This results in normalized signals with the same intensities 41,n during diastole.
`The normalized signals of the transmitted red and infrared light are independent
`of the incident light levels and photodetector nonlinearities as shown in figure
`4.6. The AC components of the normalized signals represent only changes of
`transmitted light caused by the pulsation of blood in the arteries and can be
`compared with each other. They depend on the absorbers present in the arterial
`blood (ideally Hb and HbO2) and the actual optical path length d through the
`volume changing part of the arteries.
`
`4.4.2 Ratio of normalized signals
`
`The absorbance of the light is derived by calculating the natural logarithm of the
`measured and normalized transmitted light level. Dividing the raw signal by the
`iransmitted light during diastole 4i as in equation (4.15) and calculating the total
`absorbance then is comparable to calculating the total absorbance only due to the
`AC components in the pathway. The transmitted light during diastole represents
`the new nonchanging incident light level and the ratio R of these normalized
`
`66
`
`MASIMO 2053
`Apple v. Masimo
`IPR2022-01299
`
`
`
`RX-0035.0067
`
`APL_MAS_ITC_00015684
`
`50
`
`Design of pulse oximeters
`
`absorbances at the red (R) and infrared (IR) wavelengths depends only on the
`light absorbers present in the arterial blood (see equation (4.3))
`
`/H,IR
`
`t
`41.R
`1 .R
`
`t
`41,n
`44,n
`6.,AA
`6.t€h.Op
`646:· Id 19023
`
`Normalized signals
`'Raw' signals
`Figure 4.6 The nonnalization of the signals The [ransmitted light from the red LED (R) and
`from the infrar'ed LED CIR) is divided by its individual DC component. 11,us. both normalized light
`intensities have [lic same magnitude during diastole. The normalized signals determine the basis for
`the calcul•lion al- the arterial oxygen samration.
`
`ALR ln(IL,R /41,R)
`At, IR lnUL, IR /Ill,IR )
`
`By using equation (4.15) the ratio can be derived as
`
`R=
`
`[(£Ht,(AR )CHb +(EHbO, (AR k.'Ht-,02 ~AdR
`[(CHI)(AIR)(Hb + (EnbO2 (kIR)('HbO·, ]Ad]R
`
`(4.16)
`
`(4.17)
`
`Assuming that the optical path lengths d~ for red light and dIR for the infrared
`light are equal, only the arteries change their diameter, and using equation (4.ID
`
`R=
`
`£Hb(AR)+[€HbC)2 (AR 1 -Ellb (AR )]SaO2
`£Hb (AIR ) + [£Hboa C XIR 1 - €111, ( AIR )]S:102
`
`(4.18)
`
`In this form the ratio R is not a function of the optical path length and can be
`derived from the arterial oxygen saturation instead of the concentration of the
`hemoglobins in the blood (see de Kock and Tarassenko 1993).
`
`4.4.3 Theoretic calibration citrve
`
`Equation (4.18) can be rewritten in a form where Sa02 is a function of the
`measured and calculated ratio R
`€Hb (AR)- €Hh (AIR)R
`£Hb (AR ) - £HbO'l (XR ) + ~€Ht,02 (AIR ) - €Hb (AIR )]R
`
`Sa02=
`
`x 100%. (4.19)
`
`Therefore. the functional oxygen saturation in arterial blood can be derived
`theoretically by calculating the ratio R of measured and normalized total light
`
`67
`
`1
`
`MASIMO 2053
`Apple v. Masimo
`IPR2022-01299
`
`
`
`RX-0035.0068
`
`APL_MAS_ITC_00015685
`
`Light absorbance in pulse oximetry
`
`51
`
`absorbances in the red and infrared region and using equation (4.19). Figure 4,7
`plots this relationship as the theoretical calibration curve.
`
`- theoretical I
`Amparl/=1 1
`
`1
`1
`
`t
`2
`3
`Ratio of normalized absorbances
`
`1
`4
`
`1
`5
`
`100 -
`90 -
`80 -
`9 70 -
` 60-
`2 50-
`e
`40 -
` 30 -
`* 20 -
`0
`10 -
`0
`0
`
`C6
`
`0
`
`00
`
`1
`
`Figure 4.7 Calibration curves for pulse oximeters: the solid line is the theoretical curve by Beer's
`law lind the dashed I ine is the empirical curve. The difference between these curves is due mainly
`to light scatte,ing effects. This empirical calibration curve is derived by a second order polynomial.
`
`4.5 VALIDITY OF BEER'S LAW IN PULSE OXIMETRY
`
`Incident light passing through human tissue is not split only into absorbed light
`and transmitted light as proposed by Beer's law. Some parts of the light are
`reflected and others are scattered.
`Light reflection at the skin surface and light absorbance due to tissue other
`than the pulsating arterial blood are overcome by using the plethysinographic
`waveform. However, the skin surface, tissue, muscle, bone and eapecially blood
`cause light scattering which increases the absorbance of light (see following
`section). Blood is a nonhomogeneous liquid, which is capable of nonlinear
`absorbance of light, e.g. as the concentration of hemoglobins varies (Wukitsch et
`al 1988).
`The variation in light absorbance is not entirely due to the increased optical
`path length during systole. If the change in diameter were the only reason, the
`variation would be much less. The reason is a change in the axis of the red blood
`cells, which changes their absorbance as well. Red blood cells have the shape of a
`biconcave disk. Their major diameter is aligned parallel to the direction of blood
`flow during diastile and aligns perpendicular to the direction of flow during
`qystole. l'herefore. the optical path length is larger during systole and increases
`light absorbance. Even the light reflectance changes with the axis of the red blood
`cells, which is important for the use of reflectance probes. As a result of these
`properties, the absorbance and reflectance of blood in motion varys within the
`cardiac cycle and with the velocity of blood flow (Moyle 1994).
`
`68
`
`MASIMO 2053
`Apple v. Masimo
`IPR2022-01299
`
`
`
`RX-0035.0069
`
`APL_MAS_ITC_00015686
`
`1 52
`
`Design of pulse oximeters
`
`4.6 LIGHT SCATTERING
`
`The results of oximetry measurements with whole blood differ from the results
`of the theory based on Beer ' s law . A physical phenomenon called light scattering
`highly increases the absorbance of light. Nevertheless, pulse oximeters read the
`arterial oxygen saturation of the blood accurately enough for clinical use under
`normal circumstances. This is due to the fact that most of the commercial pulse
`oximeters use a calibration curve based on empirical data, because modeling the
`problem of light scattering mathematically for different conditions is very
`complex. Several approaches have been made to create models which describe the
`real process within certain limits of accuracy.
`
`4.6.1 Light absorbance in whole blood
`
`Unfortunately Beer' s law does not apply for whole blood. The absorbance of
`light is not simply proportional to the concentration of hemoglobin or to the
`length of the optical path. Beer's law assumes no light scattering, which is not
`true in whole blood, besides the fact that the LEDs do not emit monochromatic
`light.
`Shymada and Yoshida (1984) verified that the influence of multiple
`scattering can not be overcome be subtracting the DC level as had been expected.
`Kramer et al (1951) stated that the absorbance of light due to oxyhemoglobin and
`reduced hemoglobin is increased in whole blood compared to hemolyzed blood
`by factors of the order of five.
`The reasons for the increased absorbance are mainly scattering and multiple
`scattering . Light scattering causes the deviation of a light beam from its initial
`direction. It occurs when light is refracted by an object of a size similar to the
`magnitude of the wavelength of the light and a change in the index of refraction
`at the interface of this object. The wavelengths of red and infrared light do have
`the same order of magnitude as the geometric dimensions of red blood cells
`(approximately 7 pm in diameter). The discontinuity in the index of refraction at
`the interface between plasma and red blood cells and the great proportion of red
`blood cells in blood yield a highly light scattering medium. Light that is scattered
`once will likely be scattered again by cells and therefore multiple scattering
`occurs (Steinke and Shepherd 1986). Multiple scattering increases the optical path
`length and therefore increases the absorbance.
`The intensity of the light scattered by the tissue depends on such factors as
`the red blood cell concentration in the blood; on the size, shape, orientation, and
`index of refraction of the scattering particles; on the tissue thickness; and on the
`aperture cone of the detector (Fine and Weinreb 1995). The thickness of the
`tissue, the distance between the LED and the photodiode, and the concentration of
`hemoglobin will vary from patient to patient and the shape and orientation of the
`red blood cells is irregular. Thus it is difficult to develop a physical model which
`can be used under different circumstances.
`
`4.6.2 Models for light absorbance including scattering
`
`It would be very useful to find a relationship between SaO2 and the ratio R of
`normalized absorbances for whole blood instead of only for hemoglobin
`solutions. An accurate scattering theory for whole blood could replace the
`
`69
`
`MASIMO 2053
`Apple v. Masimo
`IPR2022-01299
`
`
`
`RX-0035.0070
`
`APL_MAS_ITC_00015687
`
`Light absorbance in pulse oximetry
`
`53
`
`empirical calibration curves used for the SpO2 readings. A few attempts are
`described below.
`
`4. 6. 2. 1 Twersky 's multiple scattering theory. Twersky ( 1962 , 1970a, b) has
`developed an analytical theory to describe the scattering of light by large, low-
`refracting, and absorbing particles. It is based on electromagnetic field theory
`and uses statistical averages to expand the theory for scattering and absorbing
`valid for a single particle, to find a formulation valid for multiple scattering (de
`Kock and Tarassenko 1993).
`The total absorbance of whole blood can be expressed as the sum of
`absorbance as described by Beer's law and a second term representing the
`attenuation of light due to scattering. These two processes can be treated as
`independent processes. The intensity of scattering depends on variables such as
`those mentioned i n section 4.6.1. The theory can be adapted for a special setting
`and will provide accurate results, but once the physiological conditions change,
`recalibration is required (Fine and Weinreb 1995). Hitachi, Ltd uses Twersky's
`approach in one of their US patents (Ito et al 1993).
`
`4. 6 2. 2 Comparison of dijferent models. Steinke and Sheperd (] 986) compared
`Twersky's theory of radiation scattering and photon diffusion equations. They
`found Twersky's original equation to give the best fit for the measured data.
`Marble et al (1994) found the three dimensional photon diffusion theory to
`be useful for modeling tissue optics although the pulse oximeter system violates
`many of the requirements of the model. However, they came to the conclusion
`that this theory can not replace clinical calibration studies.
`De Kock and Tai·assenko ( 1993) also found Twersky's theory to give the best
`fit to the experimen[al data. They compared results of this model with the photon
`diffusion theory and the Kubelka-Munk theory.
`
`4.6.3 Influence of scattering on pulse oximeter readings
`
`Although the assumptions of Beer's law are violated in pulse oximetry, the actual
`readings of the devices show a good correlation between the measurement and the
`actual arterial oxygen saturation.
`Steinke and Sheperd (1986) found that the scattering effects of the light
`passing through whole blood depend on the wavelength of the light and the
`oxygen saturation. The relationship between oxygen saturation and total
`scattering effects {absorbance due to hemoglobin plus multiple scattering) is
`approximately linear and so scattering does not influence the IiIlearily of the pulse
`oximeter in a negative way. In contrast, the total absorbance has a larger slope
`than that due only to the absorbance of hemoglobin following Beer's law.
`Therefore, light scattering increases the sensitivity of the whole blood nximeter.
`Fine and Weinreb (1993, 1995) demonstrate that the ratio of total
`absorbances is a function of the effective blood layer thickness and the
`concentration of hemoglobin. Therefore physiological factors such as
`temperature or peripheral vasoconstriction reduce the accuracy of saturation
`readings. The error increases as the level of arterial oxygen saturation decreases.
`This is dangerous because the clinician has to question the readings of the oxygen
`saturation when it is most critical for the patient.
`
`70
`
`MASIMO 2053
`Apple v. Masimo
`IPR2022-01299
`
`
`
`RX-0035.0071
`
`APL_MAS_ITC_00015688
`
`4
`4
`4
`
`44
`
`444
`
`4
`
`44
`
`4
`
`54
`
`Design of pulse oximeters
`
`4.6.4 Calibration curves used ,for pulse oximeters
`
`Commercial pulse oximeters are calibrated from iii vir,-0 data (see section 10.1 )
`A large set of data obtained in clinical studies is collected containing information
`about the ratio R of' the absorbances calculated by the pulse oximeter and the
`actual arterial oxygen saturation Sa02 measured by a very accurate method such
`as the CO-oximeter (see section 3 3 I. Lookup tables or equations are used to find
`the relatinnship of these two variables for a pulse oximeter reading.
`To relate the measured values of the ratio R to the reading of the pulse
`oximeter, the equation of the theoretical calibration curve based on Beer's law
`can be modified as Mendelson and Kent (1989) described
`
`ki - k2R
`402=4-k#R
`
`(4.20)
`
`In this equation the extinction coefficients from equation (4.19) are replaced by
`constants ki. These constants are determined by clinical studies to give the curve a
`to the in ritro measured data. Another approach for a mathematical
`best fit
`representation is the use of a polynomial such as found for example in the
`Ohmeda 3700 and Radiometer OX100 pulse oximeters (Fine and Weinreb 1995)
`
`Sp02 =4 +k,R+43*2
`
`(4.21)
`
`Figure 4.7 provides an example of a calibration curve used in pulse oximeters in
`comparison to the theoretical calibration curve.
`
`REFERENCES
`Barker S J and Tremper K K 1987 Pulse oximetry : applications and limitations /nt. Anesthesiol.
`Clinics 25 155-75
`Bunn H F 1986 He,nogiobin: Molecular, Genetic, and Clinical Aspects (Philadelphia PA-
`Saunders)
`Fine I and Weinreb A 1993 Multiple scattering effect in transmission oximetry Med. Biol. Eng.
`Comput. 31 5 16-22
`Fine I and Weinreb A 1995 Multiple scattering effect in transmission pulse oximetry Med. Biol.
`Eng. Comput. 33 709-12
`de Kock J P and TarassenkoL 199 ! in vitro investigalion of tlic factors affecting pulse oximetry J.
`Biomed. En,t: 13 6 1 -6
`de Kock J P und Turassenko L !993 Pulse oximetry: [hcoretical and experimental models Med.
`Biol. Eng. Compur. 31 29 1 -300
`Ito Y, Kawaguchi F Yoshida M and Kohida 11 1993 Method and equipment for measuring
`absorptance of light scattering materials using plund wavelengths of light US patent 5.239. 185
`Kramer K, Elam J O. Saxion G A and Elam W N Jr 1951 influence of oxygen saturation.
`crythrocyte concentration and optical depth upon the red and near-infrared light transrnittance of
`whoic blood Am J Physiol. 165 229-46
`Malinheimcr P D, Casciana J R. Fein M E :ind Nierlich S L 1997 Wavelength selection for low-
`s,ituration pulse oximetry IEEETrans. Bionic {L Eng. 44 148-58
`Marble D R, Bums D H and Cheung P W 1994 Diffusion-based model of pulse oxiinetry: in vitro
`und in vivo comparisons Appl Op, 33 1279-85
`Mendelson Y and Kent JCI 989 Variationx in optical absorption spectra of adult und fetal
`hemoglobins and its effect on pulse oximetry /EEE Truns. Bionied Eng. 36 844-8
`Moyle JTB 1994 Pul.ve Oximaers iLondon: BMJ)
`Nellcor 1993 Hemoglobin and the principles of pulse oximetry Referenec Note: Pulse Oximetry
`Nofe Number ! (Picasanton , CA: Nellcor)
`
`71
`
`MASIMO 2053
`Apple v. Masimo
`IPR2022-01299
`
`
`
`RX-0035.0072
`
`APL_MAS_ITC_00015689
`
`Light absorbance in pulse oximetry
`
`55
`Pologe JAI 987 Pulse oximetry : technical aspects of machine design Int. Anestliesio!. Clinics 15
`(3) 137-53
`Shymoda Y und Yoshida I 1984 Effects of multiple scaltering and peripheral circulation un arterial
`oxygen saturation measured with a pulse-type oximeter Med. Bial. Eng. Comptit. 11 475-8
`Steinke .1 M and Shepherd A P 1986 Role of light scatteri ng in whole blood oxi mctry IEEE Trans.
`Binmed. Eng. 33 294-301
`Twersky V 1962 Multiple scaillering of waves and optical phenomena J. Opt. Soc. Am. 52 145-71
`Twersky V 1 970a Interface cffects in multiple scattering by large, low refracting, absorbing
`particles J. Opt. Soc. Ant. 60 908-14
`Twersky V 1970b Absorption and multiple scattering by biological suspensions J. Opt. Soc. Am.
`60 1084-93
`Wukitsch M W, Petterson M T, Tobler D R and Pologe J A 1988 Pulse oximetry: analysis of
`theory , technology, and practice J. C/in Monitoring 4 290-301
`Zijlstra W G, Buursma A and Meeuwsen-van der Roest WP 1991 Absorption spectra of fetal and
`adult oxylicmoglobin, de-oxyhemoglobin , corboxyhemoglobin , and methemoglobin Cim.
`Chem. 37 1633-8
`
`INSTRUCTIONAL OBJECTIVES
`
`4.1 Describe the properties and limitations of Beer's law.
`4.2 Describe different species of hemoglobin and their effect on the oxygenation of blood.
`4.3 Describe the functional and the fractional hemoglobin saturation and their difference.
`4.4 Describe the properties and assumptions of the spectrophotometric method to determine
`oxygen saturation in hemoglobin solutions.
`4.5 Describe the pnociples of pulse oximetry :ind what a pulse oximeter measureN.
`4.6 Describe why and how a pulse oximeter measures the ahsorbance in the arterial blood only.
`4.7 Describe tile normalization of the signals mid the reasons for this normalization.
`4.8 Explain how and why the ratioof the normalized signals is calculated,
`4.9 Explain errors in the spectrophotoinetric method when used for whole blood samples.
`4.10 Describe the different physical phenomena occurring when light travels through tissue and
`blood.
`4.11 Describe what light scattering is and where it occurs in pulse oximetry.
`4.12 Describe the influence of light scattering on the accuracy of a pulse oximeter.
`
`1
`h
`i
`e
`
`0
`ya
`11
`e
`
`11
`
`4.
`
`1
`
`1
`
`72
`
`g,
`/t.
`J'
`
`dn
`
`g15In,of
`
`W-
`tro
`:tal
`
`try
`
`MASIMO 2053
`Apple v. Masimo
`IPR2022-01299
`
`
`
`RX-0035.0073
`
`APL_MAS_ITC_00015690
`
`..
`
`CHAPTER 5
`LIGHT-EMITTING DIODES AND THEIR
`CONTROL
`Brad W J Bourgeois
`
`In order to make pulse oximetry practical in the modern medical environment, a
`light source is required that is powerful enough to penetrate more than a
`centimeter of tissue yet diminutive enough to fit in a small probe. Chapter 4
`shows that it also is desirable for the light source at each desired wavelength to
`have a very narrow emission spectrum, which minimizes error
`in the
`measurement of arterial oxygen saturation (Sa02)· Fortuitously, light-emitting
`diodes (LEDs) fulfill all the requirements for the light source in a pulse
`oxtmeter.
`pulse oximeter designers is how to deal with variations and shifts in the peak
`However, LEDs are not without drawbacks. The primary problem faced by
`wavelength of each LED . Because the main function of a pulse oximeter,
`measuring arterial oxygen saturation, is so heavily dependent upon accurate
`values for the two wavelengths of light. a design which does all it can to
`compensate for LED wavelength changes will outperform its competition.
`This chapter discusses important characteristics of LEDs, a LED driver
`circuit in a pulse oximeter, and various problems with the use of LEDs in pulse
`oximetry.
`
`5.1 AN INTRODUCTION TO LIGHT-EMITTING DIODES
`
`Light-emitting diodes are the light source of choice for all pulse oximeters on the
`market today. Their small size, excellent drive characteristics, and large light
`output over a very narrow ban