Compensation of human variability in pulse oximetry

Abstract

The invention relates to a method of calibrating a pulse oximeter, in which the effects caused by tissue of a subject can be taken into account. A detector output signal is measured when living tissue of the subject is present between emitters and the detector in a sensor. Nominal calibration and nominal calibration characteristics are read from a memory, whereupon values for the same nominal characteristics for the sensor on living tissue of the subject are established using the detector output signal. Then, changes in the nominal calibration characteristics induced by the living tissue are calculated and a subject-specific calibration is formed by correcting the nominal calibration with the changes. Finally, the hemoglobin fractions are solved using the corrected nominal calibration. The invention also relates to a pulse oximeter having pre-stored data in a memory comprising data of initial characterization measurements, data of nominal characteristics describing calibration conditions under which a predetermined calibration of the apparatus has been applied, and data of nominal calibration and nominal calibration characteristics. An extinction coefficient compensation block is operatively connected to the first signal processing means and to the memory for reading data, said block comprising first calculation means adapted to correct the nominal characteristics of the sensor on living tissue of the subject. A transformation compensation block is operatively connected to the first signal processing means for receiving the DC signals and to the memory for reading data, said block comprising second calculation means adapted to correct the transformation values based on the changes in the DC signals induced by tissue of the subject. Alternatively, said data may be stored in the sensor part of the pulse oximeter.

1 . A method for compensating for subject-specific variability in a pulse oximeter intended for non-invasively determining in in-vivo measurement the amount of at least two light-absorbing substances in the blood of a subject and provided with emitters for emitting radiation at a minimum of two different wavelengths and with a detector for transforming the radiation received into an electrical output signal, the method comprising the steps of 
 measuring detector output signal when living tissue of the subject is present between the emitters and the detector in a sensor, wherein the detector output signal depends on tissue,    reading a nominal calibration and nominal calibration characteristics from a memory,    establishing values for the same nominal characteristics for the sensor on living tissue of the subject using the detector output signal,    calculating changes in the nominal calibration characteristics induced by the living tissue,    forming a subject-specific calibration by correcting the nominal calibration with the changes,    solving the hemoglobin fractions using the corrected nominal calibration.    
     
     
         2 . The method as in  claim 1 , wherein the nominal calibration characteristics include factors of external origin.  
     
     
         3 . The method as in  claim 1 , wherein the nominal characteristics include factors of internal origin.  
     
     
         4 . The method as in  claim 2 , wherein the factors of external origin are associated with calculation of numeric values of extinction coefficients in the Lambert-Beer model.  
     
     
         5 . The method according to  claim 1 , wherein determination of wavelength shifts in relation to the nominal wavelengths are incorporated into said establishing step.  
     
     
         6 . The method according to  claim 5 , wherein the effect of wavelength shift is derived from the transmitted signal through the tissue.  
     
     
         7 . The method according to  claim 5 , wherein effects of the wavelength shifts are incorporated into calculation of subject-specific extinction coefficients.  
     
     
         8 . The method according to  claim 3 , wherein effects of the internal factors are used in defining a subject-specific transformation that is used to transform in-vivo measurement results to the Lambert-Beer model.  
     
     
         9 . The method of  claim 8 , wherein the internal factors affecting the transformations of the modulation ratios from the in-vivo values to the Lambert-Beer values are derived based on the transmitted signal through the tissue.  
     
     
         10 . The method according to  claim 4 , wherein tissue filter effect is corrected according to the equation  
       
         
           
             
               
                 E 
                 Eff 
               
               = 
               
                 
                   E 
                   kl 
                   0 
                 
                 ⊗ 
                 
                   ( 
                   
                     1 
                     + 
                     
                       S 
                       · 
                       
                         ( 
                         
                           
                             Tissue 
                             SHIFT 
                             
                               SLOPE 
                               = 
                               1 
                             
                           
                           - 
                           1 
                         
                         ) 
                       
                     
                   
                   ) 
                 
               
             
           
         
       
       where E kl   0  is included in the nominal calibration, S denotes the column array in Eq. 8 and the matrix multiplications are performed element by element ({circle over (x)}) or element by row (•), respectively.  
     
     
         11 . The method according to  claim 4 , wherein external temperature and the LED drive power induced wavelength shift is corrected according to the equation  
       
         
           
             
               
                 E 
                 TEMP 
                 EFF 
               
               = 
               
                 
                   E 
                   kl 
                   0 
                 
                 ⊗ 
                 
                   ( 
                   
                     1 
                     + 
                     
                       
                         ( 
                         
                           Δ 
                           ⁢ 
                           
                               
                           
                           ⁢ 
                           
                             λ 
                             / 
                             5 
                           
                           ⁢ 
                           
                               
                           
                           ⁢ 
                           nm 
                         
                         ) 
                       
                       · 
                       
                         ( 
                         
                           
                             Temp 
                             SHIFT 
                             
                               
                                 Δ 
                                 ⁢ 
                                 
                                     
                                 
                                 ⁢ 
                                 λ 
                               
                               = 
                               
                                 5 
                                 ⁢ 
                                 
                                     
                                 
                                 ⁢ 
                                 nm 
                               
                             
                           
                           - 
                           1 
                         
                         ) 
                       
                     
                   
                   ) 
                 
               
             
           
         
         where Δλ is the array in Eq. 11 and  
         
           
             
               
                 Temp 
                 SHIFT 
                 
                   
                     Δ 
                     ⁢ 
                     
                         
                     
                     ⁢ 
                     λ 
                   
                   = 
                   
                     5 
                     ⁢ 
                     
                         
                     
                     ⁢ 
                     nm 
                   
                 
               
             
           
         
          is as in equation 12.  
       
     
     
         12 . The method according to  claim 8 , wherein 
 the nominal calibration characteristics include nominal values for the Functional Light Transmission (FLT) of the apparatus,    the establishment step includes calculating new values for the Functional Light Transmission (FLT) of the apparatus based on tissue-induced changes, and    the subject-specific transformation is established on the basis of the nominal and new values    
     
     
         13 . The method according to  claim 8 , wherein 
 the nominal calibration characteristics include nominal values for function F kl  of the apparatus,    tissue-induced changes includes calculating new values for the function F kl  of the apparatus, and    the establishment step includes determining the subject-specific transformation on the basis of the nominal and new values,    wherein the function F kl  corresponds to the ratio                    f   a     ⁡     (       μ   a   k     -     μ   v   k       )       +     μ   v   k             f   a     ⁡     (       μ   a   l     -     μ   v   l       )       +     μ   v   l         ,            where μ v  and μ a  are the absorption coefficients of venous and arterial blood, respectively, as determined in the Lambert-Beer domain, f a  is the volume fraction of arterial blood, and the superscripts k and l indicate the wavelength.    
     
     
         14 . The method according to  claim 13 , wherein the nominal and new values for the Function F kl  are calculated on the basis of measured fluctuation of the DC component of the radiation received by the detector.  
     
     
         15 . The method according to  claim 1 , wherein the at least two light absorbing substances include oxyhemoglobin (HbO2) and reduced hemoglobin (RHb).  
     
     
         16 . The method according to  claim 1 , wherein all the method steps are performed at each heart beat of the subject.  
     
     
         17 . A pulse oximeter for non-invasively determining in in-vivo measurement the amount of at least two light absorbing substances in the blood of a subject, the pulse oximeter having an interface to a sensor for receiving an output signal at at least two distinct wavelengths and controlling operation of the sensor, the pulse oximeter further comprising 
 signal processing means for extracting AC and DC signals for each wavelength, each wavelength signal representing the absorption caused by the blood of the subject,    pre-stored data in a memory, comprising    data of initial characterization measurements,    data of nominal characteristics describing calibration conditions under which a predetermined calibration of the apparatus has been applied, and    data of nominal calibration and nominal calibration characteristics,    an extinction coefficient compensation block operatively connected to the first signal processing means, and to the memory for reading data, said block comprising first calculation means adapted to correct the nominal characteristics of the sensor on living tissue of the subject,    a transformation compensation block operatively connected to the first signal processing means for receiving the DC signals and to the memory for reading data, said block comprising second calculation means adapted to correct the transformation values based on the changes in the DC signals induced by tissue of the subject, and    a hemoglobin fraction calculation unit connected to the extinction coefficient compensation block and the calibration values compensation block, said unit comprising means for applying the corrected nominal characteristics and the corrected nominal calibration values in solving the hemoglobin fractions.    
     
     
         18 . The pulse oximeter as in  claim 17 , wherein the pre-stored data include nominal transformation values of the Lambert-Beer model.  
     
     
         19 . The pulse oximeter as in  claim 17 , wherein the pre-stored data include nominal calibration values of the Lambert-Beer model.  
     
     
         20 . The pulse oximeter as in  claim 17 , wherein the extinction coefficient compensation block and the transformation compensation block are time synchronized by synchronization means and the nominal calibration is compensated on heart beat-to-beat basis.  
     
     
         21 . A sensor for collecting measurement data for a pulse oximeter intended for non-invasively determining in in-vivo measurement the amount of at least two light absorbing substances in the blood of a subject, the sensor comprising 
 at least one light emitter for emitting light at a minimum of two different wavelengths,    a light detector for receiving said light at each of said wavelengths and producing an electrical output signal indicating the received light at the said wavelengths,    a pulse oximeter readable memory    data pre-stored in the memory, comprising 
 data of initial characterization measurements,  
 data of nominal characteristics describing calibration conditions under which a predetermined calibration of the pulse oximeter has been establihed, and  
 data of nominal calibration values,  
 wherein said data allows the pulse oximeter connected to the sensor to determine tissue-induced changes in the nominal characteristics when light is propagated through said tissue.  
   
     
     
         22 . A sensor as in  claim 21 , further comprising a temperature sensitive element operatively connected to the light emitter for producing electrical signals responsive to the temperatures of the semiconductor junction.  
     
     
         23 . The sensor according to  claim 21 , wherein the light emitters are Light Emitting Diodes.  
     
     
         24 . The sensor according to  claim 21 , wherein the light emitter is a laser.  
     
     
         25 . The sensor according to  claim 21 , wherein the light emitter is light conduction means for conducting radiation from the emitting component to the tissue site, at which the measurement is performed.  
     
     
         26 . The sensor according to  claim 21 , wherein the means for receiving radiation is radiation conduction means for conducting radiation from the tissue site to the detector component.  
     
     
         27 . The sensor according to  claim 21 , wherein the data of nominal characteristics includes the Functional Light Transmission (FLT) of the pulse oximeter.  
     
     
         28 . The sensor according to  claim 21 , wherein the data of nominal calibration values includes function F kl  of the pulse oximeter in nominal conditions, 
 wherein the function F kl  corresponds to the ratio                    f   a     ⁡     (       μ   a   k     -     μ   v   k       )       +     μ   v   k             f   a     ⁡     (       μ   a   l     -     μ   v   l       )       +     μ   v   l         ,            where μ v  and μ a  are the absorption coefficients of venous and arterial blood, respectively, as determined in the Lambert-Beer domain, f a  is the volume fraction of arterial blood, and the superscripts k and l indicate the wavelength.