Untitled record

[0001] [Technical field to which the invention belongs] The present invention generally relates to a measurement system for an analysis object in a body fluid, and more particularly to a non-invasive measurement system for an analysis object in a body fluid. [0002] [Conventional technology] Patients with abnormal blood glucose levels are medically required to monitor their blood glucose levels on a regular basis. Abnormal blood glucose levels are caused by a variety of causes, including diseases such as diabetes. The purpose of monitoring blood glucose levels is to measure blood glucose levels and take appropriate action based on their levels to return them to the normal range. Failure to take appropriate measures can have serious consequences. When blood glucose levels drop abnormally (a condition known as hypoglycemia), the patient becomes tense, shivering, and confused. Such a person may lose judgment and faint. Abnormally high blood glucose levels (a condition known as hyperglycemia) also impair a person's health. Both hypoglycemic and hyperglycemic conditions are potentially life-threatening emergencies. [0003] The usual method of monitoring human blood glucose levels is inherently invasive. For example, blood is obtained from the fingertips using a puncture device to check blood glucose levels. Blood droplets are generated on the fingertips and the blood is collected by a test sensor. The test sensor is inserted into the test device and comes into contact with blood droplets. The test sensor draws blood into the test device, and the test device measures the glucose concentration in the blood. [0004] One problem with this type of analysis is some pain associated with fingertip puncture. Diabetics must regularly self-measure themselves several times a day. This includes cases where a separate puncture is required for each test and the user feels pain. In addition, each puncture causes a laceration in the user's skin, which, like a normal wound, takes longer to heal and is more susceptible to infections. [0005] Other methods of analyzing human blood glucose levels are inherently non-invasive. In general, this technique analyzes spectral information related to light transmitted through or reflected from human skin. This type of non-invasive analysis is excellent in that it is painless and does not cause skin lacerations. However, to date, such techniques have proven to be unreliable because they are unaware of many of the issues that affect their analysis. For example, in many non-invasive systems based on reflectance or transmission, the resulting spectral data includes glucose information from the part of the body tissue being analyzed as a whole and is not limited to blood glucose. Does not consider the facts. Other methods include instrument drift during or during analysis, temperature changes in the tissue under analysis, changes in the spectral characteristics of the tissue due to changes in pressure, etc. Does not consider irregularities. Such irregularities affect the quality of calibration models and algorithms used to measure the concentration of objects of analysis from non-invasively collected spectral data. Spectral data with these irregularities cannot be used in algorithms to determine the concentration of the object to be analyzed. [0006] [Problems to be Solved by the Invention] A reliable, non-invasive system for measuring objects to be analyzed in body fluids is needed. [0007] [Means for solving problems] Systems that measure the concentration of objects of analysis in at least one type of body fluid within body tissue include an infrared light source, body tissue interface, detector, and central processing unit. The body tissue interface is adapted to contact the body tissue and irradiate the contacting body tissue with light from an infrared light source. The detector is adapted to receive spectral information corresponding to infrared light transmitted through a part of the body tissue to be measured and convert it into an electrical signal indicating that the received spectral information is the received spectral information. .. The central processing unit is adapted to compare the electrical signal with an algorithm created based on the correlation with the object to be analyzed in the body fluid, and this algorithm transfers the received spectral information to the object to be analyzed in at least one body fluid. It is adapted to convert to the concentration of things. [0008] [Example] First, with reference to FIG. 1, a permeation-based non-invasive system 10 (System 10) for measuring objects to be analyzed in body fluids is functionally shown. Although the present invention will be described in the context of measuring glucose levels in a patient, the invention is applicable to the analysis of any analytical object in body fluids exhibiting spectral properties. In short, System 10 irradiates a portion of the skin with near-infrared light, eg, the "web" of the skin between the patient's index finger and thumb, and passes through the portion of the skin to measure the patient's glucose levels. And record the transmitted light. Usually, the patient's glucose level is called the patient's blood glucose level. However, when referring to the blood glucose level, it includes ignoring the amount of glucose contained in the extracellular substance and the intercellular substance of the patient. Therefore, the inventors of the present invention prefer to refer to the glucose level of a patient. [0009] About 50-60% of human skin is an intercellular substance, and the rest is an extracellular substance. Extracellular substances include about one-third of plasma (blood) and about two-thirds of interstitial fluid (ISF). Therefore, when examining the spectral properties of glucose from light transmitted through the patient's skin, it is important to consider the glucose as a whole in that portion of the skin, rather than just the glucose in the patient's blood. Most of the transmitted light is the light that has passed through the ISF, not the blood. Conversely, for example, in the case of an invasive test in which 10 μl of blood droplets are taken at the patient's fingertips, the measured glucose concentration primarily represents the concentration of glucose in the patient's blood. [0010] System 10 is used to obtain transmission spectrum information from the patient. For example, System 10 is used in tests where a subject's glucose concentration is adjusted to multiple different concentration levels. One such test is the glucose clamp test, in which a subject's glucose levels fluctuate to different levels throughout the test period. According to one embodiment, the glucose clamp test is designed to guide the subject's glucose levels to six constant regions distributed over a concentration range of 50-300 mg / dl. Each of the regions is separated by about 50 mg / dl so that each of the regions can be clearly distinguished. ISF and plasma samples are taken during the clamp test. Samples are taken every 5 minutes and glucose content is analyzed. This measurement regulates glucose or insulin infusion to maintain plasma glucose levels for about 25 minutes in a particular target constant region. Most commonly, spectral data obtained over the course of the test is compared to actual glucose levels (measured using invasive techniques). From this data, a calibration algorithm is assembled to predict the patient's actual glucose level based on the spectral characteristics of the light transmitted through the patient's skin. This calibration algorithm can be incorporated into the portable version of System 10 shown in FIG. [0011] Such portable devices allow the user to non-invasively monitor the user's glucose concentration level. The user brings the device into contact with his or her skin and obtains spectral information from the user's skin. The device can then give the user a measurement of the user's glucose concentration level after a short period of time. [0012] Returning to FIG. 1, the acoustic-optical tunable wavelength filter (AOTF) spectrometer is schematically shown by the dashed line 12. The AOTF spectrometer 12 outputs the monochromatic modulated light beam 14 to the optical fiber cable 16 via the lens 18. The AOTF spectrometer 12 includes a light source 20. In one embodiment, the light source 20 is a tungsten halogen light source, which is an inexpensive and stable light source that outputs a sufficient amount of light (275 watts, etc.). Alternative light sources include light emitting diodes (LEDs), dope fibers, including uranium-doped fibers, and laser diodes. The light source generates a light beam 22 in the near infrared region (wavelength 750 to 2500 nm). [0013] In general, the AOTF spectrometer 12 functions as an electronically tunable wavelength passband filter and outputs a monochromatic beam 14 having a wavelength within a desired range. The AOTF12 is a solid-state photoelectric device consisting of crystals 19 in which acoustic (vibration) waves at radio frequency (RF) are used to separate single wavelength light from a broadband light source. Wavelength selection is a function of the RF frequency applied to crystal 19. Crystals 19 used in AOTF devices can be made from many compounds. According to one embodiment of the invention, tellurium dioxide (TeO)₂) Crystals are used. TeO₂Crystals give good results when using light in the spectral region of 1200-3000 nm. In one embodiment, the crystals 19 are used in a non-identical arrangement and the sound waves and light waves (paths) that have passed through the crystals 19 have very different angles from each other. A transducer (not shown) is attached to one side of the crystal. This transducer produces vibrations (sound waves) when RF is applied to the transducer. When the sound wave from the transducer reaches the crystal 19, the crystal 19 alternately compresses and relaxes, and the refractive index fluctuates to operate like a transmission diffraction grating. However, unlike a classic grating, a crystal diffracts only one particular wavelength of light and acts more like a filter than a diffraction grating. The wavelength of the diffracted light is TeO₂It is determined by phase matching conditions based on the birefringence of the crystal, the velocity and frequency of the sound waves, and the parameters specific to the AOTF design. The wavelength selected varies by simply changing the frequency of the applied RF. The diffracted light is guided by two primary beams, which we call positive and negative beams. The remaining non-diffraction light passes as a non-diffraction zero-order beam. The two positive and negative beams are orthogonally polarized. The positive beam is sent to the optoid as described below, and the negative beam is used as a reference beam to correct fluctuations in the intensity of the light source or to correct the efficiency of the AOTF as described below. [0014] According to one embodiment, the light beam 14 output from the AOTF spectrometer has a resolution or bandwidth of about 4-10 nm. This bandwidth is swept (back and forth) over a wavelength range of approximately 1400-2500 nm. In other words, the AOTF spectrometer 12 outputs light with a wavelength that extends continuously over 1400-2500 nm and has a resolution of 4-10 nm. The timing of the sweep ranges from about 1 second to several seconds. A suitable AOTF spectrometer is Crystal in Palo Alto, California. It is marketed as OTF Model 2536-01 by Technologies Inc. The AOTF spectrometer includes an RF driver, mixer, and RF oscillator (not shown) to modulate the monochromatic beam light 14 at approximately 20,000 Hz. A voltage controlled oscillator (not shown) is controlled by frequency, modulation and power levels in the range 0-5.0 watts. A suitable voltage controlled oscillator is commercially available from Inrad Corporation in Northvale, NJ as Model DVCO-075A010. Power is sent to the acoustic converter to produce sound waves that change the properties of the birefringent crystal 19, so that the entire spectrum of light is separated into wavelengths associated with a particular frequency and the rest of the light is zero. It passes as the next light. [0015] Crystal 19 of the AOTF spectrometer 12 separates the light beam 22 into a first beam 14 and a second beam 23. The second beam 23 is guided to a reference detector 24 for measuring and recording the light shining on the skin. In addition, the reference detector 24 measures and records light 23 due to equipment drift associated with the light source and AOTF that can occur due to the length of operating time and fluctuations in the temperature of the equipment during that time. [0016] The light 14 output by the AOTF spectrometer 12 is guided to a lens 18, which reduces the diameter of the light beam and focuses the light beam 14 on one end of the fiber optic cable 16. The lens 18 effectively couples the AOTF spectrometer 12 to the fiber optic cable 16. The fiber optic cable 16 is a low OH (preferably about 0.3 ppm in silica ) fiber optic cable and has a high amount of attenuation over the length of the cable. The higher the OH, the higher the intrinsic absorbance of the fiber, especially in the wavelength region longer than 2100 nm. In another embodiment, the fiber optic cable has less than about 0.12 ppm OH. The quality of the light input to the fiber optic cable 16 substantially maintains the quality when it reaches the patient's skin at the other end 33 of the fiber optic cable 16. The output end 33 of the fiber optic cable 16 is connected to a device that the inventor calls the optoid 34. Generally, optoid 34 consists of hardware that acts as an interface to the patient's skin. The optoid 34 includes a first plate 46 and a second plate 48, which clamp the tissue to be analyzed, eg, the web 52 of the skin of the human hand between the index finger and the thumb. The optoid 34 includes a sapphire rod 42, which sends light from the fiber optic cable 16 to the web 52. The sapphire rod 42 has a diameter of about 3 mm in one embodiment, increasing the diameter of the light beam input to the web 52. Fiber optic cables are generally limited in diameter to about 2 mm. The wider diameter sapphire rod 42 allows beam diameters up to 3 mm and provides an effective means of bound light delivered to the skin. By sending a wide beam of light (eg, 3 mm of sapphire rod for 2 mm of fiber optic diameter), it covers a wider range of skin and limits the effects of slight irregularities in skin properties. The sapphire rod 42 is coplanar with the inner surface of the first plate 46. [0017] The light guided to the web 52 via the sapphire rod 42 passes through the web 52 and is guided to the second sapphire rod 54 (also 3 mm in diameter) arranged in the second plate 48. Light that has passed through web 52 is generally indicated by arrow 56. The amount of light transmitted through the web 52 is very small. Typically, less than about 2% of the light exiting the first sapphire rod 42 is incident on the second sapphire rod 54. The light transmitted through the web 52 is guided to the detector 58 by the second sapphire rod 54. In one embodiment of the invention, the detector 58 is an elongated indium gallium arsenide (InGaAs) detector, having a circular active surface with a diameter of 3 mm and being able to respond over a spectral region of 1300-2500 nm. Such detectors are commercially available from Hamamatsu Corporation. In one embodiment of the invention, the reference detector 24 and the detector 58 are similar types of detectors. Other types of detectors that can be used in other embodiments of the invention are indium arsenide (InAs), indium selenium (InSe), lead sulfide (PbS), mercury cadmium telluride (MCT), and DTG. It is a detector. Other detectors can also be used depending on the desired spectral region to be analyzed to measure glucose concentration levels. Glucose exhibits unique spectral properties in the spectral regions of about 1450 to 1850 nm and about 2200 to 2500 nm, as described in detail later in connection with FIG. The detector 58 generates an electrical signal indicating the detected transmitted light, which is processed as described in detail below. [0018] In addition to providing a mechanism for transmitted light to pass through the web 52, the optoid 34 performs other mechanical functions. First, the movable first and second plates 46, 48 (also called jaws) maintain a constant optical path through the web 52 by applying pressure to compress the web 52. Improve the integrity of the testing process by compressing the web 52. In one embodiment, plates 46,48 compress the skin tissue by about 6%. Also, by compressing the tissue, the air gap between the skin web and the plates 46, 48 is removed and an interface of the same height is formed between the skin web and the plates 46, 48, and the first sapphire rod The transmitted light from 42 is directly incident on the web 52. Optoid 34 includes a load cell 56 that measures the contact pressure on the skin web 52. During the analysis, pressure and temperature measurements are obtained and the anomalies associated with changes in pressure and temperature become apparent as described in more detail below. [0019] Second, the plates 46, 48 each include a thermoelectric heater (not shown), which heats the skin web 52 to a uniform temperature. In one embodiment of the invention, the thermoelectric heater heats the web to about 100 degrees Fahrenheit ± 0.1 degrees Fahrenheit. The thermoelectric heater built into each plate is capable of very accurate temperature control. Typically, the temperature difference between the skin surface and the interior is 5-7 degrees Fahrenheit. Heating the skin to a substantially uniform level reduces the scattering of light through the skin due to temperature gradients, resulting in a more stable analysis. In addition, heating the skin to about 100 degrees Fahrenheit dilates the capillaries, increasing blood flow within the capillaries by about 300%, resulting in more glucose being delivered to the area of analysis. [0020] As mentioned above, the AOTF 12 modulates the light beam 14 to produce a light beam to be modulated that passes through the skin via the optoid 34. Modulation helps solve some problems related to instrument drift that can affect the quality of spectral information. The modulated transmitted light is received by the detector 58, and the modulated transmitted light reaches the active substance of the detector 58 and is converted by the detector into a current indicating the amount of received light. In one embodiment, the electrical signal generated by the detector 58 is amplified by an amplifier (not shown), sent to the lock-in amplifier 70, and demodulated. In one embodiment of the invention, a suitable lock-in amplifier 70 is available from Stanford Research Instruments as a Model SR 810 DSP. Alternatively, the lock-in amplifier can be incorporated on an integrated circuit constituting the above-mentioned electronic hardware of the present invention. [0021] [0021] The demodulated signal is then digitized by an analog-to-digital (AD) converter 72. In one embodiment of the invention, the AD converter is a 16-bit converter, available from National Instruments Corporation, Austin, Texas, USA. Alternatively, digitization can be incorporated on an integrated circuit that includes the electronic hardware described above of the present invention. In another embodiment, digitization is performed at a bit rate of 18 bits or higher. [0022] The spectral data optionally removes high frequency noise through a high frequency filter 74 and then low frequency drift through a low frequency filter 78. This low frequency drift is caused by gradual changes in the patient's skin over the course of the measurement and drifts that can be observed in the instrument or fiber optics. By filtering the signal in this way, the overall signal-to-noise ratio can be improved. [0023] The signal is then sent to the central processing unit (CPU) 78. The CPU78 averages the received signal every minute and obtains about 500 data points over the course of about 500 minutes of testing. The data points are then stored in the memory of the CPU 78. The data is stored with tracking the wavelength of the light input to the optoid 14 and the corresponding spectral signal generated by the detector 58. Spectral signals are also stored over time in association with skin temperature, room temperature, pressure applied to the skin during measurement, and blood pressure values. This information is useful in determining whether any anomalies in the spectral signal are due to changes in these types of factors or changes in glucose concentration. The data is then processed to improve the signal-to-noise characteristics of the data and to remove artifacts that can degrade the quality of the spectral data. In another embodiment of the invention, the method of improving the signal-to-noise ratio can be done in various ways. For example, in one embodiment, the wavelet transform is used to remove high frequency noise and low frequency baseline drift noise (unrelated spectral variation determined by the information entropy corresponding to the glucose level) to result in a signal pair of signals. The quality of noise can be improved. In another embodiment, classical methods such as Savitzky-Golay multipoint smoothing can be used to improve signal vs. noise quality. In yet another embodiment, first-order differential analysis can also be used to address baseline problems such as baseline drift noise. [0024] Furthermore, noise in the signal can be improved by removing spectral information that does not correspond to the relevant glucose information according to another embodiment of the present invention. It is done by applying a genetic algorithm to select the wavelength region most relevant to glucose change and remove the other regions. This process leads to the development of robust calibration algorithms that significantly reduce overfitting problems. In yet another embodiment, Orthogonal Signal Correction (OSC) helps remove non-glucose spectral information from the signal. This method has been found to be useful in removing traces of temperature and time drift related changes in glucose-related data. By removing the data related to pressure and temperature changes over the course of the analysis, a better calibration algorithm is obtained, resulting in better glucose prediction based on the spectral data. By using a combination of a plurality of approaches, the signal can be further improved as compared with the case where different approaches are used individually. For example, it was confirmed that excellent results can be obtained by using the wavelet process and OSC together. Furthermore, the present inventor has confirmed that excellent results can be obtained by using the genetic algorithm in combination with OSC. [0025] Similarly, the reference detector 14 detects a light beam 23 representing the light 14 supplied to the optoid and generates a "reference signal". The reference signal is processed in the same manner as the signal output by the detector 58. [0026] Referring to FIG. 2, a graph showing the percentage of light passing through the web plotted against wavelength (nm) is shown. The peaks located at about 1450 to 1850 nm and about 2200 to 2500 nm in the graph show the high absorbance of light 56 transmitted through the tissue. The high absorbance in these spectral regions is partly due to absorption by the water contained in the skin. Glucose in the skin is mostly where the water in the skin is located. Glucose exhibits unique spectral characteristics in these two wavelength regions. [0027] As mentioned above, during the glucose clamp test, in addition to the transmission spectrum data, blood and a sample of ISF were obtained from the subject (ie, the patient under test) to obtain the subject's actual blood glucose. The level is measured. In one example of a glucose clamp test, the test is performed over a length of about 500 minutes. Blood and ISF samples are obtained approximately every 5 minutes, for a total of approximately 100 samples. These values are interpolated over a 500 minute test period to give about 500 glucose concentration values. [0028] The transmitted light digital spectral signal is averaged and stored every minute, resulting in approximately 500 data points over the test period. This data is analyzed and processed (detailed below) to create a calibration algorithm for predicting the actual glucose concentration level from inspection of the spectral characteristics of transmitted light. [0029] As shown in FIG. 3, a graph plotting the predicted glucose value (obtained from the spectral characteristics of transmitted light) against the measured glucose value is shown. As shown in the graph of FIG. 3, there is a high correlation between the predicted glucose concentration and the measured glucose concentration. [0030] In order to obtain the predicted values plotted in FIG. 3, a calibration algorithm must be created to predict the glucose concentration from the transmission spectrum signal (the signal generated by the detector 58). After the spectral signal is filtered by the high and low frequency filters 74,78, the signal is normalized by the spectral scattering of light transmitted through the web 52 and the action of the pressure of the optoid 34 clamping the web 52. Correct the resulting spectral signal variation. Without correction for these fluctuations, glucose-related spectral information can be obscured. As mentioned above, less than about 2% of the light input to the skin web 52 penetrates into the detector 58. Therefore, it is necessary to clarify these types of fluctuations and anomalies that can lead to errors. The unprocessed signal from the AOTF spectrometer described above is first normalized to a constant energy and then removed a certain region of the spectrum around the mean to create a normalized and preprocessed spectral sequence. , This spectral sequence is checked for outliers by known standard methods. In addition, pretreatment with OSC reduction and wavelet analysis filtering enhances the glucose signal and reduces moisture and other background noise. The resulting spectrum set is calibrated using partial least squares (PLS) regression using Venetian blinds cross-validation in some or all of the above data. Another embodiment of the above data preparation includes general methods for reducing or eliminating background signals, including first-order differential smoothing, second-order differential smoothing, wavelength selection by genetic algorithms, wavelet processing, and principal components. It is not limited to component analysis. Another embodiment for generating a calibration model can be achieved by many different forms of regression techniques, ridge regression methods, and conventional (inverse) least squares regression. [0031] A calibration algorithm for predicting glucose concentration is then created from the normalized signal. The orthogonal signal correction process combines time-related temperature and pressure information to recognize spectral parts that are associated with these factors and not strictly related to changes in glucose concentration. This process is used in combination with correlation data (invasively measured plasma and ISF fluid glucose levels) to remove spectral data information that is associated with changes in other measurements rather than changes in glucose. The result is a less calibrated algorithm that is more clearly associated with changes in glucose concentration and is associated with artifacts that are accidentally correlated with glucose concentration. Other data improvement processes include the use of more common chemometrics such as genetic algorithms and wavelet analysis, and further improve spectral information to the most efficient information. Genetic algorithms and wavelet analyzes can select wavelengths that are particularly relevant to glucose in the spectrum that focus the calibration algorithm on specific changes in glucose concentration. The wavelength is selected based on the region of the spectrum in which the largest glucose-related peak is located, but the wavelength is also selected based on the region of the spectrum associated with the change in the refractive index of the tissue due to changes in tissue concentration. This wavelength selection process retains wavelength information to generate the best calibration algorithm. [0032] With reference to FIG. 4, a flowchart showing a method for generating a glucose calibration algorithm according to one embodiment of the present invention is shown. First, as described above, a glucose clamp experiment is performed to obtain spectral information from at least the body tissues of the first and second subjects. This information is stored in the first data set 82 and the second data set 83. In one embodiment, the first and second datasets 82,83 each contain spectral information obtained from multiple subjects. Other information, such as body tissue temperature, pressure applied to tissue, and invasively measured glucose concentration levels, is obtained from each subject during the glucose clamp test at predetermined intervals. [0033] Combined datasets (first and second spectral datasets 82,84) consisting of spectral data from one or more subjects were created to provide a useful model for predicting glucose levels in all subjects who provided the data. Used to generate. The unprocessed signals stored in the first and second datasets 82,84 obtained from the AOTF spectrometer described above are first normalized to a constant energy for the data from each subject in step 84. .. Part of the data for each subject was combined in step 85 to form a single combined spectrum set, which was averaged in step 86 to remove certain regions of the spectrum, normalized and pre-normalized. A processed spectral sequence is created and the spectral sequence is checked for outliers by known conventional methods. In addition, pretreatment with OSC reduction and wavelet analysis filtering enhances the glucose signal and reduces moisture and other background noise. In step 87, using the resulting spectral set, partial least squares (PLS) regression creates a calibration model using Venetian blinds cross-validation in some or all of the above data. Another embodiment of the above data preparation includes general methods for reducing or eliminating background signals, including first-order differential smoothing, second-order differential smoothing, wavelength selection by genetic algorithms, wavelet processing, and. It is not limited to principal component analysis. Another embodiment for generating a calibration model includes many different forms of regression techniques, principal component regression methods, ridge regression methods, and the usual (inverse) least squares method. [0034] The PLS model generated in step 87 is applied to the orthogonal signal correction and normalized first dataset in step 89, and the glucose calibration algorithm is obtained in step 90. The glucose calibration algorithm 90 is used to predict the glucose concentration based on the spectral information obtained from the subject. In other words, the glucose calibration algorithm can determine a subject's glucose concentration based on spectral information (transmission or reflection spectral information). The glucose calibration algorithm 90 is then applied to the orthogonal signal corrected and normalized second spectrum dataset in step 91, and the glucose concentration level of the subject in the second step dataset 83 is predicted in step 92. .. The glucose concentration level predicted in step 92 is then compared to the glucose concentration measured invasively during the glucose clamp test in step 93 to check the accuracy of the glucose calibration algorithm. [0035] In another embodiment of the invention, when creating the glucose calibration algorithm, OSC step 88 is followed by wavelet analysis on each dataset and the data is filtered. In yet another embodiment of the invention, the spectral datasets 82,83 include modeled spectral data for concentration levels that are outside the range of glucose concentration levels obtained during the glucose clamp test. In one embodiment, the AOTF spectrometer 16 can be used to generate out-of-range spectral data obtained during the glucose clamp test. [0036] With reference to FIGS. 5a and 5b, a reflex-based, non-invasive system 90 (system 90) for measuring objects to be analyzed in body fluids is functionally shown. Briefly, the system 90 measures a patient's glucose level by irradiating a portion of the skin, such as the patient's forearm, with near-infrared light and recording the amount of reflected light from the skin. [0037] A monochromatic light beam is input to bundle 100 of the optical fiber cable 101. In FIG. 5b, bundle 100 shows two concentric circles or rows of fiber optic cable 101, but an appropriate number of rows of fiber optic cables can be used. The monochromatic beam is generated in the same manner as described with respect to FIG. An AOTF spectrometer (not shown) outputs an optical beam 94 with a resolution of 4-10 nm and is input to the fiber optic cable bundle 100 while sweeping (back and forth) over a frequency range of approximately 2200-4500 nm. Fiber optic cable bundle 100 sends optical 94 to optoid 104. Optoid 104 constitutes the hardware that binds to the patient's skin. The optoid 104 includes a plate 106 with a window 108. The light 102 is guided to the patient's skin 110 through the window 108. In one embodiment, the window 108 is a sapphire window. [0038] When used, the optoid 104 contacts the skin 110, such as the patient's forearm, which is supported by the plate 106 and the window 108. The light 102 is guided to the skin 110 through the window 108. Light reaches a depth of about 300 microns of skin 110 and is then reflected from inside skin 110. The reflected light is indicated by arrow 112. The reflected light 112 is guided to the detector 114 via the sapphire rod 116 disposed in the optical fiber bundle 100. The reflected light 112 is detected by the detector 114 in the same manner as the transmitted light 56 described with respect to FIG. [0039] In another embodiment of the reflective-based non-invasive system 90, only part of the fiber optic cable 101 is used to direct light to the optoid 104, resulting in varying lengths of the optical path delivered. .. For example, in one embodiment only the inner ring of the fiber optic cable 101 is used and in another embodiment only the outer ring is used. By changing the optical path length, reflected light from various depths in the tissue can be sampled. In some other embodiments, changes in optical path length can be utilized to correct individual tissue properties such as scattering. [0040] The optoid 104 of the non-invasive system 90, which is based on the reflection type, controls the temperature of the skin area that captures the reflection signal. In one embodiment of the invention, the optoid plate 106 comprises a thermoelectric heater for heating the skin to about 100 degrees Fahrenheit ± 0.1 degrees Fahrenheit. Again, heating the skin to a uniform temperature reduces light scattering, which is a function of temperature. Furthermore, as described above, heating of the skin dilates the capillaries and increases blood flow in the blood vessels by about 300%. [0041] In one embodiment of the invention, the refractive index matching material 112 is disposed between the skin 110 and the sapphire window 108 to keep the refractive index of the light 102 guided to the skin 110 and the reflected light from the skin 112 constant. Maintain a consistent state. Refractive index matching gels reduce the large changes in refractive index that normally occur between the skin and the air layer. Such large changes result in Fresnel loss, which is especially noticeable in reflection-based analysis, resulting in large changes in the spectral signal. In one embodiment of the invention, the refractive index matching material 112 is a chlorofluorocarbon gel. This type of index material has some advantageous properties. First of all, the effect of the chlorofluorocarbon gel on the spectral signal passing through it is extremely small. Second, the index matching material has a high fluid temperature point, which allows it to maintain a gel state during analysis and under test conditions. Third, because the gel is hydrophobic, it seals the sweat glands and prevents sweat from fogging the sapphire window 108 (gas-liquid formation). Fourth, this type of index matching material is not absorbed by the stratum corium during the analysis. [0042] The output of the detector 114 is filtered and processed in the same manner as described with the transmissive based system 10 described above. [0043] With reference to FIG. 6, a graph of absorbance vs. wavelength of light incident on the skin is shown. As can be seen in FIG. 6, high absorbance is observed in the spectral regions of 1350 to 1600 nm and 1850 to 2100 nm. [0044] Calibration algorithms for reflective-based systems are created using similar data processing techniques described for transmissive-based systems. FIG. 7 shows a graph of glucose concentration measured paired with (invasively obtained) glucose concentration predicted using a calibration algorithm. As in the case of the transmission-based system 10, the reflection-based system 90 can accurately predict the glucose concentration of the subject. [0045] Although the present invention has been described above with one or more specific embodiments, it will be appreciated by those skilled in the art that many modifications can be made without departing from the spirit and scope of the invention. Each of these examples and obvious modifications is intended to fall within the spirit and scope of the claimed invention. [Simple explanation of drawings] FIG. 1 is a functional block diagram of a transmission-based system for measuring an object to be analyzed in a body fluid according to an embodiment of the present invention. FIG. 2 is a graph showing the absorbance of transmitted light vs. the wavelength of transmitted light according to an embodiment of the transmission type-based system shown in FIG. FIG. 3 is a graph showing predicted glucose concentration vs. measured glucose concentration according to an embodiment of the permeation type based system shown in FIG. FIG. 4 is a flowchart showing a method for creating a glucose calibration algorithm according to an embodiment of the present invention. FIG. 5a is a functional block diagram of a reflex-based system for measuring objects to be analyzed in body fluids according to an embodiment of the present invention. 5b is a cross-sectional view taken along line 5b of FIG. 5a. FIG. 6 is a graph showing the absorbance of reflected light vs. the wavelength of reflected light according to an embodiment of the system based on the reflection type shown in FIG. 5a. FIG. 7 is a graph showing predicted glucose concentration vs. measured glucose concentration according to an embodiment of the reflection type based system shown in FIG. 5a. [Explanation of symbols] 10 Permeable-based non-invasive measurement system 12 AOTF spectrometer 14 1st light beam 16 fiber optic cable 18 lens 23 2nd light beam 34 Optoid 46 1st plate 48 2nd plate 52 skin web 54 Sapphire rod 56 load cell 58 detector 90 Reflective-based non-invasive measurement system 100 fiber optic bundle 101 fiber optic cable 102 light beam 104 Optoid 106 plate 108 windows 110 skin 112 Refractive index matching material 114 detector 116 Sapphire rod