TW200538083A - Method of sample control and calibration adjustment for use with a noninvasive analyzer - Google Patents

Abstract

A method and apparatus for easing the use of an optically based noninvasive analyzer is presented. More particularly, a simplified algorithm is used that removes the daily requirement of collecting and using a noninvasive spectrum to update a calibration model. In another embodiment, a guide is used to substantially reduce variation in sample probe placement in relation to a skin tissue sampling site, resulting in the ability to maintain calibration performance with the use of a reference analyte concentration, with or without the use of a reference spectrum collected nearby in time.

Description

200538083 (1) Description of the invention: [Technical field to which the invention belongs] The present invention relates to a non-invasive method for measuring an analyte in a human body. More specifically, the present invention relates to methods for calibrating and predicting analytes, which can adjust state changes that cannot be compensated in a static calibration mode. [Prior technology] Non-invasive Non-invasive analyzer designed based on spectral analysis can transfer external energy in the form of light to a certain area of the body, where photons can interact with the chemical composition and physiological state of the sampled tissue . Part of the incident photons will be scattered or conducted outside the body's detection area. The information about the incident photon and the measured photon can be used to understand the chemical and / or structural basis of the sampling site. The special advantages of non-invasive systems include the use of a small amount of consumables to perform an in vivo chemical concentration determination in a painless manner without creating a biological hazard. In addition, this technique allows the concentration of multiple analytes to be determined simultaneously. Common non-invasive analyzers are magnetic resonance imaging machines (MRI), X-rays, pulse oximeters, and non-invasive glucose analyzers. With the exception of X-rays, these measurements are performed at relatively harmless wavelengths of radiation. The examples here focus on non-invasive glucose concentration measurement, but the principles can be applied to the detection of other analytes, such as lipid splashes, proteins, water, and blood or tissue components. Diabetes 3 200538083 Diabetes is a chronic disease, which is caused by imbalance in the production and utilization of Tengdaosu. Insulin is a hormone that helps cells take up grape bran. Although the exact etiology of diabetes is unknown, genetic factors, environmental factors, and obesity seem to play important roles ^ Diabetes increases the risk of developing three major types of disease: cardiovascular heart disease, visual network disease, and neuropathy. Diabetes usually has one or more of the following complications: heart disease and stroke, two blood pressure, kidney disease, neuropathy (neuropathy and resection), retinopathy, diabetic ketoacidosis, skin disease, gum disease, impotence, and fetal complications disease. Diabetes is a major cause of death and disability around the world. In addition, diabetes is only one of a variety of glucose metabolism abnormalities including abnormal glucose tolerance, hyperinsixlinemia, or hypoglycemia. The prevalence and trend of diabetes Diabetes is a very common disease. The World Health Organization (WHO) estimates that 154 million people worldwide currently suffer from diabetes. Of these, 54 million people with diabetes live in developed countries. It is estimated that by 2015, the number of people with diabetes will grow to 300 million. It is estimated that approximately 15.7 million people or 5.9% of the total population in the United States have diabetes. Within the United States, 'diagnosed adults had an increase of 6% in 1999' and an increase of 33% between 1990 and 1998. This phenomenon is consistent with the approximately 800,000 cases added each year in the United States. The total expenditure of the U.S. economy each year here ’is estimated to exceed $ 90 billion (Diabetes National Institutes of Health, Publication No. 98-3926 200538083

Bethesda, MD (November 1 997)) 〇 Long-term clinical studies show that proper control of blood glucose concentration can effectively inhibit the occurrence of complications (The Diabetes Control and Complications Trial Research Group, The effect of intensive treatment of diabetes on the development and progression of long-term in insulin-dependent diabetes mellitus, N Eng J of Med, 329: 977-86 (1 993); UK

Prospective Diabetes Study (UKPDS) Group, Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes, Lancet, 352: 837-853 (1998); W & Y. Ohkubo, H .Kishikawa, E. Araki, T. Miyata, S. Isami, S. Motoyoshi, Y. Kojima, N · Furuyoshi, M. Shichizi, Intensive insulin therapy prevents the progression of diabetes microvascular complications in Japanese patients with non-lnsulin-dependent diabetes mellitus: a randomized prospective 6-year study, Diabetes Res Clin Pract, 28: 1 03- 1 1 7 (1 995)). An element of diabetes management is the self-monitoring of blood glucose levels by a diabetic in a home environment. However, the existing monitoring technology requires blood to be drawn before analysis, which causes inconvenience and pain, which hinders the frequent use of blood glucose monitoring technology (The Diabetes Control and Complication Trial Research Group, wpra). Therefore, non-invasive glucose measurement is considered a very beneficial development in diabetes management. It is also currently expected to exhibit 5 200538083 to display the implantable glucose that can be coupled with 縢 shimasu to prepare μ ± X delivery system as an artificial pancreas. There are many types of sampling methods. Technology can be broadly divided into two categories: invasion technology requires the acquisition of any measuring device by itself and the intrusion front method. • Invasiveness: An example of analysis of whole blood, serum, plasma, interstitial fluid, a mixture of said samples, or selective sampling of components. In vivo glucose concentration measurement using invasive biotechnology. Electrochemical, electroezymatic, and / or colorimetric methods are used to separate samples. For example, 'enzyme and colorimetry can be used to determine the glucose concentration in tissue. • Non-invasive: A variety of techniques have been developed to measure glucose concentration in biological samples using spectrophotometry. These techniques include R.aman and fluorescence spectrum analysis, as well as some techniques that use light between the infrared wavelengths [ultraviolet (200 to 400 nm), (400 to 700 nm), and near infrared (700). To 2500 nm or to 4000 cm · 1), and infrared (2500 to 1, 4,285 nm to 700 cm-1)] instrument, which can be used to analyze chemical incorporation and non-invasiveness in vivo. In this article, when a biological sample is to be analyzed, some parts must penetrate the body. These, when the technology is classified. The classification of enzymes and the enzymes are traditionally analyzed by enzymes. Sample internal method: Raman outside line to visible light 14,286 or 40000538083 Non-invasive glucose measurement has already existed a variety of methods. Non-invasive determination of glucose concentration. Consumption of these methods is very different, but they all have at least two co-synchronous steps. First, in the case where it is not necessary to obtain a biological sample, a signal is obtained from the body using the device. No.] 'This signal is converted into a glucose measurement value using an algorithm. A class of non-invasive glucose analyzers are designed based on the principle of spectroscopy. Traditionally, non-invasive devices use some form of spectrometer to obtain the signal or spectrum of a part of the body. Spectral analysis techniques used include Raman and fluorescence, and technologies that use light between the ultraviolet and infrared wavelengths [ultraviolet (200 to 400 nm), visible light (400 to 700 nm), and near infrared (700 to 700 nm) 2500 nm or 14,286 to 4000 cm.1), and infrared (2500 to 14,285 nm or 4000 to 700 cnT1)). When the diffuse reflectance mode is used for non-invasive glucose measurement, its specific wavelength range is between about 1100 to 2500 nm or a wavelength range within this range (κ · Hazen, Glucose Determination in Biological Matrices Using Near-Infrared 3pectroscορχ? Doctoral dissertation, University of Iowa (W5).) Modes There are generally three modes that can be used to collect non-invasive spectra: transmittance, reflectance, and / or diffuse reflectance. For example, the collected signal is usually light or spectrum. After it penetrates an area in the body (such as a fingertip), 200538083 may diffuse reflection or transflect. Here, rate refers to the anti-collection at the point of incidence or the area of incidence and the sample (that is, the signal collected is not a diffuse reflectance and a signal collected in a body where the transmittance and diffuse reflectance are collected. For example Reflected light enters from the fingertips or anterior region and exits in another area 5 mm or more away from the light depending on the wavelength used. Therefore, absorbable light (such as near the maximum absorption wavelength of water) 1450 or light) will be collected after a small amount of divergence, while relatively light (such as light with a minimum absorption wavelength of 1 300, 1 600, or near water) will be at a lighter location farther from the incident photon or The location of the non-invasive technique is not limited to the different locations of fingertip invasive measurement, including: hand, finger, base of palm, forearm, back of forearm, upper arm, inside of upper arm, head, eyes, tongue, chest, Torso, abdominal area, thigh, foot, sole area, and toe. Although the instrument focuses on optical non-invasive analysis, non-invasive technology is not necessary Based on spectral analysis. Lifting impedance analyzer can also be regarded as a non-invasive device. The signal transmission rate is transmitted on the reflective surface, and a certain point of the arm enters an area 0.2 into the body. 1950 nm is not easily absorbed or 2250 nm penetrates the reflection site. Non-regional, thumb, earlobe, calf, foot, and instrument can be performed, but it is necessary to note that in the present invention, 200538083, any signal can be read by the body without penetrating the skin and collecting biological samples. The device is then called a non-invasive glucose analyzer. For example, a bio-impedance analyzer is a non-invasive device. When a near-infrared analyzer is used for non-invasive glucose measurement, a near-infrared (NIR) electromagnetic radiation is usually used to irradiate a small area of the body. The near-infrared wavelength used is between 7000 to 2500 nm or one or more of the above wavelength ranges, such as 1100 to 1 800 nm. The incident light is reflected back to the optical connection. Before reaching a detector or a light collection device directly connected to a detector, light will be locally absorbed and scattered due to interaction with tissue components. The detected light contains quantitative information that reflects known interactions between incident light and body tissue components, including water, fat, protein, and glucose. Non-invasive glucose concentration analyzers can have one or more light paths from the light source to the detector. The types of light sources include black body light sources, toothed crane light sources, one or more LED light sources, and one or more laser diodes. For multi-wavelength spectrophotometers, a wavelength selection device or -series optical filters can be used for wavelength selection. The wavelength selection device includes one or more gratings, chirps, and filters with wavelength selectivity. .Can also use different light sources (such as LED or diode to change the light source) for wavelength selection. The detector can be in the form of one or more unitary detectors, or one or more detectors in the form of an array or a beam. Detector types include InGaAs, PbS, PbSe, gi, HgCdTe, MCT or similar. The detector array includes Shishen 200538083 Indium gallium sulfide, sulfide, selenium alloy, silicon, mercury-cadmium alloy or the like. Different shapes of light collection optics, such as optical fibers, lenses, and mirrors, are often used as output elements in a spectrophotometer to guide light from the light source and through the sample to the detector. Calibration The glucose concentration analyzer needs to be calibrated. This is true for all types of glucose analyzers, such as traditional, alternative, non-invasive and implantable analyzers. The fact related to non-invasive glucose analyzers is that they are secondary 'in nature, i.e. they do not directly measure blood glucose concentrations. This means that a rudimentary method is needed to calibrate these devices in order to properly determine the blood glucose concentration. Various calibration methods are available. A non-invasive technique called "near-infrared spectroscopy" has been developed that requires a mathematical relationship that can relate near-infrared measurements in the body to actual blood glucose concentrations. The blood glucose concentration can be determined in advance using a glucose concentration analyzer such as HEMOCUE (YSI INCORPORATED, Yellow Springs OH) or any appropriate and accurate traditional invasive reference device, such as THERASENSE EREESTYLE (THERASENSE, INC., Alameda CA), and collect and correspond This mathematical relationship is obtained from the far-infrared measurement of the blood glucose concentration measured above. For the spectrophotometer, several single-variable and multi-variable methods can be used to construct the mathematical relationship between the measured signal and the actual blood glucose value. However, the known basic formula is the Beer-Lambert rule. The rule states that the intensity of the measured absorbance / reflectance is positively proportional to the concentration of the analyte being measured. 200538083, as shown in Formula One, A = sbC (1) #where / is the absorbance / reflectance under a specific wavelength of light The measured value is the molar absorption of the target molecule at the same specific wavelength, 6 is the distance the light moves, and C is the concentration of the target molecule (glucose). The calibration technique uses different signal processing and calibration methods, including one or more mathematical modes, to obtain glucose-related signals from the measured spectrum. The model is developed through a set of standard spectrometric measurements (or calibration sets) and a set of calibration procedures that refer to blood glucose-value-related data obtained from fingertip microvascular fluid, venous blood, or intercellular fluid analysis. A common multivariate method requires a standard glucose concentration reference vector as a reference standard when calibrating each sample and the spectrum. These multivariate methods include partial least squares (PLS) and principal component regression (P0! R). Many other methods of calibration or optimization are known to the skilled artisan. A device capable of measuring infrared output generally includes an energy source capable of emitting infrared energy of multiple wavelengths, an input element, an output element, and a spectrum analyzer. After the tissue is irradiated with multiple wavelengths from the output element, at least part of the photons are scattered and absorbed by the tissue at the irradiated area. Part of the above photons leaving the tissue sample will be collected by the output element and guided to a detection device for detection. Subsequently, a pattern is used to determine the concentration of the analyte. Calibration maintenance 11 200538083. In the material. In general, examples of the instrument such as the interference species and the state of the test are important. However, the target is in the state and group. The optical element includes degree, and the multi-variable mode is used to reduce the number of variables to reduce the complex measurement. Under a series of conditions, the data collected to construct the original mode is processed. When the series of conditions are changed, the original mode cannot be operated properly. For example, the ambient temperature will affect the light collection performance of the spectrophotometer. In addition to environmental and environmental impacts, changes in the sample can also affect the model. • When changes in the concentration of interferences fall outside the measured range, or when new ones are introduced. Another important and relevant factor when using non-invasive methods to determine glucose concentration in the body is that the sample itself is living and mobile and requires a new calibration of the analyzer. H Combining a subject's non-invasive spectrum to perform the calibration can determine the analyte concentration. F-state features The dynamic feature of skin tissue is a feature often overlooked in non-invasive glucose measurements. In a specific measurement site, except for changes in the precipitate and other interfering substances, the skin group is usually assumed to be stationary. However, in a relatively short period of time, differences in the distribution of physiological fluids in the body's tissues can greatly affect the quality of the tissue layers and body cavity. There are many factors that can affect Pi Guang's physiological and chemical state. These are due to environmental and physiological factors. Such factors include, at a minimum, body temperature, ambient temperature of food, medicines or medications, and the application of pressure to the sampling site. 12 200538083 When one part of the body is affected, it affects many other parts of the body. For example, ingestion of food into the digestive tract causes water to move between chambers in the body. Or ingestion of caffeine or stimulants can change blood pressure or cause microvascular dilatation. Non-invasive glucose measurement

There are many reports on non-invasive glucose measurement techniques. Some of these are related to the general instrument configuration required for non-invasive glucose measurement. The other is about sampling techniques. Here is only a brief review of the documents most relevant to the present invention: General Instruments British Patent Application No. 2,033,575 entitled Pj Rolfe titled α j (8

(January 24, 1979) describes a device that can be used to introduce light into the body and measure weak backscattered light, and use the collected signals to measure in or near the blood flow Glucose concentration. C. Dahne and D. Gross titled Specimphotometric method and apparatus for the United States Patent No. 4,655,225 (April 7, 1 987) describes a method and device that can direct light into a The patient collects reflected or backscattered light, and measures glucose from a selected near-infrared wavelength band. Wavelengths include 1560 to 1590, 1750 to 1780, 2285 13 200538083 to 2115, and 2255 to 2285 nm. At least one additional reference signal between 1000 and 2700 nm is required.

In M. Robinson, K. Ward, R. Eaton, and U. Haaland, the title is Method and apparatus determining the similarity of a biological analog from a model constructed from known biological fluids ^% 4,975,581 ^ U.S. Pat. 1990) describes a method and a device that use an infrared spectrum analyzer and a multi-variable mode to measure a biological analyte, such as glucose concentration. This multivariate model is constructed from a variety of known biological fluid samples. In J. Hall and T. Cadell Method and device for measuring concentration levels of blood constituents «ow-Mvahve / y US Patent No. 5,361,758 (November 8,

1 994) describes a non-invasive device and method for determining the concentration of an analyte in a living body by using a polychromatic light beam, a wavelength separation device, and an array detector. The device uses a receiver with a shape that accepts fingertips to block external light. Also in R. Barnes, J. Brasch, D. Purdy and W. Lougheed entitled Non-invasive determination of analytical concentration ζ · «乂 y 〇 / wczwma / s US Patent No. 5,379,764 (January 10) No. 1 995) describes a non-invasive glucose analyzer that uses a data pre-processing process in conjunction with a multivariate analysis to determine blood glucose concentration. And U.S. Patent No. 6,040,5 7 8 (March 21) in S. Malin and G. Khalil entitled awe / 14 200538083 apparatus for multi-spectral analysis of organic blood analyses in noninv as iv e infrared spectoscopy (2000, 2000) describes a method and apparatus for measuring a beta-organic blood analyte using near-infrared multispectral analysis. Inject a variety of light with unique and non-overlapping wavelengths on the same surface, collect the diffusely reflected light, and use chemometric techniques to determine the concentration of the analyte. Temperature · It is well known that many physiological components have near-infrared absorption spectra, and these spectra are very sensitive to the intensity and location of the regionalization temperature. Some literature points out that the above factors will affect non-invasive glucose measurement (see Ke Hazen,

Glucose determinations in biological matrices using ne ar -infr ar e d spectroscopy, Doctoral Dissertation,

University of Iwa (August, 1995)) Coupling fluid

In T. Blank, G. Acosta, M. Mattu, and S. Monfre, US Pat. No. 6,415,167 (June 2, 2002), a fiber optic probe guide placement guide, describes the use of one or more processes. One of the coupling fluids is composed of decayed compounds. The coupling liquid of sound volume is placed on the interface between the optical probe and the measurement site. Perfluoride has no toxicity associated with fluorocarbons. 15 200538083 Calibration adjustment Various methods are known to compensate the dynamic variation of some tissue samples. A method for non-invasive glucose concentration measurement and used to generate a calibration mode 'is to model a body's glucose concentration over a short period of time. (See also H.zen, Glucose determinations in biological matrices using near-infrared spectroscopy, Doctoral Dissertation, University of Iowa (August > 1995), and J. Burmeister, / zwmaw

noninvasive blood glucose measurements, Doctoral Dissertation, University of Iowa (December 1997)). This approach ruled out the possibility of incorporating patient-to-patient differences into the model, but it was not possible to generalize it to more patients. This method can not solve the problem of physiological variation in short-term measurement, and cannot compensate for the variation caused by the dynamic movement of water molecules in the body fluid cavity in the literature.

Another method to overcome the impact of modules on organizational changes is cross-validation. In one study, food tolerance tests were used to affect the glucose concentration of three subjects, and multiple calibration modes were established for each subject during the day, and these were tested through cross-validation Modules (see Robinson MR; Eaton RP ·; Haaland DM; Keep GW; Thomas EV; Stalled BR; and Robinson PL Non-invasive glucose monitoring in diabetic patients: a preliminany evaluation, Clin Chem 1 9 9 2 i 3 8: 1 6 1 8-2 2). This method models the differences between some patients. The intention is to model the variation so that the original model can be used to predict it. 16) 200538083 The method also fails to solve the problem of short-term measurement.

His patient. Similarly, the problem of physiological variation, the variation caused by dynamic movement in the fluid cavity

situation. When incorporating many different situations, it usually takes weeks to build the model. To date, such a large number of calibration and inspection procedures have been difficult to succeed. Another method to overcome the impact on the model due to organizational changes is to use an intelligent identification system to compensate for changes in organizational structure and state. 'The system can determine the most appropriate schooling model for the subject at the time of the measurement (see Malin, S. F ·; et · al · / or

Noninvasive 5 / 〇 with J «α / less 0w, U.S. Patent No. 6,280,3 81). A representative patient population is divided into several groups, and the calibration pattern is established from its absorption spectrum. These groups or categories are classified by the similarity of structure and state, so that the variation within a group is smaller than the variation between groups. Classification is based on characteristics obtained from tissue absorption spectra that are related to the patient's current state and structure. _ There is another type of method to overcome the impact on the mode due to the change of the level is the calibration transfer. This method uses a variety of spectral data techniques before processing. Appropriate pre-processing has the following common but incomplete steps including: sorting, wavelength 17 200538083 selection, centering, scaling, normalization, selection of the n-th derivative function (n $ i), smoothing, Fourier transform, principal component selection, Finite Pulse Response Filtering, Linearization and Transformation. It has been found that the application of such general techniques to non-invasive glucose concentration analyzers has been limited. There is another additional way to overcome the impact of organizational changes on the model. This type of technology is based on the principle of using a single spectrum to perform regional centering (see Lorbei: et. Al ·, h

Multivariate Calibration, Journal of Chemometrics, 1 996, 10, 215-220). In this method, a spectrum is selected to average the centralization of the calibration data set that is closest to the spectrum of the unknown sample (on Mahalanobis distance). Then construct a part of the least squares model for each unknown sample. However, this technique cannot reduce the variability of the calibration group in the spectral analysis. Another method that can overcome the impact of organizational changes on models is the technique of average centralization (see E. Thomas, R. Rowe, Methods and ^ PP ^ ratus f〇r tailoring spectroscopic, US Patent No. 6 528 No. 809 (March 4, 2003), and E. Thomas, R. Rowe, apparatus for tailoring Spectroscopic calibration)). This method uses spectral camera technology with an improved calibration mode tailored to the subject. In the calibration data, the model data can be changed to reduce or exclude the characteristics related to individual test subjects, so a calibration data set model constructed under the test subject's physiological variation and instrument variation is generated. In the prediction phase, the minimum light 18 200538083 spectral measurement is performed on each subject to modify the prediction process for each target subject. However, this method cannot solve the main problems of short-term physiological and chemical changes related to the dynamic nature of tissues and the variation in a single patient due to the heterogeneity of tissue samples. References to U.S. Patent No. 6,528,809 to E. Thomas and U.S. Patent No. 6,1,57,04 to F. Thomas use / infrared-based non-invasive grape bran concentration analyzers and A mode is used in combination to obtain the absorption spectrum of human tissues. This mode can periodically modify a spectrum collected from the subject and the invasive reference glucose concentration measured from the subject. Although the subtracted spectrum is not an average spectrum and the glucose value is used to correct the compensation part of the predicted value, this technique can still be roughly classified as the average centralization method. Because collecting the reference spectrum is very time-consuming, the user must also have special expertise, and also need data collection software and a microprocessor or other operator to perform the calibration. In addition, the sample spectrum used as the reference spectrum contains data collection errors and spectral characteristics that cannot be explained by the original mode. Therefore, including the reference spectrum of individual individuals in the model will become a significant source of energy error during calibration, which will also be directly converted into subsequent errors in glucose concentration prediction. In a non-invasive glucose measurement, this can lead to incorrect glucose concentrations reported to aid insulin therapy. For these reasons, it is very beneficial to reduce the steps of collection and utilization-reference sample light dancing. 19 200538083 Here, T. Blank, G. Acosta, M. Mattu, and S. Monfre are titled US Patent No. 6,415,167 (July 2, 2002) and T. Blank by Fiber optic probe guide placement guide , G. Acosta, M. Mattu, M. Makarewicz, S. Monfre, A. Lorenz and T. Ruchti are titled 5 Less * yk / w

For / «Fivo A / ewwremeW 〇 / Hut / e US Patent Application No. 10 / 70,921 (application dated June 12, 2002) is incorporated by reference in its entirety. A non-invasive glucose analyzer to increase the accuracy of the location of the sampling site. Therefore, the accuracy and precision of non-invasive glucose concentration measurement will be improved. Over a period of time, such as a part of the day, a day, or a few days, the guide is used to increase the accuracy of sampling during the sampling period. balance

Many literatures have pointed out the differences (or lack of differences) between traditional glucose determinations and alternative site glucose determinations. Some believe that this possible difference will have an impact on non-invasive glucose calibration and maintenance (see U.S. Patent No. 10/377, 9 16 · Application). Some literatures indicate that heat, lubrifractants, or other agents or vasodilators can be used, such as acid test, methyl end test, minoxidil, nitroglycerin, stigmine, menthol, Capsaicin and its mixtures to accelerate the equilibrium of glucose concentrations in blood vessels and intercellular fluids. (See Rohrscheib, Mark; Gardner, Craig; Robinson, Mark R. Method and Apparatus for 20 200538083

Non-invasive blood analysis measurement with Fluid Compartment Equilibration, US Patent No. 6,240,306 (March 29, 2001) and Robinson, Mark Ries; Messerschmidt, Robert G. Method for Non-invasive Blood Analyte Measurement with Improved Optical Interface , US Patent No. 8,152,876 (November 28, 2000)). There are also references in the literature that the release of nitrogen oxides through light stimulation combined with non-invasive glucose measurement can be used as a method to balance the glucose concentration between areas with poor perfusion in the body and areas with better perfusion. (T. Blank, S.

Monfre, M. Makarewicz, M. Mattu, K. Hazen, and R.

Henderson, P hotostimulation method and apparatus in combination with ghicose determination, 2QQ4 May 6 3 application (agent file number SENS0034). All related technologies mentioned in this section do not point out that it is only necessary to use a reference glucose average and use a guide to eliminate the necessary average centralization method using a spectral reference. In addition, it does not point out that by using a guide to reduce the need for the technique of averaging centralization by using spectral reference values, the use of biological analyzers such as near-infrared-based non-invasive glucose analyzers is simplified. Furthermore, no literature indicates that light stimulation can be used to minimize differences in reference glucose concentrations. Finally, to date, 'there are no FDA-approved devices for non-invasive glucose determinations by people or medical professionals. Question 21 200538083 Physiological parameters that can change the skin state include: tissue hydration, skin temperature, tissue blood volume ratio, skin thickness, fat-related absorption characteristic strength, blood cell volume concentration, and surface reflectance. Several of these parameters change over a period of one day, several days, or longer, such as weeks. Changes in the skin state can cause changes in various properties, such as: water concentration, concentrations of other analytes such as protein f, fat, keratinocytes, and glucose, skin wide scattering, skin absorption, reflection coefficients of skin layers, and tissue layers Thickness, light emission energy from the body, mechanical properties of the tissue, moisture-related absorption characteristic intensity, * white matter-related absorption characteristic intensity, and the size and distribution of scattering centers. Non-invasive spectra (such as near-infrared diffuse reflectance spectra) are representative of skin tissue characteristics. Because there are many states that change, and each state change will affect a variety of skin tissue characteristics, the _ specific skin sample sampling section: the non-invasive spectrum variation will often show _ kinds of highly non-linear and Complex form. In addition, multivariate patterns based on factor analysis can lead to abstract features. The change of the state will greatly affect the prediction result of the multivariate model. Since near-infrared-based non-invasive glucose analyzers often use multivariate analysis, which is susceptible to changes in state, they need to be modeled or updated at any time. The focus of this issue is on the maintenance of one or more standards. Calibration is a costly and time consuming process that requires some expertise. > The steps required to reduce or adjust the maintenance of a non-invasive glucose analyzer have at least one of the following advantages:-increase the marketability of the analyzer, 22 200538083 increase the number of people who can operate the analyzer, reduce the time required to enter a # And increase the glucose concentration determination ^ grape bran concentration determination. In particular, it would be helpful to reduce any data collection steps, such as precision and / or the advantages described correctly. The collection of illuminating light 4 in the present invention provides simple calibration for the use of the upper glucose concentration analyzer to j kinds of non-invasive early maintenance methods. [Summary of the Invention] The present invention proposes a method and device, which can simplify the use of an optical-based non-knowledge: intrusive analyzer. More specifically, using a simplified algorithm to eliminate the need to collect and use non-intrusive, and & nine spectrum to update the calibration mode. In another specific embodiment, a guide is used to greatly reduce the variation between the sample detection site and a skin tissue sampling site, and a reference analyte concentration is used to collect a reference obtained in a timely manner regardless of whether it is used in conjunction with or not Under the condition of the spectrum, the analyzer has the ability to maintain the calibration performance. [Embodiment]

The preferred embodiment of the present invention uses a near-infrared-based glucose concentration analyzer to obtain a spectrum of a sampling site of a body, and cooperates with a model and a reference analyte concentration correction to predict / measure the subject Glucose concentration. y For example, U.S. Patent No. 6,528,809 to E. Thomas and U.S. Patent No. 6,157,041 to E. Thomas, which describe a calibration method using a reference spectrum from one of the test objects and a direct reading of glucose from the individual 23 200538083. Traditional compensation correction method.

The term "reference spectrum" in Lt 4ά refers to a tissue spectrum, such as a day's initial spectrum, a spectrum from a data temple, or a "quasi-valued spectrum" relative to a reference library. It is said that the collection of the reference spectrum must pay for the instrument and δ. The user's operating technology and other costs, such as the intensity of the day, between, and the collection of the reference spectrum may become the second standard value. The pre-rigid value of grape bran is good. It is better than the above reasons, and the reference of collection and use is eliminated. It is very beneficial. The sample spectrum step is very good. An ancient method and device are proposed in the preferred embodiment. It can remove the step of periodically updating the calibration mode to collect M β ^ non-bismuth introspective reference spectra. In other words, it is not necessary to use the reference spectrum of the user, the reference spectrum of the user's spectrum, or a reference spectrum of the individual. 禋 Pairing the first spectrum, you can use it every day, week, or month. Periodically to update the pattern. The need for the reference spectrum to be successfully eliminated is based on at least one of the following reasons: First, gathering the reference spectrum is not beneficial for the net prediction of glucose concentration. -In general, the purpose of using a reference spectrum is to provide a compensation constant. The compensation that can be provided by the direct reference glucose measurement value is also a constant. … These debts are all constants, so a single correction using direct glucose indifference measurements shows that the total error also includes the above two constants. The detailed description is as follows. In particular, in Equation 1, Xmeasured is the vector of the predicted spectrum, WT is the vector of the correlation coefficient of the regression model, and eight is the predicted value of the analyte (glucose) concentration obtained before any calibration. It should be noted that 24 200538083 y hat has an error associated with it. The yhat prediction can be regarded as an uncorrected estimate of the glucose.

Yhat = (Xmeasured) The estimated yhat glucose concentration of the WT has an error-based reason, including the instrument, environment, and sampling. The reason that is particularly important in the measured spectrum is that when the tissue volume changes, parameters such as optical path length, absorption, etc. Coefficient and scattering coefficient. The method of correcting the error of the predicted glucose concentration yhat was first proposed in E. Thomas US Patent No. 6,528,809 by E. Thomas US Patent No. 6,1 5, 7,041. This method uses non-invasive reference spectrum Xref to regionalize the measured sample spectrum. Consider Equation 2, using the reference spectrum Xref will make a compensation for the predicted glucose concentration, and its error compensation value is Y b i a s 1. The new predicted value of the bran concentration after partial correction is marked here as y 'hat. It is generally an average spectrum, but as described below, there are many other sources of this Xref spectrum.

yh at = yh at + y bi as 1 — (X me as u re d — Xr ef) WT Basically, XrefWT is used as a compensation for yhat, which is equivalent to the above-mentioned correction method under 'reasonable operating conditions' Make ybiasl equal to the actual grape bran concentration. For example, the analyzer may have less stringent requirements, such as determining only low analyte concentrations. This makes it possible to limit the correction of errors by using only a reference spectrum. However, usually the reference spectrum obtained after applying this mode will produce a relative analyte that still needs to be compensated for correction (culture sugar concentration (1) in the following response. Its deviation, number, and one use one. Please refer to degree y Hat Inter Grape ^ Xref can be used as (2) Υ bias 1 〇yhat + for accurate high or to adjustable light $ grape bran) 25 200538083 concentration prediction. This compensation correction is performed using a direct reference glucose concentration. Traditionally, the direct reference glucose concentration is determined by traditional fingertip grape bran, but additional options will be discussed below. This direct glucose concentration reading or reference glucose measurement is referred to herein as ybias2. Combining the direct glucose concentration determination step with the deviation correction step in Equation 2 will produce a measured analyte (glucose) concentration, ymeas, as shown in Equation 3. ymeas = yhat + ybiasl + ybias2 = (Xmeasured —Xref) WT + yb 丨 "2 (3) As mentioned above, the first method requires the collection and use of a non-invasive reference spectrum and a reference glucose concentration. Back to the formula 1.It must be noted once again that the measured spectrum should be thrown to the calibration mode to produce a glucose prediction value yhat, and this prediction value has an error associated with it. In the second method, only the direct bran reference measurement is used but The step of pre-subtracting the average spectrum was not performed to correct this error. In this case, the direct reference glucose measurement labeled 'yerr' can provide the predicted glucose concentration yhat —the estimation error, using formula 4 to estimate a glucose concentration. Ymeas = yhat + yerr = XmeasuredWT + yerr (4) After comparing formulas 3 and 4, it can be found that the estimated error yerr obtained by using this second method can represent both ybiasl and ybias2 in the more complicated two-step method described above. Proposed here A different embodiment of the present invention that utilizes one or more stored reference spectra. In the method described above, because the general technique A spectrum is subtracted from the spectrum, and an average spectrum is usually subtracted, so the reference light 26 200538083 spectrum is often used as an average spectrum. However, the reference spectrum is only one of many spectrums, which include a period of time including a day , The first spectrum of a week, a month, or even a spectrum collected when the instrument is transmitted to a subject. Traditionally, this is done once a day, during the awake period of the day, or during the main part of the day Correction. In addition to using the first spectrum over a period of time, you can also use the average value of η spectrums over a period of time, or a spectrum obtained by linearly combining the spectrum of any number of adjacent times or individual times in a period. The reference spectrum can also be selectively generated from a database, such as a spectral database, and selected based on spectral characteristics or a calculation method such as the Mahalanobis distance. In other specific embodiments, a basic gold standard is used Spectra, such as water, are used as reference spectra. The advantage of fixed gold standard spectra is that they are controlled Collected under the environment, and stored in the analyzer in several ways for subsequent mode update. In another specific embodiment of the present invention, a glucose concentration analyzer based on near-infrared is used to obtain a body sampling site A spectrum, combined with a guide, a model, and a reference analyte concentration to predict or measure a glucose concentration of the individual. The light output of the spectrophotometer is affected by a variety of factors, including the state of the environment and the state of the sample In addition, changes in the state of the environment and changes in the state of the body cause dynamic changes in the sampling site of a living tissue. Examples of changes that affect the state of the tissue include changes in temperature and / or distribution, changes in local pressure or distribution, movement of water molecules, movement of glucose, and Changes in body composition such as protein, fat and blood cell volume. These changes affect the chemical and physical characteristics of the body. Physical changes include local and / or regional changes in refractive index 27 200538083, changes in & script coefficients, anisotropy, and effects on detecting photon penetration, radiative transfer, and #woven # ^ & path length The scattering coefficient. The combination of the above state changes will make the non-invasive 4 + b a _ — two-degree nonlinear relationship in the nine spectrum. Many of the changes described above with respect to the sampling product are highly dependent on the optical sampling probe. '' Dry sampling probes include spectrophotometer stimulation and / or collection elements. The physiological condition of the skin is not uniform. The chemical and physical composition of the skin varies with the position of the skin & t π ^. / L ^, > ^. The structure of the skin will vary depending on the location W, α, ′, including the thickness of the skin layer and the skin composition. The non-aggression of the body, [Li Gumei said, it varies with the sampling site. For example, the light 4 of the forearm skin tissue sample will vary with the position of the skin sampled on the forearm. For example, a skin sample will vary with components such as water, fat, protein, and properties such as scattering coefficient at different locations. Guide In the broadest sense, a guide is an element that limits the positioning of the sampling probe relative to a sampling site. The guide can transfer 70 input and output to ~ target tissue volume to control and reduce spectral variation, and can reduce the error in the measured analyte concentration. The guide is used as half of a lock and key mechanism to limit the positioning of the sampling probe (key) relative to the sampling site. Guides are helpful for positioning the reproducibility between sampling and collection elements, such as the relationship between the sampling probe of an analyzer based on a non-invasive spectrophotometer and the fiber relative to a sampling site; and because it can be associated Ground to reduce non-linear variability, so it is also helpful for the determination of a non-invasive analyte 28 200538083. A guide was taught in U.S. Patent No. 6,415,167 (July 2, 2002) and U.S. Patent Application No. 10 / 170,921 (filed on June 12, 2002), which incorporates the entirety of the two documents. Reference.

Various guides and attachments are described here. In the preferred case, the add-ons have the same interface so that a single guide element can be used with each add-on. Similarly, preferably, each part of the device that can be connected to an attachment should have the same interface. For example, in the preferred case, the guide has the same interface as a reference sample, so both can be coupled to a sampling probe. The commonly used interface can be used as a bonding interface for any guide or a reference combined with any additional objects, such as connectors, photon stimulators, sampling modules, or miniaturized light sources. Lock (Guide)

Sampling locations vary from individual to individual and can be expressed in circles or radius of curvature. As far as the body part is concerned, the radius of curvature in the abdominal area and the finger can range from flat to 0.375 inches. Even in a body part, the radius of curvature varies from individual to individual. For example, some people have smaller arms and others have larger arms. The shape of the guide is determined according to the structure of the sampling site, which will increase the accuracy of subsequent / continuous optical sampling. Examples of guides are a flat sampling interface and a radius of curvature of 6.0 inches, 4.5 inches, and 3.0 inches. Figure 1 presents a guide element 29 200538083 with a flat radius of curvature. When using one arm as the sampling site, the thinner the arm, the smaller the radius of curvature of the best guide. The guide primer presented in Figure 1 is about a linker, which is one of the guide additions to be discussed below. The main feature of the guide element is that it forms a combination of a lock and a key. That is, there is a surface that can repeatedly guide the other half of a lock and key element to a repeatable installation position. In this case, the lock element is located in the guide, but it may alternatively be located in the add-on. Here, the lock element is a hole in the guide. The guide is substantially rectangular and has two circular opposite sides. This rectangular shape limits its rotation arrangement. In the preferred case, the guide cannot rotate freely. For example, the guide shown in the figure has a C2 symmetry 'so that it still has the same shape after being rotated 180 degrees. This freedom of rotation can be achieved in a number of ways, such as flattening one of the rounded ends. In the second case, the locking element protrudes along the top of the lock element into the relevant hole in the 孔 4 (attachment) element, so that the key can be positioned to the guide in a repeatable manner. A variety of lock element shapes can be used. Embodiments can include holes of almost any geometry or any shape (not necessarily a hole) that provides reproducible positioning capabilities, and in the best case, these shapes prevent free rotation. In the proposed specific guide element, holes or grooves which can be optionally added are traced. Its functions are mainly to reduce weight, minimize surface anomalies (such as dent marks on the sampling site), and maintain strength while limiting the freedom of distortion of the guide. An additional non-essential element is depicted on these guides, the magnetic weave. These magnets are used to control contact forces and / or assist lock 30 200538083 head and key mechanism alignment. In the guide, the non-essential box ^ ΐ •, a heteropolar magnet is introduced. In this pair of magnets, the pairing can be a metal substance, such as a metal plate or stainless steel, which can be used at low cost and / or to reduce weight. Those skilled in the art can understand that there are various 4 # φ ^ blade edge mechanical structures to connect the lock element to the guide element. A guide is attached to a sampling site with a kind of instrument. ΛΟΟ jMr ”If using rope, strap and hook and loop technology, a double-sided adhesive can be used preferentially. Generally, S ′ puts the adhesive firmly on the sample ^ ^ ^ m and then places the guide on the adhesive. This sequence can reduce separation of the adherend from the sampling site. Alternatively, an adhesive can be attached to the guide, and the integrated adhesive and the guide can be brought into contact with the sampling site. This method can eliminate the step of arranging the guide to the adhesive. Or the adhesive obtained by the user has been attached to the guide element. The guide and the adhesive are semi-permanently and removably attached to the sampling site. The guide will usually be left in place during periods other than the sampling period, such as during a sober day or during a data collection period such as 2, 4, or 8 hours. A selective interposer or guide extension can be used between the guide and the double-sided adhesive attached to the sampling site. In essence, this is a semi-elastic material, such as acetate. The material provides some elasticity to allow the skin at the sampling site to stretch. This reduces transient changes in sampling caused by the subject's movement. Conversely, for subjects with less severe swelling and edema, because the skin is too elastic, a more rigid interposer, such as a plastic sheet, is required. Priority should be given to the use of a thermoplastic material to make the guide, for example 31 200538083 polycarbonate is now more transparent. The parts are good, and the material is used to make the head of the key, and the sample is taken. The joint sample is an interactive linker or polyurethane. However, Baizhi artisans can easily develop available materials. Because the guide will come into contact with the sampling site (sometimes an intermediary adhesive), the thermal properties of the guide are not critical. ; The guide is not thermally conductive to reduce the temperature rise of the sampling site. However, in some cases ... For example, when it is necessary to transfer thermal energy :: sample or from the sampling site, use a thermally conductive guide. Compare. The guide material is preferably biocompatible. In the preferred case, a compatible refractive index medium such as a fluorine, Fluorinert, FC-40, fc-70, or 1 & ^ #Specific Object is optically coupled to the guide Sampling site. (Attachment) Here, the other half of the guide lock and key mechanism is called a guide element attachment. Examples of attachments attached to the guide include a top of a photon stimulator, a miniaturized light source, and a spectrophotometer. In view of these attachments, a joint attachment is described in the following. Fig. 1 proposes a joint attachment coupled to a guide. The head can achieve at least one of the following functions: hydration of sampling by latching, protection of the sampling site from physical 丨 1 protection of the position from contamination, the arrangement of the guide α, and maintaining a good appearance, for example Shape, ring, or graphic symbol. The joint is depicted with two protruding elements with supporting structures to assist in joining. This watch can optionally be made into a watch shape or be equipped with the function of a watch, and it may or may not be provided. 32 200538083 It shall be provided with a 0 9 arm, a viscous ° hole. Sampling sample is therefore a concentration of bran matter in the target portion. In the ‘Interface, a strap similar to a strap. In a variant embodiment, the connector can be miniaturized to make it look like a decoration, such as a ring. The connection shown in Figure 1 has an optional central channel. The lane is used to place the guide in its original position. In this method, a double-sided adhesive is connected to the sampling site, the adhesive has an opening and the opening is slightly larger than the optical detection element. After placing an adhesive on it, 'connect the guide to the connector and lower the guide rod to the object' so that its optical path is located in the wide vent of the adhesive. The guide rod of the penetrating joint and the guide is used as an inspection. Fig. 2 proposes a photon stimulator attachment, which is coupled to a guide having a curvature radius of 4.5 inches on an interface. Take at least one

Light stimulation is performed at or near the position to enhance perfusion of the sampling site and reduce sampling-related errors. Increased perfusion at the sampling site can lead to an increase in the volume percentage of the analyte 'and / or make the blood or tissue composition concentration more correct and / or accurate to track the sample components corresponding to the better perfused body parts, such as arteries, blood vessels, Or fingertips. In a specific example, by analyzing the light-stimulated area and cooperating with a dialysis analyzer, it is possible to more easily, accurately, or accurately measure the glucose index 'and to measure another non-exposed material | The analyte of the part is described in the U.S. Patent (Attorney's Slot Number SEN describes this technique, and it is hereby incorporated by reference as a whole. A sampling module is proposed, which is coupled to a curvature taken. On a guide with a radius of 6.0 inches. The sampling module is described in US Patent Application No. 10 / 472,8 33 200538083 (application dated September 18, 2003, agent file number SENS0011). In a preferred case, the sampling module may be part of an analyzer, which may additionally include a base module and a communication beam. The sampling module is continuously or semi-continuously connected to a human receiver The tester collects spectral measurement values used to determine a biological parameter in the sampled tissue. The preferred target analyte is glucose. The preferred analyzer is a near-infrared-based glucose analyzer that can be used for Given the concentration of glucose in vivo.

Figure 4 proposes a miniaturized light source attachment and incorporates a reference guide element. A reference element is used to provide a signal that is representative of the current state of the analyzer. In a preferred case, the reference component is included in a component having an interface. The function of the interface is the same as that of the guide. The reference component can be coupled to a sampling mode in the same manner as the guide. Group or a miniaturized light source. Guide information

Figure 5 shows the standard deviation of a series of far-infrared non-invasive spectra obtained from the spectrum collection of an arm of a body, where 501 is a situation where a guide is thrown away, and 502 is a case where a guide is not used Situation. It can be found that when a guide is used, the standard deviation of the absorption unit is significantly smaller. Those skilled in the art can understand that the main cause of this deviation $ is the path length and the detected water concentration, and the lesser variables are related to protein, fat, and temperature. Therefore, this guide increases the accuracy of the position of the optical probe relative to the sampling site, thereby reducing the sampling variation caused by the position, which makes the 34 200538083 spectrum measurement source reference source and the school equipment to predict the subject's acceptance. The variation of the completed reading of a watch with a poor wavelength detection is small. 1 Figure 6 presents a set of exemplary non-invasive glucose concentration prediction sets. The above-mentioned preferred mode correction technique is used to process these representative materials. “Stomach better mode correction technique refers to a technique that uses only a direct glucose concentration and does not utilize a non-invasive (or related) reference spectrum. The instrument for collecting authorization data includes three types of optical devices equipped with a toothed tungsten filament light stimulation guide, a wavelength filter, a collection guide, a slit, a grating, an array detector, and related electronics. Near-infrared-based analyzer. During calibration and subsequent measurements, a guide is placed on each recipient on a specific day when all measurements are to begin. During a period of at least one day, all will undergo one or more glucose determinations. The calibration was performed with eight subjects. Collected 24 i copies using a calibrated instrument over a two-month period. A 27-point Savitsky-Golay-order derivative was used to preprocess the data of the spectrum from 12 00 to 1 800 nm to construct a model. In vitro procedures include modules to detect and exclude samples with surface contact between the optical device and the sampling site, modules to detect and exclude samples with larger tissue transients, and to detect and eliminate excessive tissue distortion This module. Use a multivariable partial Fen least squares model. Predicted spectra were collected with an analyzer similar to a glucose analyzer, but only with a different set of stimulus and collection optics. Seven months after the collection of the school's material # 仪 + 贝 料, the forecasted light level is collected over a period of five weeks. "The forecast resource 35 200538083 The type and testability of the material is not a value A This school Pre-exam scores included

The group consisted of 3 6 8 predicted spectra obtained from 2 1 a μ f ^ A 1 test subjects with three instruments. Using self-calibration with caution2, the pattern generated by the Pinggong group is used in a single-blind form without any modification. Li and Yu—x 4, W use a compensation switch, which uses the glucose reference concentration of the first sample collected on a specific day. Prediction 'does not use any reference spectrum from any of the predicted spectra, nor does it use any non-intrusive spectrum (into a non-intrusive spectrum) in any it segment of the prediction asset. The model is updated again with a reference glucose concentration so that a non-invasive analyzer makes a subsequent prediction of the glucose concentration and

Use a corresponding non-invasive reference spectrum. In Figure 6, the resulting prediction is plotted on a Clarke error grid. A total of 98.7% of the obtained glucose concentration predictions fell into the clinically acceptable or region B of the Clarke error box, and the standard deviation generated by the predicted value was 26 3 mg / dL. A result clearly shows the performance of the analyzer. It is worth noting that 100% of the measured glucose concentration falls in the B area of the Clarke error grid, and the standard deviation of the calibration is 270 mg / dL. It can be found that the measured performance is almost the same as the calibrated performance. In the method described here, one parameter is subtracted because

Spectrum, and is usually subtracted from an average spectrum, so a reference light is usually called an average spectrum. However, alternative reference spectra include one of the following, including the first spectrum in a period, which includes a day, week, month, or even when the instrument is delivered to a body A kind of spectrum. Traditionally, corrections can be made once a day, at awake time of the day, and in the main part of the day. Alternatively, the removed spectrum may be the first spectrum in a period, the average value of the first η 36 200538083 spectra in the period, or any number of adjacent 1 or individual times in a period The spectra in the spectrum are obtained in a linear combination. In a specific embodiment, the reference spectrum is from a database, such as a "spectral database" and is selected based on spectral characteristics or a calculation method such as a Mahalanobis distance meter, for example, the gold spectrum described above can be used. In the above methods, the direct reference analyte concentration is usually based on the traditional enzyme method, electroenzyme method, or colorimetric method for blood glucose measurement in microvessels. However, a minimally invasive glucose concentration meter can also be used Or a non-invasive concentration meter to generate the reference glucose concentration. Usually a reference method approved by the FDA is used. Although the invention is described herein with reference to a linear model, a non-linear model such as a nerve Analyze the concentration of analytes using a network. Although the present invention has been described in terms of several preferred embodiments, those skilled in the art can easily find that other applications can replace the ones listed here without departing from the present invention. The spirit and scope of the invention. Therefore, the present invention should be limited only by the scope of the following patent applications. FIG. 2 shows a flat guide coupled to a connector according to one of the inventions; FIG. 2 shows an LED accessory connected to a connector having a guide with a curvature radius of 4.5 inches according to the invention; The figure shows a miniature light source coupled to a guide with a radius of curvature of 60 inches 37 200538083 inches according to the present invention; Figure 4 shows a miniature light source coupled to a flat guide according to the present invention; Figure 5 shows the spectral variation with and without the use of a primer and primer according to the present invention; and Figure 6 shows the prediction result plotted in the Clarke error table according to the present invention. 】 ♦ None

38

Claims (1)

200538083 The scope of patent application: 1. A non-invasive method for estimating a biological property of the human tissue of a subject, which includes at least the following steps: Provide a device for measuring light output, the device includes at least one An energy source of light, an input element, an output element, and a spectrum analyzer; coupling the input and output element to a target of the human tissue that controls a sample volume; using the input element to make the light energy Multiple wavelengths are irradiated to the tissue, wherein the sample volume will cause partial absorption of the light; use the output element to collect at least a portion of the unabsorbed light energy; determine the intensity of the collected portion; generate a pattern, wherein the pattern Using a plurality of calibrated non-invasive spectra and a plurality of calibrated reference concentrations; adjusting the mode with an estimation error, which uses the spectrum of the target with a control sample volume and at least one direct reference measurement value to generate the estimation error; collecting At least one non-invasive prediction of the subject's target of controlling the sample volume Spectrum; and using the model to estimate the adjustment of the concentration of the biological properties of the subject test subject. < 2. The method according to item 1 of the scope of patent application, wherein the biological property includes at least a glucose concentration. 39 200538083 3 · The method according to item 1 of the scope of patent application, wherein the light contains at least near-infrared energy. 4. The method according to item 3 of the scope of patent application, wherein the interval of near infrared energy includes at least an interval from 11000 to 1 900 nm or at least one interval therebetween. 5. The method according to item 1 of the scope of patent application, wherein the calibrated non-invasive spectra include at least any one of the following spectra: a single individual; a plurality of individuals; a single analyzer; and a plurality of analyzers. 6. The method according to item 1 of the scope of patent application, wherein the above mode includes at least one multivariate analysis. 7. The method according to item 6 of the patent application scope, wherein the multivariate analysis includes at least one of principal component regression and partial least squares. 8. The method according to item 1 of the patent application scope, wherein The direct reference assay contains at least one glucose concentration. 40 200538083 9. The method as described in the first item of the scope of patent application, wherein the above adjustment mode includes at least: y meas-yhat + yerr = Xmeasured W + yerr where Xmeasured contains at least one vector of the predicted spectrum, and WT at least A vector containing a correlation coefficient with the regression model, yhat at least contains an analyte concentration generated by the model before any correction is performed, an estimation error of the analyte concentration provided by yerr, and ymeas contains at least the biological attribute concentration. 10. The method according to item 1 of the scope of patent application, wherein the period of using the adjusted mode includes at least one of the following periods: a period of a day; a working day; a sober period of a day; a week; and a month . 11. The method as described in item 10 of the scope of patent application, wherein the direct reference measurement includes at least any of the following: a single reference glucose concentration; < a reference glucose concentration measured before or at the beginning of the period ; And a set of data for at least two glucose concentrations. 41 200538083 1 2 · The method described in item 1 of the scope of patent application, further comprising at least a guide having the same surface and an attachment surface, wherein the sample surface of the guide is detachably connected to The human tissue, and the attachment surface of the guide is detachably connected to at least one of the input element and the output element. 1 3. The method according to item 11 of the scope of patent application, wherein the sample surface of the primer includes at least one of the following: a flat surface; and a surface having a radius of curvature between a plane and 1.5 cm. 1 4 · A non-invasive method for estimating a biological property concentration in a human tissue of a subject, which includes at least the following steps: using a non-invasive spectrum and a related reference concentration to generate a pattern; using at least one direct Adjusting the mode with a reference measurement without using a corresponding reference spectrum; using an analyzer to collect at least a non-invasive spectrum of the tissue of the subject; and using the adjusted mode to estimate the subject's The biological attributes. 15. The method according to item 14 of the scope of patent application, which at least further comprises a guide, wherein the guide couples the analysis instrument to the human tissue in an alternative manner. 16 · — A non-invasive method for estimating a biological property concentration of a subject's human tissue, which includes at least the following steps: Provides a device for measuring light output, which device includes at least one energy that can emit light Source, an input element, an output element, and a spectrum analyzer; A guide is used to couple the input and output elements to the human tissue, wherein the guide includes at least an attachment surface and a same surface; the input element is used to irradiate the tissue with a plurality of wavelengths having the light energy, and At least a part of the wavelengths are absorbed; using the output element to collect at least a part of the unabsorbed light; determining the intensity of the collected light; using a plurality of calibrated non-invasive spectra and a plurality of calibrated reference concentrations to generate a pattern ; Using at least one direct reference measurement and a corresponding reference spectrum to adjust the model; collecting at least one non-invasive prediction spectrum of the subject's tissue; and using the adjusted model to estimate the subject's organism Attributes. 1 7. The method according to item 16 of the scope of patent application, wherein the biological property includes at least a glucose concentration. 43 200538083 1 8 · The method as described in item 16 of the patent application range, wherein the light includes at least a wavelength ranging from 1 100 to 1 900 nm or at least one interval range therebetween. 19 · The method as described in item 16 of the scope of patent application, wherein the non-invasive calibration spectrum includes at least any one of the following spectrums: a single individual; a plurality of individuals; a single analyzer; and a plurality of analyzers. 20. The method of claim 16 in the scope of patent application, wherein the mode includes at least one multivariate analysis. 2 1. The method according to item 20 of the scope of patent application, wherein the multivariate analysis includes at least one of a principal component regression and a partial least square. 22. The method according to item 16 of the scope of patent application, wherein the period of using the adjusted mode includes at least one of the following periods: a period of a day; a working day; a sober period of the day; 44 200538083 a week; And a month. 23. The method as described in claim 22, wherein the direct reference measurement includes at least any of the following: a single reference glucose concentration; a reference glucose concentration measured before or at the beginning of the period; the period The first n reference glucose concentrations in; and a set of data for at least two glucose concentrations. 24. The method as described in claim 22, wherein the reference spectrum includes at least any of the following spectra: a stored reference spectrum; a first spectrum in the period; a spectrum of the subject; in the period An average of the previous n spectra; a set of data consisting of at least two spectra of the subject; and a spectrum from a database, which is selected based on at least one spectral feature. 25. The method according to item 16 of the scope of patent application, wherein the sample surface of the guide is detachably attached to the human tissue, and the attachment surface of the guide is It is detachably attached to at least one of the input element and the output element. 26. The method according to item 16 of the scope of patent application, wherein the sample surface of the guide includes at least one of the following: a flat surface; and a radius of curvature between flat and 1.5 cm. A surface. 27. The method according to item 16 of the scope of patent application, wherein the attachment surface of the guide is engaged with at least one of the following objects: a connector; a light stimulator; a sampling module; the input element; and the output element. 2 8. — A non-invasive method for estimating a biological property concentration of a target human tissue of a subject, which includes at least the following steps: using a non-invasive spectrum and related reference concentrations to generate a pattern; The correlation spectrum is used to adjust the mode, wherein the correlation spectrum is selected based on at least one of the following spectral characteristics that can represent the target tissue; a spectral database; y-a linear combination of spectra; and at least one spectrum of the subject. 46 200538083 29 · A non-invasive method for estimating a biological attribute concentration of a human body tissue of a subject, which includes at least the following steps: Provide a device for measuring light output, the device includes at least multiple wavelengths of infrared radiation An energy source, an input element, an output element, and a spectrum analyzer; coupling the input and output elements to the human tissue; using the input element to irradiate the tissue with multiple wavelengths of infrared light energy and causing at least a portion The wavelengths are absorbed; the output element is used to collect at least a portion of the unabsorbed infrared light energy; determine the intensity of the infrared light energy; use a calibrated non-invasive spectrum and a calibrated reference concentration to generate a mode; use at least A direct reference measurement and not using a corresponding reference spectrum to adjust the model; collecting at least a predicted non-invasive spectrum of the subject's tissue; and using the adjusted model to estimate the subject's organism Property concentration. 47 '