JP2003532461A - Non-invasive measurement method using retinal image - Google Patents

Abstract

(57) Abstract Illumination light (12) of a selected wavelength in the visible or infrared region is projected onto the fundus through the pupil of the eye (10). Light that is later reflected back out of the pupil is detected (22) and analyzed using the area of the optic disc to analyze retinal vessels preferably located above the optic disc. A particular wavelength of the illumination light is selected for each blood component to be analyzed, depending on the spectral characteristics of the substance to be analyzed. Reflection images from the retina are used to measure non-photoreactive blood components, such as hemoglobin, and photoreactive components, such as bilirubin. For measurement of the photoreactive component, an image is taken before, after or during illumination of the eye using light of a wavelength that affects the photoreactive analyte and allows the concentration of the analyte to be measured.

Description

Detailed Description of the Invention

TECHNICAL FIELD The present invention relates to the field of non-invasive in vivo measurement of blood components such as glucose, hemoglobin and bilirubin.

BACKGROUND OF THE INVENTION Measuring the concentration of blood components such as hemoglobin and glucose required the withdrawal of blood samples for in vitro analysis. The need to draw blood for analysis is of several reasons, including discomfort to the patient, the time it takes medical personnel to draw and handle the sample, and the potential risk of spread of the disease due to skin puncture. Not desirable because. Repeated withdrawal of blood samples is not desirable, especially in children. Many diabetics test their blood up to six times a day,
The glucose level in their blood must be monitored. Thus, it would be desirable to have a quick and reliable assessment of the concentration of blood components in the blood, such as hemoglobin and glucose, by a simple, non-invasive technique. Previous efforts have involved the examination of skin or terminal blood such as fingers and ear lobes, or of visible superficial blood, but these efforts have involved tissue components that prevent accurate readings of blood component concentrations only. Has had limited practical success.

DISCLOSURE OF THE INVENTION Each year alone, approximately four million newborns are born in the United States. About 50% of newborns
Jaundice clinically due to high bilirubin levels. If serum bilirubin reaches very high levels during the late birth period, persistently high levels of serum bilirubin may cause nuclear jaundice, or nerve damage. Frequent monitoring of serum bilirubin is important for the care of these infants. 1.7 million newborns found to have jaundice during the first five days of life will have at least one blood test for bilirubin. Approximately 700,000 of the newborns tested are treated with phototherapy, and these babies have an estimated two to three additional blood tests. Currently, blood is drawn from the heel of the newborn, resulting in accidental infections and other complications. Another drawback of this procedure is its high cost and delayed laboratory results that reach the doctor. The currently introduced non-invasive device for measuring bilirubin does not provide the level of accuracy required to diagnose or treat high serum bilirubin levels, making the device practically useless in practice. .

SUMMARY OF THE INVENTION The present invention combines the accuracy of in vitro laboratory testing of blood components with the advantages of rapidly repeatable, non-invasive techniques. The present invention utilizes a handheld or stationary instrument for retinal imaging that allows non-invasive measurement of certain blood components within retinal blood vessels. Illumination light of a selected wavelength in the visible or infrared region is projected onto the fundus into the eye and then reflected back (eg, from the pupil) is preferably the most to be measured. Is detected and analyzed using the area of the optic disc to analyze retinal vessels located on the optic disc for blood components of The particular wavelength of illumination light is selected for each blood component to be analyzed, depending on the spectral characteristics of the particular substance being analyzed. Reflection images from the retina are used to measure blood components such as hemoglobin, glucose and bilirubin.

The use of the retina as a site for obtaining blood component data has several advantages, including the ease of visualizing the data due to the natural window provided by the eye. Light reflected from the fundus of a visually effective wavelength is much less scattered than light reflected from the skin or mucous membranes because the eye is inherently immune to scattering. . Since the retina creates a uniform background for imaging and the ophthalmologist needs to image the retina for diagnosis of disease states, the optical devices and techniques needed to obtain retinal images are widespread. Has been researched and developed. Moreover, blood flow to the retina is highly uniform and reproducible in many medical conditions. For example, a shocked patient has reduced blood flow to the skin and mucous membranes, and current techniques for examining the skin and mucous membranes give false data, but the body is exposed to extremely low blood pressure. Maintain a uniform blood flow to the retina. Moreover, the present invention eliminates the required physical contact between the device and the skin or mucous membranes, thereby eliminating the potential transmission of pathogens associated with the device requiring contact with the patient. The device is placed in the patient's eye for ease of use, and if desired, a disposable plastic cover can be used to further minimize the risk of pathogen infection. Such a cover is transparent and may or may not completely cover the device, depending on the needs of the imaging device and the need to prevent accidental contact with the device.

According to the invention, a handheld or stationary instrument for retinal imaging is used to obtain a non-invasive measurement of a photoreactant, an example of which is serum bilirubin. Illumination light of a selected wavelength in the visible region is projected onto the fundus into the eye. The particular wavelength of the illumination light is chosen so that serum bilirubin can be measured. Analysis of the reflex image from the fundus is used to measure bilirubin. Although measurement of reflected light from blood vessels located on the optic disc is preferred according to the invention, it is also generally possible to obtain bilirubin measurements from light reflected from the fundus of the eye.

During conditions of illness where serum bilirubin levels are above normal (eg, neonatal jaundice), bilirubin is extruded from the choroid into the neural layer of the retina. During the newborn jaundice,
This nerve layer is colored yellow due to high bilirubin levels. This yellow color is
It is directly proportional to high serum bilirubin levels and changes rapidly with changes in serum bilirubin. The bilirubin molecule exhibits a maximum absorption of light at 470 nm. However, when exposed to light at or near this wavelength, the molecule decomposes into optically inactive molecules. The complete bilirubin molecule reflects light at and near the wavelength of about 550 nm (yellow light) and is unaffected by this light. In the present invention, the retina of the patient's eye is first imaged with light at a wavelength of 550 nm with little or no light that does not degrade bilirubin, eg, 470 nm light. The intensity of the reflected light at or near the maximum reflection wavelength of 550 nm is detected. The retina is then followed by light that is projected into the eye, which decomposes bilirubin, for example, followed by light at 550 nm,
Or it is imaged again using 470 nm light combined with 550 nm light. Reflected light at 550 nm from the pupil is detected for the second time to image the retina. With the addition of 470 nm light, the bilirubin molecule is rendered optically inactive and 55
No longer reflects at 0 nm. The difference in reflected image intensity at 550 nm from the first image to the second image is a function of bilirubin concentration. Neural networks or other processing techniques may be used to analyze the two data sets of images captured by the retinal camera.

Further objects, features and advantages will be made apparent from the following detailed description when taken in conjunction with the accompanying drawings.

DETAILED DESCRIPTION OF THE INVENTION The present invention is used to measure a variety of analytes. These analytes may be photoreactive or non-photoreactive. Table 1 below summarizes examples of preferred methodologies according to the invention for the determination of photoreactive and non-photoreactive analytes. However, it should be understood that these are merely examples of analytes that can be measured using the present invention.

[Table 1]

Device Description Referring to the drawings, FIG. 1 illustrates an inventive blood analysis device with the patient's eye illustratively shown at 10 in FIG. A blood analyzer is used to project illumination light, though not necessarily directly, through the pupil to the fundus, to receive light reflected from the fundus and out of the pupil, and to focus the light to form an image. The optical device 11 is composed of a lens for The lens is preferably near the cornea of the eye (eg,
It has a final lens that can be positioned approximately 3 mm from the cornea and that provides a 10-30 degree conical view of the retina to be illuminated and imaged. Such a lens system is of conventional design and is used in the macroview lens system of retinal video cameras.

Illuminator 12 provides selected illumination light for viewing and imaging the retina. The illuminator preferably provides light for viewing and imaging the retina,
It is a monochromatic or multi-discontinuous wavelength light source. Preferably, the illuminator provides viewing light and imaging light coaxially to reduce the possibility of extraneous reflections from inside or outside the eye. Light from the illuminator is projected through the pupil. The frequency components of this light source are selected depending on the compound to be analyzed. The illumination light may consist of two (or more) separate illumination devices, such as a xenon strobe or multi-laser diode for imaging and a halogen light source for viewing. Infrared imaging may be performed using a filtered halogen light source or laser diode light source. The light is reflected from the fundus of the eye 10, passes through the aperture of the pupil of the eye to the optical device 11, and then passes through the illumination device 12, eg, a charge coupled device (CCD) detector 22.
to go into. The illuminator 12 is preferably the same device used in existing non-mydriasis fundus cameras, modified to provide a coaxial design for illumination and imaging. A viewing device 14, such as a liquid crystal display (LCD) screen, receives the image data and displays, in real time, an image used by an operator to initially position the patient's retina based on the image from the optical device. A coaxial "scene" or visual target is included in the field of view of the device so that the patient can gaze at the scene and reduce eye movement. In addition to reducing eye movement, the location of the visual target will bring the optic disc to approximately the center of the CCD detector. For children's devices, the scene may include visually pleasing objects such as familiar animals. In currently commercially available video cameras designed for retinal imaging, an LCD (or other display) screen is typically installed on a tabletop power supply attached to the handheld camera by a cable. Although such a display device may be used in the present invention, the LCD screen (or other) may be used to have the patient's eye and the LCD screen in line of sight to allow the operator to more easily position the retina. Display device) is preferably arranged after the handheld camera unit. The current retinal video camera device modified and utilized in the present invention and including the optical device 11 and the illuminator is a NIDEK NM100 handheld non-mydriatic fundus camera and TOPCON TRC-50EX (TRC-NW5S / TRC-). NW5SF) non-mydriatic retinal camera. The Nidek NM100 camera utilizes a coaxial imager with an infrared source and imaging is accomplished by reflected light outside the optical device. Although the present invention may be performed with the dilated pupil of the eye, it is preferred that imaging of the retina be performed without requiring dilation of the pupil for measurement speed and patient convenience. The camera preferably has a shield (not shown) that prevents ambient light from entering the optical device 11 in order to minimize the introduction of inconsistent reflections and optical noise.

The optical device 11 is also provided for automatically finding and focusing the optic disc,
This also connects to a positioning focus device 16 which utilizes feedback from an image capture device 17 connected to the optical device 11. A convolver or other pattern recognition software may be used to search the optic disc area by finding a circular pattern of the optic area. Using pattern recognition information,
After positioning the optic disc more accurately in the center of the visual field,
The image may be magnified using a series of lenses of the optical device 11 so as to substantially fill the active area (or other detector). The optical device preferably tracks the movement of the fundus while zooming the optics so that the optic disc is centered in and occupies most of the visual field. The optics may be configured to track fundus movements by a motorized drive that slightly supports the lens system with gimbals. The motion device is driven and controlled in a closed loop method utilizing the feedback of pattern recognition software.

The image capture device 17, selectively controlled by the operator, uses feature and pattern recognition to drive the positioning and autofocus device 16 to capture and store an image suitable for analysis. . The image capture itself resembles the function performed by a digital still camera. The image capture device can utilize feature and pattern recognition to drive a positioning and autofocus device to capture and store an image suitable for analysis. It is preferable to use commercially available pattern recognition software. The actual imaging is preferably timed based on the patient's blood flow as sensed by the systolic sensing device 13 from blood vessels through the patient's skin around the structure of the eye. The light reflected from the retina is preferably detected and an image is formed during systole, thus ensuring maximum blood flow within the retinal vasculature. Detection of the systolic state of the patient may be done by any of the commonly known means, such as commercially available blood pressure transducers. The image analyzer 18 is connected to the image capture device 17 and analyzes the light reflected from the retina to quantitatively determine the amount of the particular target analyte compound present. The result may be displayed to the operator via the output device 20. This output device presents the result, as well as any feedback associated with the capture of data, and may include an LCD display screen or other display device.

A modified imaging methodology as shown in FIG. 2 may be used, in which case the imaging device 21 functions in scan mode (which may be similar to that used in a normal video camera). Then, multiple images of the retina are provided to the image capturing device 17. The image recognizer 15 then acts to select a valid image from the series of images stored by the image capture device 17. Once the appropriate image selection has been made, the calculation of the desired analyte size is performed using the image analyzer 18 as previously described.
The effectiveness of an image is based on the focus, the proximity to the center of the fundus, and the tilt of the fundus in the image.

FIG. 3 shows a handheld camera and illumination unit for the analysis device of the present invention with the optical device 11 attached, and FIG. 4 shows the operator observing the image of the retina obtained on a real-time basis. Unit 25 for holding the output observation device 14 that enables the
Shown at the rear end. If desired, a disposable transparent plastic (eg polyethylene, polystyrene, polypropylene, etc.), selected to transmit the required wavelength of light, is used to cover the lighting unit 25 during use, It can then be disposed of to minimize the risk of transmission of the infectious disease. Eye infections are common, especially in newborns, and covers help prevent these infections. Moreover, the cover may include a soft portion (eg, foam) that contacts the skin around the orbit. This makes the cover comfortable and also prevents ambient light from entering the camera lens during the measurement.

Currently available CCD detectors may not be sensitive to wavelengths longer than about 1000 nm. Other strategies may be used when measuring analytes that require sensing of these long wavelengths. Instead, the illumination source is scanned across the region of interest on the retina and the reflected beam is read by a single sensor (or multiple small sensors) sensitive to infrared (IR) wavelengths above the sensitivity of standard CCDs. To be This method may be used to digitally reproduce the retinal image if desired.

As shown in FIG. 5, image processing and analysis centralizes data from image capture device 17 to a remote location (ie, anywhere in the world linked by the Internet) where image analysis device 18 is executed. It may be done at a location remote from the clinical environment by using a wired or wireless internet link (or a dedicated communication link) to send to the computer. The output data from the output device 20 may be returned to the clinic's viewing device 14 (or, if desired, elsewhere) via the access link 29.

Measurement of Non-Photoreactive Analyte (Hemoglobin) In one implementation of the device for detecting hemoglobin using a multiple beam splitter, an image of the optic disc that returns from the eye Divided into three simultaneous equal images. Each image is then 640,766 for the image.
And at a preferred wavelength of 800 nm. Different wavelengths of light are used to analyze other substances such as glucose. These preferred wavelengths would result in an accurate measurement of hemoglobin, but also other wavelengths in the visible or near infrared,
The present invention can be effectively used for hemoglobin.

In order to obtain an accurate measurement of hemoglobin concentration without the need to calculate different levels of oxygen saturation, the hemoglobin measurement is carried out by measuring the wavelength of light at which the extinction counts of oxyhemoglobin and deoxyhemoglobin match. It can be used and practiced in the present invention. Thus, the total absorption of energy at that wavelength is proportional to the total hemoglobin. Examples of such wavelengths are 550 nm and 800 nm. The absorption curves for oxyhemoglobin (HbO ₂ ) and deoxyhemoglobin (Hb) as a function of illumination wavelength are shown in FIG. As shown there, the absorption curve is about 3
Within the wavelength range of 80 nm to 580 nm, especially about 530 nm to 580 nm, there is a near match, and most (or all) wavelengths in these ranges are used.

The area preferably used for analysis is the retinal artery and vein located on the optic disc. Previous work by Hickam et al. Proposed that when the optic disc is illuminated with light directed to the fundus, the optic disc serves as a light source that directs light to the retinal vessels located on the optic disc. "Study of oxygen saturation of retinal venous blood of subjects by photographic method", March 1963, Volume XXVII, pp. 375-3
See 85. At that time, Hickam et al. Believed that these vessels would function as blood-filled cuvettes, which could be analyzed by a spectral pattern measured from the light emitted from the pupil through these vessels. It was In fact, light does not pass through these vessels, does not reflect from the optic disc, and does not pass through these vessels again, as suggested by Hickam et al. Instead, light reflects directly from the surface of these blood vessels and blood components within these blood vessels. In the present invention, the retinal artery is a sub-target area that is analyzed by measuring the magnitude of the signal at each selected wavelength. This is done by determining the effective reflection image intensity of the pixel for the sub-target area corresponding to the retinal artery at each selected wavelength. Within the sub-target area, the maximum concentration of blood is determined by observing the maximum absorption at each measurement wavelength. The sub-target areas are then normalized by this highest concentration and may be averaged using any conventional digital averaging technique such as a convolver. Within the sub-target area, a histogram is created to represent the different signal levels after averaging. These data, which then consist of histogram values and values representing the rate of change of the local intensity, are the direct inputs to an analysis program for analysis, eg a neural network. A computer implemented neural network is preferably utilized to analyze the pixel data for comparison with known blood concentrations. A preferred type of neural network is backpropagation.

The effect of the difference in illuminance from one retinal image to another is preferably eliminated by comparing the background illuminance of the optic disc with the vascular illuminance and “zeroing” the illuminance difference. This ensures that the inputs from the sub-target areas are matched between patients and between cameras.

The present invention is suitable for being implemented on a “real time” basis using the apparatus of FIGS. 1 and 2 as described above, but using scanned and processed photographic fundus images. You may execute it. As an example of the present invention using photographic images, a group of individuals is examined to utilize the present invention to measure hemoglobin levels. Routine in vitro tests were also performed on blood samples drawn from each individual. Clinical data were obtained as follows. Each ophthalmologist agreed to refer patients who had already dilated their eyes for the study. The doctor asked the patient if they could agree to have their eyes photographed and blood drawn for the test for hemoglobin. Informed consent documents were signed by each patient. The retina photographer as a member of the clinic took a picture of each retina. The slide film used was Kodachrome 64 select with 24 frames per roll. The retina photographer took each eye with the fovea in the center of the picture. Exactly the same conditions were used for each picture, including aperture, shutter speed and illumination. Each patient was given a code that assures confidentiality and allows accurate tracking of each patient. The code appeared in each picture and the same code appeared in the blood tube.

A qualified flevotomist obtained a forearm blood sample from each patient. These blood samples were aspirated into purple-headed tubes for later analysis using vacutainers. The tubes were refrigerated and spent the same night. Following the analysis, save a blood tube if further analysis is required. Blood samples were analyzed for hemoglobin content on a HemoCue® B-hemoglobin analyzer. The machine was calibrated with a calibration slide. The device was CLIA approved and the technician who analyzed the sample was also CLIA approved. Each sample
The machine was flushed three times to ensure accurate results. These results are used to identify the patient's ID
Recorded with the code. Kodachrome 64 Select slide film was used for each retinal photograph. The film was sent to Kodak for processing. Kodak Labs used the same lot of chemicals for all of the slides. The developed slides were then scanned in digital format (jpeg file) with a Polaroid Sprintscan 4000 scanner.

Multiple images of each subject were taken, and images were selected for analysis based on illumination uniformity, focus quality, foveal tilt with respect to the illumination axis, and systolic state. The (normalized) green part of the RGB image data in the JPEG file was used for analysis, where green = G / (R + G + B). As shown in FIG. 11, absorption of oxyhemoglobin and deoxyhemoglobin is similar in the green wavelength range (about 500 nm to 550 nm). Image analysis proceeded to the next step for each selected image.

Find the optic disc. Create a circular region of data at the optic disc that is smaller in diameter than the optic disc. The data is smoothed by wrapping the data in this circle with a small convolver (0.5 mm). Find the maximum and minimum levels of green in the pixels within the circle. This lowest level is used to standardize the values of all pixels within the circle. Calculate the gradient of green in a circle. Find the centroid of gradient magnitude. Within the optic disc, an image circle (sub-target) centered on the centroid and having a diameter of about 0.5 mm is constructed. (This procedure positions the sub-target region within the retinal vessels of the optic disc.) Use data in pixels with a green magnitude within 25-45% of the maximum level of green in the sub-target. Then, a similarity diagram (histogram) of this region is created with 5 buckets for the value and 5 buckets for the derivative size. For each size range, calculate the average derivative (slope) size and the average size.

These values are then inputs for a learning algorithm, such as a neural network simulation trained to associate the measured hemoglobin with the data set.

The neural network employed for purposes of illustration is a “backpropagation” neural network, which is included in Matlab Toolbox of MATLAB Version 5.3. .

The neural network is trained with the previously determined data set by the following technique. (A) Remove one data set (representing one image) from the total number of sets. (B) Train the neural network described above using the remaining dataset. (C) Calculate the hemoglobin value from the data set excluded in (a) above using a trained neural network

By repeating the above neural network calculations, a curve of non-invasively calculated values versus actual experimental values is created.

Correlation data showing hemoglobin concentration measurements for each individual as determined by the present invention versus hemoglobin measurements obtained by in vitro experimental measurements are shown in FIG.
Shown in. These data show a strong correlation between the measurements obtained according to the invention and those obtained by standard in vitro laboratory techniques, ie 0.89.

The result of analysis of the serum concentration value of the blood component is displayed to the operator on the output device 20. The output device 20, utilizing the LCD screen of the display device, presents the data to the operator and provides feedback on image capture or any problems with the device. If the image is not good and the image does not cause the device to calculate an accurate blood value, the output device 20 preferably causes the operator to capture another image.

In a subset of the patient population, different degrees of light scattering may be encountered due to particular anatomical eye characteristics. This light scattering may be the result of various disease states, including senile cataract. A polarized light source is included in the device to measure the degree of scattering. The orthogonally polarized return light is used as a measure of scatter. This data is fed into the neural network as a patient-to-patient scatter variable and the device is calculated for this variable.

Some analytes measured with the device of the invention are also secreted into the anterior chamber of the eye. The previously described embodiments of the device direct light through the anterior chamber of the eye to the retina. The image returned from the retina passes through this area of the eye again. The presence of analytes in this part of the eye could create erroneous data sets, making the instrument inaccurate. Therefore, when measuring analytes such as glucose present in the anterior chamber of the eye, light should be directed along the alternate path to the retina. Illumination light emitted from the device should be directed towards the retina at an angle into the eye just beside the cornea near the strong corneal junction (cornea limbus). This angle avoids light passing through the anterior chamber of the eye. The exiting light would then be detected on the opposite side of the eye, at an angle, again just beside the cornea near the scleral junctions (cornea limbus). This region near the surface of the eye, although not optically clear in the visible range of the electromagnetic spectrum, is transparent at longer wavelengths (> 1000 nm) required to measure analytes such as glucose.

(Measurement of Photoreactive Analyte (Bilirubin)) Further, according to the present invention, the concentration of the photoreactive analyte is determined by utilizing its photoreactivity. For example, the bilirubin concentration is 470
It is determined by the fact that it exhibits a maximum absorption of light in nm and that at or near this wavelength the bilirubin molecule decomposes into optically inactive molecules. Bilirubin molecules also effectively decompose when exposed to light in the range of 470 nm ± 30 nm. Bilirubin molecules are usually at or near the wavelength of about 550 nm (yellow light),
Maximum reflection of light. According to the invention, a time-lapse image of the retina is created in which the first image is illuminated at 550 nm (± 30 nm), followed by the second image at 470 nm (± 30 nm) and the third. Image is 550nm (± 30nm)
Illuminated by. The second and third images are separated by sufficient time for the photoreactive analyte to change its chemical state. In the case of bilirubin, the change is almost instantaneous and the two wavelengths should be projected simultaneously and consecutively. The first irradiation is
An image containing a photoreactive analyte (bilirubin in this example) is provided and the third irradiation is
Provide a reference point for this patient without a photoreactive analyte (bilirubin in this example). The concentration of the photoreactive analyte can be calculated based on the difference between the first irradiation and the third irradiation using a neural network or other known analysis method. For light with a wavelength other than 550 nm, such a wavelength has an analyte (bilirubin in this example).
It can be used to illuminate the retina as long as it does not promote the photochemical reaction from the.

In one embodiment of the present invention, referring to FIG. 6, using bilirubin as the target analyte, light at a wavelength of 550 nm is directed onto the partially transmissive reflective mirror 31 on the beam 30. It passes through and is directed on the optical path 32 to the beam splitter 34 and thus to the eye 10. Light with a wavelength of 550 nm is directed to the retina,
Make the first image. The light from the 470 nm light source is given to the partial mirror 31 on the optical path 36, and is reflected on the optical path 32 to the eye 10 through the beam splitter 34. The 470 nm reflected light is directed by element 34 on path 37 to absorber 38. The 470 nm light is pulsed onto the retina at very short time intervals. Four
With each pulse of light at 70 nm, the bilirubin molecule is rendered optically inactive, and the bilirubin molecule is not reflected from the retina at 550 nm. As shown in FIG. 7, the intensity of the reflected light of 550 nm is high in the interval 40 when there is no illumination of the light of 470 nm and low in the interval 41 when the light of 470 nm is applied. The difference in reflected light intensity from peak to valley for each pulsation is a function of bilirubin concentration. The time intervals 40 and 41 selected depend on the blood flow to the retina. FIG. 8 shows changes in blood flow and time. The time interval is required so that there is a partial filling of the retinal circulation with blood containing bilirubin. The present invention is also practiced with the application of a single pulse of 470 nm light to measure bilirubin concentration, as shown in FIG.

The data corresponding to the 550 nm return light that makes up the pixel value is then the direct input to the neural network for analysis or other known procedure. Neural networks are preferably utilized to analyze pixel data for comparison with known blood concentrations of bilirubin. The difference in pixel data from the 550 nm reflected light before and after the 470 nm light pulse is a direct function of serum bilirubin level.

The result of the serum level of bilirubin may be displayed to the operator on the output device 20. The output device 20 may utilize the LCD screen of the viewing device to present the data to the operator as well as to feed back any image capture or any problems that may occur with the device. If the image is not good and does not allow the device to calculate accurate bilirubin blood, the output device 20 preferably causes the operator to capture another image. Other photoreactive analytes may be measured in a manner similar to the measurement of bilirubin, as discussed above.

It is understood that the invention is not limited to the embodiments described herein by way of example, but encompasses all such forms as come within the scope of the claims.

[Brief description of drawings]

[Figure 1]   1 is a schematic diagram of an apparatus for measuring the concentration of blood components according to the present invention.

[Fig. 2]   FIG. 6 is a schematic diagram of a modification of the blood component measuring device according to the present invention.

[Figure 3]   1 is a schematic side view of a handheld lighting and camera device used according to the present invention.

[Figure 4]   FIG. 4 is an exemplary front view of the lighting and camera device of FIG. 3.

FIG. 5 is a schematic diagram of a further device according to the invention having a communication link with a remote processing device.

FIG. 6 is a simplified schematic diagram of a device according to the invention for measuring the concentration of bilirubin in blood.

FIG. 7 is an exemplary graph of the intensity of light at 550 nm reflected from the retina as a function of time showing the effect of periodic illumination of the eye at 470 nm.

FIG. 8 is an exemplary graph of time and reflected light intensity, similar to FIG. 7, showing the effect of blood flow changes in retinal vessels.

FIG. 9 is a graph of the intensity of 550 nm light reflected from the retina showing the effect of 470 nm monopulse illumination.

FIG. 10 is a graph illustrating the correlation between hemoglobin concentration measured according to the present invention and concentration measured by standard in vitro techniques.

FIG. 11 is a graph of light absorption as a function of wavelength for deoxyhemoglobin and oxyhemoglobin.

[Procedure for Amendment] Submission for translation of Article 34 Amendment of Patent Cooperation Treaty

[Submission date] November 9, 2001 (2001.11.9)

[Procedure Amendment 1]

[Document name to be amended] Statement

[Name of item to be amended] Claims

[Correction method] Change

[Contents of correction]

[Claims]

─────────────────────────────────────────────────── ─── Continued front page    (81) Designated countries EP (AT, BE, CH, CY, DE, DK, ES, FI, FR, GB, GR, IE, I T, LU, MC, NL, PT, SE), OA (BF, BJ , CF, CG, CI, CM, GA, GN, GW, ML, MR, NE, SN, TD, TG), AP (GH, GM, K E, LS, MW, SD, SL, SZ, TZ, UG, ZW ), EA (AM, AZ, BY, KG, KZ, MD, RU, TJ, TM), AE, AG, AL, AM, AT, AU, AZ, BA, BB, BG, BR, BY, CA, CH, C N, CR, CU, CZ, DE, DK, DM, DZ, EE , ES, FI, GB, GD, GE, GH, GM, HR, HU, ID, IL, IN, IS, JP, KE, KG, K P, KR, KZ, LC, LK, LR, LS, LT, LU , LV, MA, MD, MG, MK, MN, MW, MX, NO, NZ, PL, PT, RO, RU, SD, SE, S G, SI, SK, SL, TJ, TM, TR, TT, TZ , UA, UG, UZ, VN, YU, ZA, ZW (72) Inventor Liu James M             United States Kentucky 40517               Lexington Appian Way 3751               # 132 (72) Inventor Williams William Tee             United States Tennessee 37659 The             Jones Born Chestnut Grove               Road 360 (72) Inventor Laut Wilson             United States Kentucky 40503               Lexington Arcadia Park             119 F term (reference) 2G059 AA01 AA06 BB12 BB13 CC16                       CC18 EE02 EE13 FF04 GG01                       GG10 HH01 HH02 HH06 JJ02                       JJ11 KK04 MM02 MM03 MM09                       MM10 PP04                 4C038 KK10 KL07 KX01

Claims (21)

[Claims] 1. A method for detecting the concentration of a blood component of an individual, comprising: (a) projecting light having a selected wavelength into the eye of the individual to illuminate the fundus; and (b) from the fundus. Detecting reflected light to form an image of a portion of the fundus containing blood vessels, and (c) analyzing selected wavelength components of the detected light from the blood vessels in the image to select the selected blood. Determining the concentration of the component. 2. The method of claim 1, wherein the selected blood component is hemoglobin. 3. The method of claim 1, wherein the selected blood component is glucose. 4. The method of claim 1, wherein the selected blood component is bilirubin. 5. The method of claim 1, wherein in the step of analyzing wavelength components only light from regions of the image corresponding to blood vessels is analyzed. 6. The light projected into the eye includes wavelengths in the visible region or near-infrared region, and detecting the light reflected from the fundus detects the light in the visible region or near-infrared region. The method of claim 1, comprising: 7. The image forming step forms an image of blood vessels in a region located on the optic disc of the eye, and the step of analyzing the wavelength component analyzes the light reflected from the optic disc. The method according to Item 1. 8. Detecting blood flow in the individual being examined to detect systole, this information detecting the light reflected from the fundus and timing the forming of the image. The method of claim 1 used to do. 9. The method according to claim 1, wherein multiple images of the fundus are formed in discontinuous time over a certain period of time, one image is selected from the multiple images, and the images are analyzed to determine the concentration of blood components. The method described. 10. The image is divided into multiple images, the multiple images are filtered through different wavelengths to obtain a filtered multiple image, and the filtered multiple images are analyzed to obtain blood images. The method of claim 1, wherein the concentration of the component is determined. 11. The blood component is hemoglobin, the multi-image is 640 nm,
11. The method of claim 10, wherein the method is filtered to pass light concentrated at wavelengths of 766 nm and 800 nm. 12. The method of claim 1, comprising sending data corresponding to the image to a remote location over a communication link, the step of analyzing selected wavelength components being performed at the remote location. 13. The method of claim 1, wherein projecting light into the eye is performed by projecting light through the pupil of the eye and onto the fundus. 14. The method of claim 1 wherein the step of projecting light into the eye is performed by projecting the light onto the track just beside the cornea, ie at an angle that is near the scleral corneal junction. Method. 15. A method for detecting the concentration of a photoreactive analyte in the blood of an individual, comprising: (a) applying light containing a wavelength that does not include a wavelength of light that decomposes the photoreactive analyte. Illuminating the fundus by projecting into it and detecting the light reflected from the fundus to determine the intensity of the reflected light at these wavelengths, and (b) light containing wavelengths that decompose the photoreactive analyte. To illuminate the fundus by projecting light into the eye, and illuminating the fundus by projecting light containing a wavelength that does not decompose the photoreactive analyte into the eye.
And detecting the intensity of such light reflected from the fundus, and (c) decomposing the photoreactive analyte, as well as the intensity of the light having no wavelength that decomposes the photoreactive analyte and is projected and detected on the fundus. A method of determining the difference between the detected light intensity and the intensity of the light when projected onto the fundus, and determining the concentration of the analyte from the difference. 16. The method of claim 15, wherein the target analyte is bilirubin or a related molecule. 17. In the step of projecting light which does not decompose bilirubin,
17. The method of claim 16, wherein the projected light does not include wavelengths in the range of 470 nm ± 30 nm. 18. The method of claim 17, wherein the projected light that does not degrade bilirubin comprises light concentrated at 550 nm. 19. The step of projecting light that decomposes bilirubin is 470n.
18. The method of claim 17, comprising projecting light having a wavelength in the range m ± 30 nm. 20. The method of claim 20, wherein step (b) comprises projecting light comprising wavelengths of 470 nm and 550 nm simultaneously. 21. Step (b) is performed by projecting light of 470 nm and then projecting light containing a wavelength of 550 nm to illuminate the fundus, and detecting light of 550 nm reflected from the fundus. The method of claim 16, wherein the method comprises: