Whole blood hemoglobin saturation and concentration optical sensor

Abstract

A non-invasive optical sensor device for determining hemoglobin concentration and oxygen saturation of blood and a method for calibrating an optical sensor device.

Description

WHOLE BLOOD HEMOGLOBIN SATURATION AND CONCENTRATION OPTICAL SENSORCROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims the benefit under 35 U.S.C. § 119 of U.S. Provisional Application Serial No. 63 / 749,142, filed on January 24, 2025, which is incorporated herein by reference.STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH

[0002] This invention was made with government support under U.S. Army Med Research grant number W81XWH2210304 / AWD00001680 / OSP00004656. The government has certain rights in this invention.BACKGROUND OF THE INVENTIONField of the Invention

[0003] The field of the invention is optical spectroscopy and, more specifically, optical spectroscopy to measure attributes of bodily fluids.Background of the Invention

[0004] The present invention is an optical sensor developed for non-invasively monitoring characteristics of whole blood, including, without limitation, oxygen saturation and hemoglobin concentration, through an optically transparent (of variable opacity) tube, such as used in dialysis, in an Extracorporeal Membrane Oxygenation (“ECMO”) setting, and other settings including but not limited to cardiopulmonary bypass, apheresis, ex-vivo perfusion, surgical drains, etc. The cross-section of the tube can have varying shapes, including, without limitation, a circle, oval, rectangle or square ( / .< ., a cuvette). There exists a need for portable and real-time analysis of blood oxygenation in circulatory support systems to replace sampling and ex vivo analysis of the blood. The present invention performs thisanalysis to assess ECMO functionality. The various embodiments of a system of the present invention are advantageous over current clinical monitors (e.g. Spectrum M4 monitor and Terumo CDI 500) as they are designed to integrate with both standard ECMO systems and newly developed wearable ECMO systems. Typically, ECMO systems are bulky and the associated monitoring equipment is not optimized in terms of footprint (size.) Wearable devices like pulmonary assistant systems (“PAS”) — consisting of an axial flow pump and a small biocompatible gas exchanger — provide the benefit of supporting patients (e.g. when performing ambulation) with reduced staff while still maintaining the same standard of care. The portability aspect of wearable ECMO necessitates that the associated bulky monitoring equipment, such as the hemoglobin and oxygen saturation monitors, be reduced in size as well. Other monitors in the field, such as the Spectrum M4 monitor weighs 4.5 kg and has a 26.4 cm display and the Terumo CDI 500 weighs 2.8 kg with a 28x32x15cm display. The present invention improves upon these monitors through the development of a miniaturized optical monitor device and system. By addressing this need, sensors of the present invention can allow wearable ECMO systems to be translated clinically.BRIEF SUMMARY OF THE INVENTION

[0005] One embodiment of the present invention is an optical sensor device for measuring characteristics of bodily fluids (in an optically transparent tube,) the sensor device comprising the following components: (i) a housing having a first half and a second half, which operate together to clamp onto tubing; (ii) two or more light sources connected to the second half of the housing; (iii) a printed circuit board connected to the first half of the housing; (iv) a microcontroller unit connected to the printed circuit board; (v) a photometric analog front end register connected to the printed circuit board; and (vi) at least one photodiode connected to the first half of the housing, positioned to receive light transmitted from the at least two or more light sources and through the tubing, and operable to generate an optical return signal. For thisembodiment, the microcontroller unit is configured to program the photometric analog front end register to (i) stimulate the two or more light sources with a predetermined number of pulses, (ii) obtain at least one measurement from the return signal, and (iii) transmit the at least one measurement to the microcontroller unit for use in computing at least one characteristic of the bodily fluid in the tubing, through a calibration of the at least one measurement.

[0006] Another embodiment of the present invention is a method of calibrating a sensor device by performing the following steps: (i) collecting intensity data of a first light source and of a second light source of an optical sensor device and collecting at least one ground truth blood gas measurement of a predetermined gas of in vitro blood; (ii) dividing the intensity data of each light source by a predetermined pulse number to compute a scaled intensity for each light source and computing logarithms of the scaled intensities of the first light source and the second light source and a ratio of the logarithms of the scaled intensities of the first light source and the second light source; (iii) inputting the logarithms, the ratio, and the blood gas measurement(s) into a preselected model for hemoglobin concentration; (iv) fitting the preselected model to the data to extract coefficients for the model and using the coefficients for a hemoglobin concentration prediction; (v) applying a linear fit of the ratio of the logarithms versus oxygen saturation from the blood gas measurement(s) for every hemoglobin concentration value from the preselected model; (vi) extracting slope(s) and intercept(s) of the linear fit; (vii) applying an additional linear fit to the extracted slope(s) and intercept(s) as a function of hemoglobin; (viii) using the hemoglobin concentration prediction to estimate a slope and an intercept for an oxygen saturation model; (ix) using the estimated slope, the intercept, and the measured ratio to predict oxygen saturation. For some embodiments, the final step is reporting the predicted hemoglobin concentration and oxygen saturation.BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWINGS

[0007] To facilitate understanding of the invention, the drawings and description illustrate preferred embodiments thereof, from which the invention, various embodiments of its structures, construction and method of operation, and many advantages, may be understood and appreciated. The drawings are incorporated by reference.

[0008] Figures 1 A through IF are various views of one embodiment of a sensor device of the present invention;

[0009] Figures 2A and 2B are various views of a second embodiment of a sensor device having three lasers;

[0010] Figures 3 A and 3B are block diagrams of two analogous embodiments of sensor devices of the present invention;

[0011] Figures 4A through 4D are graphs of the calibration factors of one embodiment of the present invention;

[0012] Figure 5 illustrates an experiment using the present invention across time intervals;

[0013] Figure 6 is a graph of the ratio R as a function of oxygen saturation at different hemoglobin concentrations;

[0014] Figures 7A and 7B are graphs of the slope and intercept as influenced by hemoglobin concentrations;

[0015] Figures 8A through 8H show the predicted hemoglobin concentration and oxygen saturation from two sensor devices of one embodiment of the present invention plotted against blood gas measurements;

[0016] Figure 9 illustrates one embodiment of a device and method of the present invention;

[0017] Figure 10 illustrates one embodiment of a calibration method of the present invention;

[0018] Figures 11A through 11D show the predicted hemoglobin concentration and oxygen saturation from two sensor devices of a second embodiment of the present invention plotted against blood gas measurements;

[0019] Figure 12 shows one embodiment of a complete sensor device with a PCB affixed to the enclosure;

[0020] Figures 13 A and 13B illustrate the tops of one embodiment of a photodiode attached to a board and a laser attached to a board, respectively;

[0021] Figures 14A and 14B illustrate the bottoms of one embodiment of a photodiode attached to a board and a laser attached to a board, respectively;

[0022] Figures 15A and 15B illustrate the interiors of the sensor device housing;

[0023] Figure 16 illustrates one embodiment of a sensor device clamped onto a blood-filled tube; and

[0024] Figure 17 is a schematic of an in-vitro recirculating blood circuit.DETAILED DESCRIPTION OF THE INVENTION

[0025] The following describes example embodiments in which the present invention may be practiced. This invention, however, may be embodied in many ways, and the description provided herein should not be construed as limiting in any way. Among other things, the following invention may be embodied as a system, method, or device. The following detailed descriptions should not be taken in a limiting sense. The accompanying drawings are hereby incorporated by reference.

[0026] The phrases “in some embodiments”, “in one embodiment”, “in various embodiments”, “according to various embodiments”, “in the embodiments shown”, “in other embodiments”, and the like generally mean the particular feature, structure, or characteristic following the phrase is included in at least one embodiment of the present invention, and maybe included in more than one embodiment of the present invention. In addition, such phrases do not necessarily refer to the same embodiments or to different embodiments.

[0027] If the specification states a component, element, part, or feature “may,” “can,” “could,” or “might” be included or have a characteristic, that particular component or feature is not required to be included or have the characteristic.

[0028] In this document, the terms “a” or “an” are used, as is common in patent documents, to include one or more than one. In this document, the term “or” is used to refer to a nonexclusive “or ” such that “A or B” includes “A but not B,” “B but not A,” and “A and B,” unless otherwise indicated. Furthermore, all publications, patents, and patent documents referred to in this document are incorporated by reference herein in their entirety, as though individually incorporated by reference. In the event of inconsistent usages between this document and those documents so incorporated by reference, the usage in the incorporated reference(s) should be considered supplementary to that of this document; for irreconcilable inconsistencies, the usage in this document controls.

[0029] For purposes of the description hereinafter, the terms “upper”, “lower”, “right”, “left”, “vertical”, “horizontal”, “top”, “bottom”, “lateral”, “longitudinal”, and derivatives thereof shall relate to the invention as it is oriented in the figures. However, it is to be understood that the invention may assume alternative orientations, variations, and step sequences, except where expressly specified to the contrary. It also is to be understood that the specific devices and processes illustrated in the attached drawings and described in this specification are simply exemplary embodiments of the invention. Hence, specific dimensions and other physical characteristics related to the embodiments disclosed herein are not to be considered as limiting.

[0030] The materials described hereinafter as making up the various elements of the embodiments of the present disclosure are intended to be illustrative and not restrictive. Manysuitable materials that would perform the same or a similar function as the materials described herein are intended to be embraced within the scope of the example embodiments. Such other materials not described herein can include, but are not limited to, materials that are developed after the time of the development of the invention, for example.

[0031] While the disclosure has been described in detail and referring to specific embodiments thereof, it will be apparent to one skilled in the art that various changes and modifications can be made without departing from the spirit and scope of the embodiments. Thus, it is intended that the present disclosure covers the modifications and variations of this disclosure provided they come within the scope of the appended claims and their equivalents.

[0032] It is to be understood that the invention may assume alternative variations and step sequences, unless specified to the contrary. It also is to be understood that the specific devices and processes illustrated in the attached drawings and described in this specification are simply exemplary embodiments of the invention. Hence, specific dimensions and other physical characteristics related to the embodiments disclosed are not to be limiting.

[0033] As a vital component in metabolism, oxygen must be reliably transported to tissue. Blood serves as the medium for oxygen transport, wherein oxygen is bound to hemoglobin within red blood cells. As such, accurate measurements of whole blood oxygen saturation (“SO2”) and hemoglobin (“Hb” or “Hgb”) concentration are critical in a variety of clinical situations (e.g., anemia. .This need for accurate measurements is especially important in the context of extracorporeal systems like dialysis, cardiopulmonary bypass, and extracorporeal membrane oxygenation (“ECMO”), which require regular monitoring of vital physiological parameters to ensure proper patient treatment and management.

[0034] Benchtop blood gas analyzers are the gold standard for measuring patient oxygen saturation and hemoglobin levels. However, this method is time-consuming, requires invasive blood sampling, and does not allow for continuous monitoring. Additionally,traditional blood monitoring systems are large and bulky. These drawbacks limit physicians’ ability to implement therapeutic changes in a timely fashion. Thus, there is a need to develop non-invasive devices that allow for real-time measurements.

[0035] Optical spectroscopy is a non-invasive technique that examines the interaction of light with tissue or bodily fluids to extract biochemical information. Depending on the wavelength of light used and the signal of interest, various optical methods can be used for diagnostics. For instance, absorption spectroscopy measures the attenuation (optical density) of light through tissue or bodily fluids for the determination of chromophore (absorber) concentrations.

[0036] For whole blood, hemoglobin is the primary absorber of light. The optical properties of hemoglobin change with oxygenation. Using at least two wavelengths, one can differentiate oxygenated and deoxygenated hemoglobin. For homogenous solutions, the absorption coefficient and thus concentration can be calculated by means of the Beer-Lambert Law (“BLL”). The equation assumes homogeneity as well as the solution being purely absorbing (no scattering). In the case of blood however, solutions cannot be assumed to be purely absorbing. As hemoglobin is encapsulated within erythrocytes (red blood cells), a mismatch in refractive index occurs between the membrane and blood plasma resulting in light scattering. This scattering causes additional attenuation of light separate from absorption that prevents the use of the BLL, making extraction of whole blood optical properties (and by extension saturation and Hgb concentration) challenging. Often light scattering in tissue can be addressed through diffuse optical methods; however, in the case of whole blood, light is subdiffuse preventing the use of these methods. One goal of the present invention is to describe optical sensor devices 1 that provide real-time measurements of oxygen saturation and total hemoglobin concentration in whole blood using an empirically developed calibration method 200. The goal of developing such sensor devices 1 is to assess the functionality of ECMO gasexchangers and patient health with a portable and light-weight device 1 unlike clinically available monitors. By clamping a pair of these sensor devices 1 to the blood-filled gas exchanger inlet and outlet tubing 55, oxygen transfer can be assessed in real-time. The accuracy of these sensor devices 1 is assessed using a novel empirical calibration method 200 developed and assess their potential for clinical use in extracorporeal system settings for direct extraction of hemoglobin saturation and concentration.

[0037] One embodiment of the present invention is an optical sensor device 1 developed for non-invasively monitoring blood characteristics, including, without limitation, oxygen saturation, hemoglobin concentration, hematocrit, flow, water content (i.e., any attribute able to be measured using optical wavelengths) of whole blood (or another bodily fluid) through an optically clear tube 55 (“tube 55” and “tubing 55” are used interchangeably herein,) as used in dialysis and in an ECMO setting. The disclosed sensor device 1, in one embodiment as described herein, uses optical transmission of two or more light sources 60 in the near-infrared range (i.e., 680 nm, 850 nm, 808 nm, as non-limiting examples) and computes oxygen saturation and hemoglobin concentration using an empirically developed calibration method 200. As used herein, “light source 60” is a generic reference to a compatible source of source for various embodiments of the invention including a “first laser 65,” a “second laser 70,” and a “third laser 808” or a “first light source 65,” a “second light source 70,” and a “third light source 808.” In other embodiments, the range of light sources 60 can vary from 630-2000 nm and the number of light sources 60 can vary beyond the one to three light sources 65, 70, 75 described herein. The sensor device 1 is lightweight and designed, in one embodiment, to be clamped onto medical-grade tubing 55 (as shown in Figure 16) without needing a specialized connector, making it appropriate for monitoring during ambulation and for use in conjunction with existing medical technologies.

[0038] Historically, various approaches to overcome the influence of sub-diffuse blood scattering for extraction of hemoglobin concentration and oxygen saturation have been proposed. These include Twersky’s multiple scattering theory for biological suspensions (e.g. blood) which provides a mathematical formulation that gives the total attenuation of light incident through a suspension of large, low-refracting, and absorbing particles. Modified photon diffusion models also delineate absorption and scattering in whole blood for optical property extraction. These approaches are complex, making them difficult to implement algorithmically. Additionally, they often require accurate knowledge of input light intensity, detector size and aperture, and prior knowledge of geometry. In contrast, empirical calibration approaches that are simple to implement have been utilized allowing for direct extraction of hemoglobin saturation and concentration.

[0039] Experimental studies demonstrate comparable performance of various embodiments of the sensor device 1 to current, clinically available technologies, specifically Terumo CDI 500 and Spectrum Medical M4 monitors. The housing 2 is designed to clamp on to standard medical-grade Tygon tubing 55, thus avoiding the need for a custom connector and making device translation simpler. Additionally, embodiments of the sensor device 1 are lightweight (weighing less than 50-grams) and smaller than comparative technologies currently available (measuring about 4.8x5-cm.) These parameters are a significant reduction in size compared to currently available clinical monitors as the smaller of the Spectrum and Terumo devices weighs 2.8 kg with its smallest dimension being 15 cm. This size advantage is useful for translation of the sensor devices 1 into existing ECMO systems and enabling use of the sensor devices 1 with wearable technologies. Current technologies cannot be transported by ambulatory patients the way embodiments of the sensor devices 1 can be. Additionally, embodiments of the sensor devices 1 can measure the amount of blood oxygenation without needing to sample the blood themselves and the related needs of transporting the sample tobenchtop analysis equipment and then reporting the oxygenation back. Embodiments of the sensor devices 1 can determine blood oxygenation non-invasively and in real-time.

[0040] In one embodiment, an empirically developed calibration method 200 is employed to develop optical sensor devices 1 that provide real-time oxygen saturation measurements and total hemoglobin concentration in whole blood (among other types of blood measurements). These sensor devices 1 are designed to assess the functionality of ECMO gas exchangers and patient health with a portable and lightweight sensor device 1, unlike clinically available monitors. In one embodiment, oxygen transfer can be assessed in real-time by clamping a pair of these sensor devices 1 to a blood-filled gas exchanger inlet and outlet tubing 55. The accuracy of these sensor devices 1 can be evaluated using the calibration method 200 and the sensor devices 1 potential for clinical use in extracorporeal system settings is established. The setting in which the sensor device(s) 1 are being applied is invasive itself as a treatment method in that blood is removed from a patient and pumped through tubing 55 to a separate device (artificial lung, dialysis machine, etc.) However, the sensor devices 1 are noninvasive in that they only make contact with the tubing 55, not the blood, and thus are considered to be non-invasive.

[0041] Various embodiments of the optical sensor devices 1 were developed for the extraction of whole blood oxygen saturation and total hemoglobin concentration but can be modified to measure other optically-determinable blood characteristics. The sensor devices 1, in one embodiment were calibrated and validated on an in-vitro recirculating blood loop 300 (an example of a schematic of an in-vitro recirculating blood circuit 300 is shown in Figure 17.) Other forms of calibration can be used in other embodiments, including, without limitation, the use of optical phantoms.

[0042] With reference to Figures 1A through IF, one non-limiting embodiment of a sensor device 1 is shown and is described below. It is to be understood that the componentscan vary in other embodiments as will be known to one skilled in the art. The characteristics of the substituted components can have an impact on the function and characteristics of the sensor device 1. Optical sensor devices 1 function based upon light transmission through a blood-filled tube 55. Again, tube 55 can have varying cross-sectional shapes. Figures 1A illustrates one embodiment of a laser board 5 with top 5A (right side of Figure 1A) and one embodiment of a photodiode board 6 with top 6A (left side of Figure 1 A). Figure IB illustrates one embodiment of a printed circuit board 35 (“PCB 35”) and a header 10 (or connector 10) to deliver power from one side of the board to the other via a plurality of wires 11. Figure IB also illustrates a first laser 65 and a second laser 70 and a photodiode 80. For different embodiments, PCB 35 can be the same as the photodiode board 6 or it can be a separate board (PCB 35) that is secured to the photodiode board 6. Additionally, it will be understood to one skilled in the art that the laser board 5 and the photodiode board 6 are both PCBs 35, meaning both are boards that support electronics. One is a PCB 35 supporting a photodiode 80 (the photodiode board 6) and one is a PCB 35 supporting a laser or light source 60 (the laser board 5).

[0043] Figures 1C and ID are two perspective views of the housing 2 of one embodiment of a sensor device 1. The housing 2 comprises two halves (for ease of reference a first half 3 and a second half 4.) The two halves 3, 4 are comprised of a strong material, like aluminum, other metal, strong plastic or a composite, that can support and give structure to the device 1, including withstanding the compression forces on the housing 2. The first half 3 of housing 2 supports the photodiode board 6. The second half 4 of housing 2 supports the laser board 5. Also shown in Figures 1C and ID are four corner screws 50, which are used to compress the tubing 55 (for this embodiment, 2-mm compression) and to hold together the first half 3 and the second half 4 of the housing 2, with two adjustment screws 51, which act as a lock around the tubing 55. A plurality of other attachment mechanism 52 can be used to attachthe photodiode board 6 and the laser board 5 to the two sides of the housing 2 as will be obvious to one skilled in the art. For other embodiments of sensor devices 1, different securing technologies (beyond screws) can be used in different locations and different numbers of attachment mechanisms 50, 51, 52 to secure the halves 3, 4 of each housing 2 together and to control for and lock at a predetermined amount of pressure / compression on the tubing 55. Such substitutions will be obvious to one skilled in the art.

[0044] Axial cross-sections of the optical sensor device 1 are shown in Figures IE and IF, which depict an optical scheme and the interface of the sensor device 1 with the tubing 55.Figures IE and IF better illustrate the first half 3 of house 2 and the second half 4 of housing 2. Also suggested or shown is the first half interior side 23 of housing 2, the first half exterior side 24 of housing 2, the second half interior side 25 of housing 2, and the second half exterior side 26 of housing 2. For illumination, and as an example of one embodiment of a sensor device 1, emitting lasers 65, 70 with wavelengths at or approximately at 680 nm (such as VD-0680C-005M-1A-2A0, BrightLaser, Hong Kong) and 850 nm (such as VD-0850C-010M-1A-2A0, BrightLaser, Hong Kong) respectively are used and located within two vertical cavity surface 130 of the housing 2 (see Figure 15.) In other embodiments, other solid state light sources 60 can be used. A 7 mm2photodiode 80 (such as BPW34, ams OSRAM, Premstatten, Austria) measures the light transmitted through the tube 55. In other embodiments, other similar photodiodes 80 can be used. The number of photodiodes 80 needed depends upon the type of photodiode 80 used and the number of light sources 60. For example, an ADPD 1080 95 allows for a single photodiode 80 to be used to measure two lasers 65, 70 in sequence. For an embodiment with 3-wavelengths 65, 70, 75, a second photodiode 80 can be necessary. Ultimately, the number of photodiodes 80 needed is determined by the desire to differentiate between various light sources 60 and the amount of signal that is needed. The number of light sources 60 and the type of light source 60 are chosen based upon the blood attribute seeking tobe measured. As non-limiting examples of other body fluids that can be evaluated, water content can be measured or glucose in saliva, both of which require different lights sources and signals and, thus, possibly one or more photodiodes 80.

[0045] A photometric analog front end 95 (such as ADPD1080, Analog Devices, Massachusetts, United States) is used in one embodiment to control laser pulse timing and readout of photodiode signals In other embodiments, other similar front ends 95 can be used as will be known to one skilled in the art. The ADPD108095, and other similar devices 95, are highly efficient, photometric front ends 95, each with an integrated 14-bit analog-to-digital converter 105 (“ADC”) and a 20-bit burst accumulator 135 that works with flexible light emitting diode (“LED”) drivers 140. The ADPD1080 / ADPD1081 95 stimulate a light source 60 and measures the corresponding optical return signal 115. For one embodiment, the data output of the sensor device 1 and functional configuration occur over a 1.8 V I2C interface 120 on the ADPD108095 or a serial port interface 120 (“SPI”) on the ADPD108195. “Functional configuration” refers to software setup of the ADPD 108095 (i.e. number of pulses to the laser, current to power the lasers, etc.) Control circuitry, which comes as part of the ADPD 95, includes flexible LED signaling and synchronous detection. In this embodiment, the components were placed on a split printed circuit board 35 for transmission measurements and housed in an aluminum housing 2. As previously mentioned, a layout of the PCB 35 is shown in Figures 1A and IB, and the housing 2 design and sensor device 1 are shown in Figures 1C and ID, respectively. This layout can vary in other embodiments. Again, the specific components used are non-limiting and substitutions will be obvious to one skilled in the art. The type and / or selection of light source 60 (laser or LED), detector 85 for measuring the light intensity, and analog front end 95 are flexible choices depending on the application, the attribute being measured, the body fluid, and the final user’s need.

[0046] Figure 12 shows one embodiment of a sensor device 1 with a PCB 35 affixed to a first half 3 of the housing 2. Also depicted in Figure 12 are the header 10 with inserted wires 11 (to power the device 1) and the corner screws 50. Figure 13 illustrates alternative views of a photodiode board top 6A (left) and a laser board top 5A (right) with headers 10. Figure 14 illustrates alternative views of a photodiode board bottom 6B and a laser board bottom 5B (left and right respectively.) Figure 15 shows one embodiment each of a housing first half interior side 23 (with photodiodes 80) (left image) and a housing second half interior side 25 (with light sources 60)(right image.) Figure 16 illustrates one embodiment of a sensor device 1 clamped onto a blood-filed tube 55.

[0047] Figures 3 A and 3B show schematic or modular diagrams of embodiments of a sensor device 1 of the present invention illustrating alternative names / labeling for the various parts of analogous embodiments of a device 1. In this exemplary, but non-limiting, embodiment, the ADPD1080 front-end 95 is programmed via a computer 90 or microcontroller 90 (for example, Arduino Nano ESP32, Arduino, UK, which are specifically labelled “Arduino 500 ”) in the Arduino Integrated Development Environment using an open-source custom library developed for this specific integrated circuit. While embodiments of the present invention utilize an Arduino 500, there are a multitude of computers 90 or microcontrollers 90 that can be used to program and control the device 1. In other embodiments, other similar microcontrollers 90 can be used. This front end 95 is used to stimulate the lasers 65, 70, using 3-ps pulses and measure the return signal 115 from the photodiode 80 through the output of a transimpedance amplifier 100 (“TIA”) with a gain of 50-kQ. The characteristics of the pulses can vary if other light sources 60 are used - with corresponding changes to Figures 3 A and 3B. Also, a predetermined number of pulses 107 are used depending upon the design of the sensor 1, the attribute being measured, the fluid being analyzed, etc. The laser pulses 107 (maximum 127 pulse number 107) are summed using an integrator to increase the signal-to-noise ratio.The output voltage from the TIA 100 is sent through a built-in bandpass filter 145 (optimized for 2-3 ps pulses to remove the influence of ambient light) and is converted by the analog-to-digital converter 105 (“ADC”) to an ADC intensity value 106. The ADC intensity values 106 scale linearly with the pulse number 107 up to a maximum value of 65535. The currents for the 680 nm and 850 nm lasers 65, 70 are set to 9.97 mA and 18.41 mA, respectively. Again, these values can vary if different light sources 60 are used. For this embodiment, the currents are selected to be above 80% of their maximum current to prevent degradation of the lasers 65, 70 due to overdriving and obtain maximal signal through blood. To ensure consistent placement and stability of the sensor devices 1, adjustment screws 51 are incorporated in the housing 2 to apply about 2 mm compression to the tubing 55. As excess compression of the tubing 55 will result in improper blood flow which could harm the patient, different levels of compression can be chosen to determine a safe level for a clinical case. This type of assessment can be performed by examining changes in blood flow resistance (calculated as a pressure difference divided by the volumetric flow rate). It was found that compression up to a maximum of 4 mm results in less than 1 mmHg L'1min increase in blood flow resistance . Typical ECMO circuits will have blood resistances that are an order of magnitude above these changes. To be conservative in the compression of the tubing 55, for one embodiment ,half of the maximum level of compression (2 mm) was selected. 1. Additionally, to prevent light distortion and damage to the lasers 65, 70, nylon spacers 110 can be added to the laser PCB 35 to prevent contact between the surface of the lasers 65, 70 and the tubing 55.

[0048] For one embodiment, the sensor devices 1 are calibrated and validated using a recirculating flow circuit 300 (shown in Figure 17) with 600 mL of bovine blood (7200801, Lampire Biological Labs, United States) at different hemoglobin concentrations and saturations. Prior to loading the blood in the circuit 300, the hemoglobin concentration can be assessed via blood gas analysis and, if necessary, adjusted to the desired level. The adjustment is achievedby centrifuging the blood to remove plasma, thereby increasing hematocrit and / or saline dilution to decrease hematocrit. The circuit 300 contains a centrifugal blood pump 320 (PediMag, Abbott Cardiovascular), an oxygenator 310 (Terumo, Capiox FX 15 oxygenator 310) connected to a gas blender 305, a sampling port, and a 400-ml soft bag reservoir 330 for blood. The blood temperature for this embodiment is maintained at 37 ± 2°C using a water bath 315 with the built-in heat exchanger of the oxygenator 310. The blood flow rate is set to approximately 2 L / min, and the gas flow rate was 4 L / min. This environment mimics the human body for the purpose of testing the sensor device 1.

[0049] For the testing of this embodiment, the gas mixture contained oxygen, carbon dioxide, and nitrogen, and the concentrations of the gases are adjusted to achieve different oxygen saturation levels for a given hemoglobin concentration. As shown in Figure 5, for each gas change, blood is allowed to circulate for 20-minutes for a measurable change in saturation and stabilization. A sample is then drawn from the circuit 300 to obtain ground truth blood oxygen saturation and hemoglobin concentration using a blood gas analyzer (BGA, ABL 825 Flex, Radiometer, Denmark). “Ground truth blood gas measurements” refers to direct, laboratory -verified blood gas values that are considered the true measurements of a patient’s blood gas status. These are used as the reference standard when validating sensors, monitors, point-of-care devices, or algorithms that estimate blood gases and are specific to whichever blood gas is being measured.

[0050] For this testing, the optical sensor devices 1 are clipped on the circuit’s tubing 55 (3 / 8" ID x 9 / 16" OD x 3 / 32" Wall, ND 100-65 Tygon) and light transmission (intensity) data is collected for each of the oxygenation states. During each blood sampling point, one minute of intensity data with a 20 Hz sampling rate is collected. Once the blood is fully saturated, high flow nitrogen is used to reduce the oxygen saturation, and the saline is added tocircuit 300 for dilution to a lower hemoglobin concentration. This step enables additional saturation measurements at multiple hemoglobin concentrations.

[0051] Figures 2A and 2B illustrate an alternative, second embodiment of a sensor device 1 according to the present invention having three light sources 60 (a first laser 65, a second laser 70, and a third laser 75.) When including additional light sources 60 in various embodiments, the number of photodiodes 80, detectors, 85, and the PCB 35 (electronics) are usually the only components of the device 1 that change (when using three or more light sources 60.) The embodiment in Figures 2A and 2B of a sensor device 1 adds an additional laser 75 (at 808 nm) and an additional photodiode 80 for improving hemoglobin estimates. As shown in Figures 2A and 2B, the optical sensor device 1 utilizes an additional laser 75 at 808 nm (VD-08081-015M-XX-2 A0, Bright Laser) and photodiode 80 to measure the intensity of the third laser 75. This 808 nm laser 75 enables improved estimation of hemoglobin concentration due to reduced influence of oxygen saturation on the optical properties of hemoglobin near this wavelength. Results of validation data using this new 3-wavelength configuration are shown in Figure 11 A-l ID. An R2>0.9 was obtained for both saturation and hemoglobin concentration predictions from both sensor devices 1, except for sensor 3 with its hemoglobin estimates (R2=0.89). The limits of agreements (“LOA”) for Hgb were within ~1 g / dL across the two sensors, while for SO2 the LOA were within 10%. Additionally, the accuracy root-mean square error (“ARMS”) for Hgb was found to be 0.53 g / dL and 0.36 g / dL for sensor 3 and sensor 4 respectively. For SO2, the ARMS was 3.04% and 3.09% for sensor 3 and sensor 4 respectively. These results are evidence of demonstrated improvement between a sensor device 1 having two lasers 65, 70 and a sensor device 1 having three lasers 65, 70, 75. As context for and an explanation of Figures 8A through 8H and 11A through 11D, the labels “Sensor 1,” “Sensor 2,” “Sensor 3,” and "Sensor 4” refer to individual sensor devices 1 whose results are plotted on the respective graph. For example, Figures 8 A through 8H show the results of two sensordevices 1 (named Sensor 1 and Sensor 2) that were developed using only two lasers 65, 70.Figures 11A through 11D show data from the second embodiment (the described in Figures 2A and 2B), where the two sensor devices (Sensor 3 and Sensor 4) are made using the three lasers 65, 70, 75.

[0052] Figures 11 A through 1 ID illustrate the predicted hemoglobin concentration and oxygen saturation from the two sensor devices 1 of Figures 2A and 2B, which are plotted against blood gas measurements (upper plots) and Bland-Altman plots are generated to assess their accuracy (bottom plots).

[0053] An empirical calibration method 200 has been developed using the collected in vitro intensity data to extract and report predicted hemoglobin concentration and oxygen saturation. Embodiments of this calibration method 200 are described below and also are depicted in the flowcharts shown in Figures 9 and 10. As illustrated in Figure 10, the method 200 can be modified depending upon whether it is used with a single sensor device 1 or multiple sensor devices 1 by incorporating step 220 for the calibration of multiple sensor devices 1. The set pulse number 107 (the predetermined pulse number 107) first normalizes the raw intensity data from the two sensor devices 1. This parameter (the set pulse number 107) is adjusted to prevent low signal (at high hemoglobin concentrations) and saturation (at low hemoglobin concentrations) at the detector 85. The average intensity of each sensor devices 1 at each wavelength is then computed, and the ratio of average intensities is used as inputs to the empirical calibration. This approach is conducted to account for manufacturing differences in certain embodiments of the sensor devices 1, enabling a single calibration for each sensor device 1 with appropriate scaling factors applied (Figures 4A through 4D show how to use asingle set calibration values for multiple sensor devices.). Appropriate scaling factors are those that correctly map raw intensity values to the reference curve.

[0054] An empirical calibration method 200 for extraction of hemoglobin concentration and oxygen saturation begins by using the collected in vitro intensity data (as illustrated at step 210 of Figures 9 and 10.) If multiple sensor devices 1 are being calibrated, then the raw intensity data of individual light sources 60 from multiple sensor devices 1 is first normalized by the predetermined, preset pulse number 107. The pulse number 107 is adjusted to prevent low signal (at high hemoglobin concentrations) and saturation (at low hemoglobin concentrations) at the detector 85. The average normalized intensity of the light sources 60 at each wavelength and the light sources’ 60 ratio of intensities are then calibrated to be used as inputs to the empirical calibration (step 220 of Figure 10.) This approach accounts for manufacturing differences in the sensor devices 1 and enables the use of a single calibration for both (or multiple) sensor devices 1 with appropriate scaling factors applied at step 220. For both the calibration of one sensor device 1 and multiple sensor devices 1, intensity of each light source 60 is divided by a predetermined pulse number 107 to compute a logarithm of scaled intensities and the intensities’ ratio of logarithms 222 (“scaled” refers to the intensity data of a light source 60 divided by the pulse number 107) (step 222 of Figures 9 and 10). Hemoglobin concentration is determined based on a sparse regression-based model, with input parameters being the logarithm of the intensities of the individual light sources 60 and the ratio of the logarithms of the intensities (steps 220 and 222 in Figures 9 and 10) (equation 1 below, represented in Figure 10 by step 230, is a selected regression model to determine hemoglobin concentration),where I680and I850are the scaled intensities through blood, R=ln(I680) / ln(I850) is the ratio, and Ui-9 are the fitted coefficients from the regression model based on the in vitro data collection. Again, the ratio is computed by dividing the logarithm of one scaled intensity from one light source 60 by the logarithm of the scaled intensity of another light source 60. In other embodiments, other regression-based models may be used. The model for this one embodiment has been derived by choosing (selecting) a model from a function library consisting of polynomials of the input data, which is the intensity data (see Figure 10) and using elastic net regression with a 5-fold cross-validation to determine the optimal coefficients. The elastic net regression is selected as opposed to ordinary least squares as it imposes sparsity (selecting only the most important predictors) and performs well when there are correlations between predictors.

[0056] In one embodiment, using the estimates of Hgb from eq. 1, the hemoglobin oxygen saturation was calculated based on equations 2-4 (below) (which occurs at step 245 in Figure 10):SO2= a(Hgb) * R + [3(Hgb (2)a[Hgb] = ma* Hgb + ba(3)0[Hgb] = mp* Hgb + bp (4)where ma, mp, ba, and bp are coefficients to be fitted for. Equation 2 reflects that the ratio R of the logarithms of intensities shows a strong linear correlation with oxygen saturation. However, the slope (a) as well as intercept (P) are influenced by hemoglobin concentration (equations 3-4) and are novel aspects of a calibration method 200.

[0057] In this embodiment of a calibration method 200, hemoglobin concentrations are examined, ranging from 6-14±0.5 g / dL, and oxygen saturations range from 50-100% to extract the coefficients (Table 1) for calibration and model validation (Figures 7A-8H). The model is assessed by computing the R2and adjusted R2and generated Bland-Altman plots to assess theagreement between the optical sensors 65, 70, 75 and the blood gas analyzer measurements. The Bland-Altman Method is a statistical technique that compares two methods of measurement to determine their agreement. This Bland- Altman Method is often used in health research to compare raters or methods for quantitative outcomes. Additionally, the accuracy root-mean-square error (“ARMS”) is computed as a metric of device performance (Table 2). In this one embodiment, a total of 86 sensor device measurements and blood gas analyzer measurements were collected from seven days of experiments and split into two groups: calibration and validation. The calibration group is used to fit for the coefficients in the model and the validation group is used to test the model on unseen data.

[0058] The results of this embodiment, shown in Figures 4A through 4D, depict the logarithms of the intensities at the two wavelengths prior to and after the application of the extracted scaling factors for each sensor device 1 are plotted against the reference line.

[0059] For this one embodiment shown in Figures 1 A through IF, applying the scaling factors for sensor 1 and sensor 2 at 680 nm (0.74 and 1.54, respectively) cleanly maps the raw data to the reference. In other embodiments of the invention, including the use of other light sources 60, the scaling factors can vary. For 850 nm, there is a minimal change in the data as the intensities for the two sensors at this wavelength are quite similar. This similarity is reflected in the associated scaling factors (0.98 and 1.03, respectively), which are close to 1. The reference line data for this embodiment is used to generate the coefficients for the hemoglobin calibration in Table 1.Table 1: Hemoglobin calibration coefficients

[0060] For the embodiment illustrated in Figures 2A and 2B, a modified calibration method 200 is employed. Only the intensity of the 808 nm laser 75 was used to determinehemoglobin concentration, while 680 nm 60 and 850 nm 70 are used for oxygen saturation. A 3-wavelength sensor device 1 is calibrated using bovine whole blood on a recirculating loop 300 at different hemoglobin levels (5-11 g / dL) and oxygen saturations (50-100%) according to the one or more methods 200 described previously herein. Using collected intensity data, the hemoglobin concentration (Hgb) is determined via regression using the logarithm of the 808 nm laser 75 intensity (I808) data as input and fitting the coefficients ( is) in eq.5230. Using the estimated Hgb and the ratio of the logarithms of 680 nm 65 and 850 nm 70 laser intensities (R = In I680 / ln I850), the oxygen saturation (SO2) is computed according to equations 6-9245, where my, ma, mp, by, ba, and bp are coefficients to be fitted for. The calibration coefficients that produced the hemoglobin concentration values shown in Figure 11A and 1 IB are as follows: ais = 1.68, -26.7, and 111 for the sensor 3 and ais = 1.6, -26, and 111 for sensor 4. The SO2 coefficients are my=-1.04, ma=1.92, mp=-0.876, by=4.31, ba=-6.9, and bp=3.49 for sensor 3 and are my=-0.558, ma=1.04, mp=-0.465, by=1.64, ba=-l .83, and bp=l for sensor 4. As mentioned previously, the oxygen saturation is computed according to the following equations:SO2= y ■ R2+ a ■ R + p (6)Y = my■ Hgb + by(7)a = ma■ Hgb + ba(8)= mp■ Hgb + bp(9)

[0061] The same scaled intensity data as well as calculated Hgb is used with equations 2-4 to extract the unknown slope (a) and intercept (P). The intensity ratio is plotted against oxygen saturation at different binned Hgb values (±0.5 g / dL) and linear fits are applied to the data (Figure 5 and step 225 of Figure 10). A strong linear correlation exists between intensityratio and oxygen saturation at each concentration of hemoglobin (R2>0.9) (Figure 6). However, the slope and intercept of this relationship change linearly (R2=0.92) with increasing hemoglobin concentrations as seen in Figures 7A and 7B. This result is consistent with what is known to those skilled in the art. The obtained linear equations for a and P (Figs. 8A through 8H) are used in combination with the estimated Hgb values to calculate oxygen saturation. The comparison of calculated versus ground truth hemoglobin concentration and oxygen saturation values are shown in Figures 8A through 8D (upper row). Results are shown for data used for calibration (8A and 8C) as well as validation (8B and 8D). Figures 8E through 8H (bottom row) show the Bland-Altman plot for the same data sets. Figures 8A through 8D are coded to show two different sensors. The R2, adjusted R2, and ARMS for each sensor device 1 and group per parameter (Hgb and SO2) are computed and reported in Table 2.

[0062] For this one embodiment, plotting the extracted slopes and intercepts (which result from step 227 of Figure 10) from the fits in Figures 7A and 7B against the average of each hemoglobin concentration reveals a linear trend can be observed (R2=0.90 and R2=0.89, respectively). The obtained equations and the estimated hemoglobin concentration from the hemoglobin calibration model are used to determine oxygen saturation (step 240 or 250 of Figures 9 and 10.)

[0063] In this embodiment, the results of applying the novel calibration method 200 to the two optical sensor devices 1 for determining hemoglobin concentration and oxygen saturation are shown in Figures 8A through 8H, respectively. Figures 8A-8D show the predicted hemoglobin concentration 290 from each sensor against the ground-truth blood gas measurement for both calibration and validation groups. Figures 8E through 8H demonstrate the deviation of sensor measurements from the blood gas measurements through the Bland-Altman method. For this embodiment, Figures 11A and 11B and Figures 11C and 11D similarly show the results for the predicted oxygen saturation 290 for each sensor along withassociated Bland-Altman plots. The R2, adjusted R2, and ARMS for each sensor and group per parameter (Hb and SO2) for this embodiment are reported in Table 2.Table 2: Summary statistics of calibration method 200, in one embodiment, for the two sensors. Adjusted R2is not computed for oxygen saturation as simple linear fits 225 are used for this parameter making R2alone a sufficient descriptor for model performance.

[0064] One embodiment of a calibration method 200 that can be used in developing one or more sensors comprises the following steps and is illustrated broadly in Figure 9 and in more detail in Figure 10. Collect intensity data of a first light source and of a second light source of an optical sensor device and collect at least one ground truth blood gas measurement of in vitro blood (step 210). Divide the intensity data of each light source by a predetermined pulse number to compute scaled intensities and compute logarithms of the scaled intensities of the first light source and the second light source and compute a ratio of the logarithms of the first light source and the second light source (step 222). As a non-limiting example, a ratio can be defined as R=ln(I680) / ln(I850), where ln(I680) and ln(I850) are the logarithms of the intensities, and 1680 and 1850 are the intensities (which have already been divided by the pulse number). Input the logarithms, the ratio, and the blood gas measurements into a preselected model for hemoglobin concentration (step 230). Fit the preselected model to the data to extractcoefficients for the model and using the coefficients for a hemoglobin concentration prediction (step 240). Apply a linear fit of the ratio of the logarithms versus oxygen saturation from the blood gas measurements for every hemoglobin concentration value from the preselected model (step 225). Extract slopes and intercepts of the linear fit (step 227). Appy an additional linear fit to the extracted slopes and intercepts as a function of hemoglobin concentration values (step 227) (there are multiple blood gas measurements for a single hemoglobin concentration and each sample is at different oxygen saturation.) Use the hemoglobin concentration prediction to estimate a slope and an intercept for an oxygen saturation model (step 245). Use the estimated slope, the intercept, and the measured ratio to predict oxygen saturation (step 250).

[0065] When calibrating more than one sensor device 1 the above-outlined method 200 is employed with additional step 220 (see Figure 10.) Step 220 is the primary difference between a method of calibrating one sensor device 1 versus more than one sensor device 1. In another words, step 220 is performed when calibrating multiple sensor devices 1 simultaneously. For Figure 10, all the steps after 220 apply to the calibration 200 of both one sensor and multiple sensor devices 1.

[0066] The present invention further comprises a method 200 for developing and calibrating optical sensors for real-time monitoring of oxygen saturation and hemoglobin concentration in whole blood. Results show an average ARMS (accuracy) of 1.30 g / dL for concentration and 4.76 % for saturation across the two devices. The consistency of the R2 and adjusted R2 across the calibration and validation groups is indicative of model stability and a lack of overfitting. The calibration and validation group results for hemoglobin concentration extraction are similar for sensor 1 as indicated by the R2 (0.79 and 0.78 respectively), adjusted R2 (0.73 and 0.72 respectively) and ARMS (1.28 g / dL and 1.30 g / dL respectively) of the calibration and validation groups. The results for sensor 2 worsen from calibration (R2=0.84, adjusted R2=0.80, ARMS=1.12 g / dL) to the validation group (R2=0.79, adjusted R2=0.73,ARMS=1.29 g / dL). However, this change is not statistically significant. The similarity between sensors is further demonstrated by Bland- Altman plots when comparing the mean biases and limits of agreement. The above values can vary in other embodiments where different components, such as light sources 60, are used.

[0067] In this embodiment, both sensor devices 1 show similar results across the calibration and validation groups for determining oxygen saturation. However, the sensor device 1 illustrated in Figures 2A and 2B demonstrates better overall calibration (R2=0.93, ARMS=3.70 %) and validation performance (R2=0.92, ARMS=3.78 %) for extraction of oxygen saturation compared to the sensor device 1 illustrated in Figures 1A through IF according to their calibration (R2=0.85, ARMS=5.45 %) and validation results (R2=0.82, ARMS=5.78 %). Additionally, the limits of agreement are narrower for sensor 2 compared to sensor 1. However, these differences between sensors 1 and 2 are not statistically significant, demonstrating reproducibility across devices.

[0068] While research on the invention has seen good agreement between sensors 1 and 2, each sensor accuracy is dependent on calibration accuracy. Given that SO2 and Hgb are coupled, a certain amount of cross talk is expected. Additionally, the Bland-Altman plots, particularly for Hgb, demonstrate deviations from blood gas values that are non-linear - making the error in sensor measurements dependent on the true Hgb concentration. One way to mitigate these effects is through the inclusion of an additional wavelength at an isosbestic point (for instance near 800 nm). Use of additional wavelengths can reduce overall error in the sensor and remove the observed non-linearity. Furthermore, since intensity changes at the isosbestic wavelength are only affected by Hgb concentration, cross talk between SO2 and Hgb can be minimized. According to various embodiments of the present invention, sensors can be reliably used for monitoring patient hemodynamics and aid in clinical decisions of blood transfusion. Sensor devices 1 of the present invention also can be used to assess ECMO oxygenator 310functionality in real-time through pre- and post-oxygenator 310 blood saturation and oxygen delivery rate. While clinically available hemoglobin concentration and oxygen saturation monitors comparable to the sensors of the present invention do exist (e.g. Terumo CDI 500 and Spectrum M4 monitor), the devices of the present invention are of a small form factor reducing the bulk that commercial monitors contribute to extracorporeal systems such as ECMO. Another advantage the sensor devices 1 of the present invention have over current devices is that the sensor devices 1 of the present invention read through standard Tygon tubing 55 forgoing the need for a specialized connector. This aspect is valuable since these custom connectors often disrupt flow in extracorporeal systems and are procoagulant. By developing a clamp-on device instead of an inline one, embodiments of a sensor device 1 avoid these risks altogether.

[0069] The performance of embodiments of these sensor devices 1 is comparable to clinically available hemoglobin concentration and oxygen saturation monitors, such as the Terumo CDI 500. The consistency of the R2and adjusted R2(for hemoglobin concentration) across the calibration and validation groups indicates model stability and a lack of overfitting. This stability, coupled with the performance of the sensors, enable their use of real-time measurements in extracorporeal system settings such as ECMO or cardiopulmonary bypass.

[0070] The benefits of sensor devices 1 in comparison to current commercial systems are the sensors 1 have a small form factor— reducing system bulk in extracorporeal setting— and that the sensors 1 are portable allowing for effortless translation into the clinic. Another advantage to the sensor devices 1 over existing technologies is that the sensor devices 1 read through standard Tygon tubing 55 forgoing the need for a specialized connector. This aspect is valuable since these custom connectors often disrupt flow in extracorporeal systems and are procoagulant. By developing a clamp-on device instead of an inline one, the sensor devices 1 avoid these risks altogether. Additionally, setup of the sensor devices 1 can be performed withina few minutes and have been developed to be compatible with medical grade tubing 55. The sensor devices 1 possess the potential to allow for translation of other miniaturized systems (e.g. PAS), while remaining compatible with existing medical systems.

[0071] It is to be understood that the invention may assume alternative variations and step sequences, unless specified to the contrary. It also is to be understood that the specific devices and processes illustrated in the attached drawings and described in this specification are simply exemplary embodiments of the invention. Hence, specific dimensions and other physical characteristics related to the embodiments disclosed are not to be limiting.

Claims

CLAIMSWhat is claimed is:

1. An optical sensor device for measuring characteristics of bodily fluids in an optically transparent tube, the sensor device comprising:a housing having a first half and a second half, which operate together to clamp onto tubing;two printed circuit boards with one printed circuit board connected to the first half of the housing and the second printed circuit board connected to the second half of the housing;two or more light sources connected to the printed circuit board on the second half of the housing;a microcontroller unit connected to the printed circuit board on the first half of the housing;a photometric analog front end register connected to the printed circuit board on the first half of the housing; andat least one photodiode connected to the first half of the housing, positioned to receive light transmitted from the at least two or more light sources and through the tubing, and operable to generate an optical return signal;wherein the microcontroller unit is configured to program the photometric analog front end register to (i) stimulate the two or more light sources with a predetermined number of pulses, (ii) obtain at least one measurement from the return signal, and (iii) transmit the at least one measurement to the microcontroller unit for use in computing at least one characteristic of the bodily fluid in the tubing.

2. The optical sensor device of Claim 1, whereinthe bodily fluid is blood;the two or more light sources are in the near-infrared range; andthe transmit the at least one measurement to the microcontroller unit for use in computing at least one characteristic of the bodily fluid in the tubing through a calibration of the at least one measurement.

3. The optical sensor device of Claim 2, wherein the characteristics being computed are hemoglobin concentration and oxygen saturation.

4. The optical sensor device of Claim 1, also comprising:a plurality of corner screws configured to secure the first half of the housing to the second half of the housing and operable to apply compression to the tubing.

5. The optical sensor device of Claim 4, also comprising a plurality of adjustment screws operable to lock the first half of the housing to the second half of the housing to maintain the compression of the tubing.

6. The optical sensor device of Claim 1, wherein the two or more light sources comprise a first laser, a second laser, and a third laser.

7. The optical sensor device of Claim 6, wherein the first laser operates at approximately 680 nm, the second laser operates at approximately 850 nm, and the third laser operates at approximately 808 nm.

8. The optical sensor device of Claim 1, wherein the housing is about 5-cm by 5-cm.

9. The optical sensor device of Claim 1, wherein the device weighs less than 50-grams.

10. The optical sensor device of Claim 1, wherein the housing is made of aluminum.

11. The optical sensor device of Claim 1, wherein each of the first half of the housing and the second half of the housing have a vertical cavity surface configured to accommodate the tubing.

12. The optical sensor device of Claim 11, wherein the two or more light sources are located in the vertical cavity surface of the second half of the housing and the photodiodes are located in the vertical cavity surface of the first half of the housing.

13. The optical sensor device of Claim 3, which has been calibrated by a method comprising:collecting intensity data of a first light source and of a second light source of the optical sensor device and collecting at least one ground truth blood gas measurement of in vitro blood;dividing the intensity data of each light source by a predetermined pulse number to a compute scaled intensity of each light source and computing logarithms of the scaled intensities of the first light source and the second light source and a ratio of the logarithms of the scaled intensities of the first light source and the second light source;inputting the logarithms, the ratio, and the blood gas measurements into a preselected model for hemoglobin concentration;fitting the preselected model to the data to extract coefficients for the model and using the coefficients for a hemoglobin concentration prediction;applying a linear fit of the ratio of the logarithms of the scaled intensities of the first light source and the second light source versus oxygen saturation from the blood gas measurements for every hemoglobin concentration value from the preselected model;extracting slopes and intercepts of the linear fit;applying an additional linear fit to the extracted slopes and intercepts as a function of hemoglobin;using the hemoglobin concentration prediction to estimate a slope and an intercept for an oxygen saturation model; andusing the estimated slope and the intercept of the oxygen saturation model and the measured ratio to predict oxygen saturation.

14. A calibration method for an optical sensor device that measures hemoglobin concentration and oxygen saturation in blood using at least two light sources, comprising: collecting intensity data of a first light source and of a second light source of an optical sensor device and collecting at least one ground truth blood gas measurement of in vitro blood;dividing the intensity data of each light source by a predetermined pulse number to a compute scaled intensity of each light source and computing logarithms of the scaled intensities of the first light source and the second light source and a ratio of the logarithms of the scaled intensities of the first light source and the second light source;inputting the logarithms, the ratio, and the blood gas measurements into a preselected model for hemoglobin concentration;fitting the preselected model to the data to extract coefficients for the model and using the coefficients for a hemoglobin concentration prediction;applying a linear fit of the ratio of the logarithms of the scaled intensities of the first light source and the second light source versus oxygen saturation from the blood gas measurements for every hemoglobin concentration value from the preselected model;extracting slopes and intercepts of the linear fit;applying an additional linear fit to the extracted slopes and intercepts as a function of hemoglobin;using the hemoglobin concentration prediction to estimate a slope and an intercept for an oxygen saturation model; andusing the estimated slope and the intercept of the oxygen saturation model and the measured ratio to predict oxygen saturation.

15. The method of Claim 14, wherein the preselected model for hemoglobin concentration iswherein I680and I850are the scaled intensities, R=ln(I680) / ln(I850) is the ratio, and a±-9are the fitted coefficients from a regression model based on the in-vitro blood.

16. A calibration method for a plurality of optical sensor devices that measure hemoglobin concentration and oxygen saturation in blood using at least two light source, comprising the steps of:collecting intensity data of at least two light sources and ground truth blood gas measurements of in vitro blood from an optical sensor device, wherein the ground truth blood gas measurements include at least one hemoglobin concentration value;averaging the intensity data at a predetermined wavelength for all sensor devices to determine an average intensity per light source and calculating scaling factors for each sensor device to the average;dividing intensity data of each light source by a predetermined pulse number to compute logarithms of scaled intensities and a ratio of the logarithms as processed data;applying a linear fit of the ratio of the logarithms for each hemoglobin concentration value;extracting slopes and intercepts of the linear fits;applying additional linear fits to the extracted slopes and intercepts as a function of hemoglobin;using the hemoglobin concentration prediction to estimate a slope and an intercept for an oxygen saturation model;using the estimated slope, the intercept, and the measured ratio to predict oxygen saturation; andreporting the predicted hemoglobin concentration and oxygen saturation.

17. The method of Claim 16, wherein the preselected model for hemoglobin concentration iswherein I680and I850are the scaled intensities, R=ln(I680) / ln(I850) is the ratio, and a^_9are the fitted coefficients from a regression model based on the in-vitro blood.

18. The methods of Claims 15 and 17, wherein the hemoglobin oxygen saturation is calculated based on the following equations:[ SO ] _2=a(Hgb)*R+P(Hgb);a[Hgb]=m_a*Hgb+b_a; andP [Hgb]=m_P *Hgb+b_P,wherein ma, mp, ba, and bp are coefficients to be fitted for the coefficients to calculate the predicted oxygen saturation.

19. The method of Claim 14, also comprising reporting the predicted hemoglobin concentration and oxygen saturation.

Images