Non-invasive imaging system for stabilizing microvessels in oral mucosa

Abstract

A system for imaging the microvasculature of a subject's tissue is disclosed. The system comprises (a) a tissue stabilizer configured to contact the tissue of the subject and maintain the position of the region of the microvasculature to be imaged, and (b) an imaging device. The imaging device comprises (i) a housing having an imaging section, (ii) an illumination device disposed within the imaging section and having a light output end for irradiating the region of the microvasculature with light, wherein the light output end is offset with respect to the optical axis of the imaging section, (iii) an objective lens disposed within the imaging section for receiving at least a portion of the light scattered by the region of the microvasculature, and (iv) an image detector disposed within the imaging section for receiving the light redirected by the objective lens and detecting a microscopic image of the region of the microvasculature.

Description

Technical Field

[0001] Cross - Reference to Related Applications This application claims the benefit of, and priority to, U.S. Application No. 63 / 344,975, filed May 23, 2022, which is hereby incorporated by reference in its entirety for all purposes.

[0002] Statement Regarding Federally Sponsored Research This invention was made with government support under Grant No. FA9550 - 20 - 1 - 0063 awarded by the Air Force Office of Scientific Research. The government has certain rights in this invention.

[0003] The present invention relates to a system and method for imaging microvessels of a subject's tissue, and more particularly, to a system and method for non - invasive imaging of stabilized microvessels of the oral mucosa.

Background Art

[0004] Sepsis is a life - threatening medical emergency, affecting over 30 million people worldwide and causing the loss of more than 8 million lives, including over 3 million children. Sepsis is the leading cause of death among hospitalized patients in the United States, accounting for approximately 50% of deaths in intensive care units. Early diagnosis of sepsis is extremely important, and it is estimated that the mortality rate increases by 7 - 10% for every hour of delay in diagnosis (see Farkas, J Thorac Dis 2020;12(Suppl 1):S16 - S21). White blood cell count is one of the most commonly used diagnostic parameters to guide intervention in sepsis patients. However, current standard blood cell measurements are invasive and require repeated blood sampling from a vulnerable patient population at high risk of medical complications (such as secondary infections, anemia, chronic pain, etc.). Furthermore, laboratory analysis takes time, and clinical blood counts are not always available in resource - poor settings.

[0005] In animal models using intravital microscopy, leukocyte rolling and adhesion phenomena (collectively referred to as leukocyte-endothelial interaction (LEI)) are well characterized, but are rarely observed in humans. Conceptually, since conventional histopathology relies on static examination of biopsy samples, imaging cell motility as a potential source of diagnostic information has not yet been explored clinically. LEI has been reported to be significantly increased in the sublingual microvessels of patients with systemic inflammation such as sepsis and ischemia-reperfusion injury. However, due to the lack of appropriate detection and analysis tools, it has been difficult to evaluate LEI in the clinical setting.

[0006] Therefore, there is a need for an improved system and method for imaging the microvessels of a subject so that the observed leukocyte-endothelial cell interactions can provide a source of diagnostic information.

SUMMARY OF THE INVENTION

[0007] To address these limitations, we have developed a system for imaging the microvessels of tissue. This system includes a tissue stabilizer and an imaging device (such as a miniaturized microscope), and by phase gradient microscopy with oblique back-illumination, in vivo blood cells can be non-invasively detected and quantified in real time without labeling. With this system, we can capture videos of blood cells moving at high speed and slow-moving leukocytes rotating and adhering to the vessel walls of the buccal microvessels of healthy human volunteers. In particular, leukocyte rolling and adhesion are new diagnostic parameters based on cell dynamics (movement), rather than conventional static parameters such as cell morphology, and can only be obtained by in vivo microscopy. Custom algorithms for the automatic quantification, movement analysis, and potential classification of various types of leukocytes have been published to provide highly reliable and practical results for clinicians.

[0008] In one aspect, the present disclosure provides a system for imaging the microvasculature of a subject's tissue. The system includes (a) a tissue stabilizer configured to contact the tissue of the subject and maintain the position of the region of the microvasculature to be imaged, and (b) an imaging device. The imaging device includes (i) a housing having an imaging section, (ii) an illumination device disposed within the imaging section of the housing and having a light output end that irradiates the region of the microvasculature with light, wherein the light output end is offset with respect to the optical axis of the imaging section, (iii) an objective lens disposed within the imaging section of the housing and configured to receive at least a portion of the light scattered by the region of the microvasculature, and (iv) an image detector disposed within the imaging section of the housing and configured to receive the light redirected by the objective lens and detect a microscopic image of the region of the microvasculature.

[0009] In one embodiment, the tissue stabilizer includes a base, a slide mechanism attached to the base, and an adapter for contacting the tissue. The adapter is attached to the slide mechanism and is movable toward and away from the base. The base may include a jaw holder, and the tissue stabilizer may further include a frame to which the jaw holder and a forehead holder are attached. The adapter may include a patterned surface finish for contacting the tissue. The adapter may include opposing tabs for contacting the tissue. The adapter may apply mechanical pressure to the edge of the oral mucosa tissue at least 2 millimeters (e.g., 5 millimeters) away from the imaging target area, minimizing the effect of the mechanical pressure on the imaging area. In one embodiment, the adapter is bendable. In one embodiment, the adapter is rigid. In one embodiment, the adapter includes a single or multiple light sources. In one embodiment, the adapter includes a single or multiple optical elements. In one embodiment, the adapter includes a vacuum system. In one embodiment, the adapter further includes a transparent sheet attached between the opposing tabs.

[0010] In one embodiment, the tissue stabilizer includes a base, a slide mechanism attached to the base, and an adapter for contacting the tissue, the adapter being attached to the slide mechanism and being movable laterally relative to the base. In one embodiment, the adapter is sized to contact the oral mucosa of the subject. In one embodiment, the adapter is sized to contact the lip of the subject. In one embodiment, the adapter includes a rod attached between opposing connectors, the rod being sized to contact the tissue. In one embodiment, the adapter includes a flexible loop, the rod being sized to contact the tissue.

[0011] In one embodiment, the objective lens is a microlens. In one embodiment, the objective lens is a gradient index (GRIN) objective lens. In one embodiment, the objective lens is a gradient index (GRIN) objective lens, and the system further includes a doublet achromat lens. In one embodiment, the image detector is a camera. In one embodiment, the image detector is a CMOS sensor. In one embodiment, the image detector is movable relative to the objective lens. In one embodiment, the image detector detects a microscopic image using Oblique Back-illumination Microscopy (OBM). In one embodiment, the image detector detects a microscopic image using Offset Trans-illumination Microscopy (OTM). In one embodiment, the imaging unit further includes a vacuum device for stabilizing the tissue to be imaged. In one embodiment, the imaging unit further includes a perfusion channel for supplying fluid to keep the tissue to be imaged in a wet state. In one embodiment, the lighting device includes a light source and an optical fiber having the light output end. In one embodiment, the imaging unit further includes an imaging chip including an objective lens, a vacuum device, a perfusion channel, and an illumination fiber of the lighting device, and the objective lens is a microlens. In one embodiment, the imaging unit further includes an imaging chip including an objective lens, a vacuum device, a perfusion channel, and an illumination fiber of the lighting device, and the objective lens is a gradient index (GRIN) objective lens. The imaging chip is disposable.

[0012] In one embodiment, the microscopic image includes an image of the interaction between white blood cells and endothelial cells in the microvessels. In one embodiment, the imaging is label-free imaging. In one embodiment, the microscopic image is a phase gradient contrast image. In one embodiment, the illumination device includes a light source and an optical fiber having the light output end, and the light source includes a light emitting diode. In one embodiment, the imaging is performed at a frame rate of 1 Hz to 1000 Hz. In one embodiment, the imaging is performed at a frame rate of 1 Hz to 300 Hz.

[0013] In one embodiment of the system, the optical power injected is automatically adjusted by a controller to prevent saturation of the pixels of the data acquisition element (e.g., CMOS). In one embodiment of the system, the scattered light collection time (exposure time) of the data acquisition element is automatically adjusted by software to prevent saturation of the pixels. In one embodiment of the system, the microscopic image includes an image of white blood cells in the microvessels, and the system further includes a controller that communicates electrically with the illumination device and the image detector, and the controller executes a program stored in the controller to (i) receive a microscopic image from the image detector, and (ii) calculate the average rotational speed of the white blood cells in the microvessels using automated frame-by-frame white blood cell tracking. In one embodiment, the controller executes a program stored in the controller to (iii) compare the average rotational speed of the white blood cells in the microvessels with the average rotational speed of white blood cells in healthy tissue.

[0014] In yet another aspect, the present disclosure provides a system for imaging the microvasculature of a subject's tissue. The system includes an imaging device including a housing having an imaging unit, a tissue stabilizer configured to contact the tissue of the subject and maintain the position of the region of the microvasculature imaged by the imaging device, an illumination device disposed within the tissue stabilizer and having a light output end for irradiating the region of the microvasculature with light, an objective lens disposed within the imaging unit of the housing and configured to receive at least a portion of the light scattered by the region of the microvasculature, and an image detector disposed within the imaging unit of the housing and configured to receive the light redirected by the objective lens and detect a microscopic image of the region of the microvasculature.

[0015] In one embodiment of the system, the tissue stabilizer includes a base, a slide mechanism attached to the base, and an adapter for contacting the tissue, the adapter being attached to the slide mechanism and being movable toward and away from the base. In one embodiment, the light output end of the illumination device is disposed within the adapter. In one embodiment, the base includes a jaw holder and the light output end of the illumination device is disposed within the jaw holder.

[0016] In another aspect, the present disclosure provides a system for imaging the microvasculature of a subject's tissue. The system includes a tissue stabilizer configured to contact the tissue of the subject and maintain the position of the region of the microvasculature to be imaged, an illumination device disposed within the tissue stabilizer and having a light output end for irradiating the region of the microvasculature with light, an objective lens disposed within the tissue stabilizer and configured to receive at least a portion of the light scattered by the region of the microvasculature, and an image detector disposed within the tissue stabilizer and configured to receive the light redirected by the objective lens and detect a microscopic image of the region of the microvasculature. In one embodiment of the system, the tissue stabilizer includes a first arm and an opposing second arm, the first arm and the second arm define a space therebetween for receiving the tissue, and the illumination device, the objective lens, and the image detector are disposed on the first arm such that the image detector detects a microscopic image using oblique back-illumination microscopy (OBM). In one embodiment of the system, the tissue stabilizer includes a first arm and an opposing second arm, the first arm and the second arm define a space therebetween for receiving the tissue, the objective lens and the image detector are disposed on the first arm, the illumination device is disposed on the second arm, and the image detector detects a microscopic image using offset trans-illumination microscopy (OTM). In one embodiment of the system, the tissue stabilizer includes a first arm, an opposing second arm, and a hinge connecting the first arm and the second arm, and a variable-size space for receiving the tissue is formed between the first arm and the second arm.

[0017] In yet another aspect, the present disclosure provides a system for imaging the microvasculature of a subject's tissue. The system includes an imaging device capable of capturing an image and an electronic processor in communication with the imaging device. The electronic processor, by executing a program stored in the electronic processor, receives the image from the imaging device and reduces the foreground of the image to a skeleton that captures one or more attributes of the foreground including at least one of curvature, connectivity, and extent. The skeleton defines a transformed coordinate system for quantifying one or more perfusion parameters within the microvasculature. In one embodiment, the skeleton is a line along the axis of the blood vessel of the microvasculature that bends according to the local curvature of the blood vessel. In one embodiment of the system, the electronic processor, by executing the program stored in the electronic processor, creates a transformed coordinate system by generating a plurality of grid lines that cover the entire width of a region of interest (ROI) of the microvasculature. In one embodiment of the system, the electronic processor, by executing the program stored in the electronic processor, creates a transformed coordinate system in which two axes are parallel and perpendicular to the blood flow respectively. The axis parallel to the blood flow is defined by the skeleton, and the axis perpendicular to the blood flow is defined by the normal of the skeleton. In one embodiment of the system, the electronic processor, by executing the program stored in the electronic processor, creates a collection of skeletons and lines created with reference to the skeleton that defines the x'-axis and the y'-grid lines of the transformed coordinate system. The y'-grid lines run in the direction of the blood flow of the microvascular ROI. In one embodiment of the system, the y'-grid lines of the transformed coordinate system have the same pixel length regardless of the curvature of the ROI.In one embodiment of the system, the electronic processor performs the step of drawing a spatio-temporal diagram of each time segment and each y'-grid line of the vascular block by executing the program stored in the electronic processor, where the vascular block is defined as a unit of length along the axis of the microvascular ROI. In one embodiment of the system, the electronic processor performs the step of integrating a plurality of spatio-temporal diagrams of individual y'-grid lines, time segments, and vascular blocks to calculate the blood flow velocity by executing the program stored in the electronic processor. In one embodiment of the system, the electronic processor performs the step of calculating the blood flow rate by multiplying the blood flow velocity by the cross-sectional area of the ROI by executing the program stored in the electronic processor. In one embodiment of the system, the electronic processor performs the step of calculating the total number of white blood cells along the slope of the spatio-temporal diagram and generating an intensity profile by executing the program stored in the electronic processor, where the intensity profile is further integrated from a plurality of y'-grid lines, time segments, and vascular blocks, and an estimated value of the number of white blood cells is obtained from the number of peaks of the integrated intensity profile. In one embodiment, the integration is performed by a dynamic time warping method to match the peaks of the intensity profiles of the vascular blocks while allowing for variations in the time delay between candidate white blood cells. In one embodiment of the system, the electronic processor performs the step of estimating the appearance time ("gate") of candidate white blood cells by determining the peak position of the intensity profile by executing the program stored in the electronic processor. In one embodiment of the system, the electronic processor performs the step of gating the approximate space and time of the appearance of candidate white blood cells in the video using the integrated intensity profile by executing the program stored in the electronic processor.

[0018] In another aspect, the present disclosure provides a system for imaging the microvasculature of a subject's tissue. The system includes an imaging device capable of capturing an image and an electronic processor in communication with the imaging device. The electronic processor, by executing a program stored in the electronic processor, receives the image from the imaging device, accesses a deep learning model trained with training data to detect perfusion and leukocyte feature data from the image input, and applies the image to the machine learning model to quantify one or more perfusion parameters within the microvasculature. In one embodiment of the system, the deep learning model is a neural network. In one embodiment of the system, the neural network is a convolutional neural network. In one embodiment of the system, the machine learning model is applied to a restricted spatial region and time (a "gate") that includes candidate leukocytes. In one embodiment of the system, the electronic processor, by executing the program stored in the electronic processor, detects the coordinates of the image in which leukocytes are detected and the probability score of the detection.

[0019] In yet another aspect, the present disclosure provides a method for in vivo flow cytometry of a subject's biological fluid. The method comprises: (a) contacting the subject's tissue with a tissue stabilizer to maintain the position of the subject's biological structure; (b) using an illumination device to provide light to a portion of the region of the biological structure and continuously illuminating the region of the biological structure; (c) using an image detector to continuously detect a microscopic image from the region of the biological structure based on light scattered by the biological structure of the subject, wherein the illumination is performed at an oblique angle by an offset shape of the illumination device; and (d) analyzing the microscopic image to identify characteristics of the biological fluid within the biological structure. In one embodiment of the method, detecting the microscopic image in step (c) includes generating an optical image by oblique back-illumination microscopy (OBM). In one embodiment of the method, detecting the microscopic image in step (c) includes generating an optical image by offset trans-illumination microscopy (OTM).

[0020] In one embodiment of the method, the biological structure is a microvessel of the subject, and step (d) includes quantifying one or more perfusion parameters in the microvessel.

[0021] In one embodiment of the method, the biological structure is a microvessel of the subject, and step (d) includes quantifying the number of white blood cells in the microvessel.

[0022] In one embodiment of the method, step (c) includes detecting the microscopic image in a label-free manner. In one embodiment of the method, step (c) includes detecting the microscopic image at a frame rate of 1 Hz to 1000 Hz.

[0023] In one embodiment of the method, the biological structure is a microvessel of the subject, and step (d) includes calculating an average rotation speed of leukocytes in the microvessel using automated frame-by-frame leukocyte tracking.

[0024] In one embodiment of the method, the biological structure is a microvessel of the subject, and step (d) includes reducing the foreground of each microscopic image to a skeleton that captures one or more attributes of the foreground, including at least one of curvature, connectivity, and extent, and the skeleton is used to quantify one or more perfusion parameters in the microvessel.

[0025] In one embodiment of the method, the biological structure is a microvessel of the subject, and step (d) further includes creating a transformed coordinate system by generating a plurality of grid lines that cover the entire width of a region of interest (ROI) of the microvessel.

[0026] In one embodiment of the method, the biological structure is a microvessel of the subject, and step (d) further includes creating a coordinate system transformed such that two axes are parallel and perpendicular to the blood flow, respectively, where the axis parallel to the blood flow is defined by the skeleton and the axis perpendicular to the blood flow is defined by the normal of the skeleton.

[0027] In one embodiment of the method, the biological structure is a microvessel of the subject, and step (d) includes creating a coordinate system transformed such that two axes are parallel and perpendicular to the blood flow, respectively.

[0028] In one embodiment of the method, step (d) includes creating a transformed coordinate system in which the axis parallel to the blood flow is defined by the skeleton and the axis perpendicular to the blood flow is defined by the normal of the skeleton.

[0029] In one embodiment of the method, the biological structure is a microvessel of the subject, and step (d) includes drawing a spatio-temporal diagram of each time segment and each y’ grid line of the blood vessel block, where the blood vessel block is defined as a unit of length along the axis of the microvessel ROI.

[0030] In one embodiment of the method, step (d) includes integrating a plurality of spatio-temporal diagrams of individual y’ grid lines, time segments, and blood vessel blocks to calculate a blood flow velocity.

[0031] In one embodiment of the method, the biological structure is a microvessel of the subject, and step (d) includes accessing a deep learning model trained with training data to detect perfusion and leukocyte feature data from the image input, and applying the image to the machine learning model to quantify one or more perfusion parameters in the microvessel.

[0032] These and other features, aspects, and advantages of the various embodiments of the present disclosure will be better understood with reference to the following description, the appended claims, and the accompanying figures.

Brief Description of the Drawings

[0033]

Figure 1A - B

Figure 2

Figure 3A - D

Figure 4A - B

Figure 5

Figure 6A - B

Figure 6C

Figure 7A - C

Figure 8A - B

Figure 9A - B

Figure 10A - B

Figure 11A - B

Figure 12A

Figure 12B

Figure 13A - B

Figure 14A - B

Figure 15A - B

Figure 16

Figure 17

Figure 18A - B

Figure 19A - B

Figure 20A

Figure 20B - C

Figure 21A - B

Figure 22

Figure 23A - D

Figure 24A - B

Figure 25A - C

Figure 26

Figure 27

Figure 28

Figure 29

Figure 30

Figure 31

Figure 32

Figure 33

Figure 34

Figure 35

Figure 36

Figure 37 - 1

Figure 37 - 2

DETAILED DESCRIPTION OF THE INVENTION

[0034] Before explaining the present invention in further detail, it is to be understood that the present invention is not limited to the specific embodiments described. It is also to be understood that the terms used herein are for the purpose of describing particular embodiments only and are not limiting. The scope of the present invention is limited only by the claims. The singular forms “a,” “an,” and “the” used herein include plural embodiments unless the context clearly dictates otherwise.

[0035] The disclosed subject matter is described herein in connection with particular embodiments and examples, but the invention is not necessarily so limited, and many other embodiments, examples, uses, modifications, and departures from the embodiments, examples, and uses are intended to be encompassed by the claims appended hereto, as will be understood by those skilled in the art.

[0036] It will be apparent to those skilled in the art that many additional changes are possible without departing from the concepts of the invention other than those already described. In interpreting this disclosure, all terms should be interpreted as broadly as possible in accordance with the context. Variations of the terms “comprising,” “including,” or “having” should be interpreted as referring to elements, components, or steps in a non-exclusive manner, such that the referenced elements, components, or steps can be combined with other elements, components, or steps not explicitly referenced. Embodiments referred to as comprising, including, or having particular elements are also contemplated as consisting essentially of and consisting of those elements, unless the context clearly dictates otherwise.

[0037] The present invention provides a system that enables stable non-invasive imaging of oral mucosal microvessels with high spatial resolution. This system may have two main components: (1) an oral mucosa stabilizer and (2) a small imaging device. By combining the two components, it becomes possible to perform label-free high-resolution imaging of exposed microvessels and their contents (red blood cells, white blood cells, platelets, endothelial tissue / cells, epithelium, etc.) while minimizing artifacts due to movement. The acquired image data is processed using a custom algorithm adapted to the system. The results (such as estimation of blood cell count, classification of movement and subtypes) provide valuable medical information regarding the patient's immune system. This system can be used as a diagnostic tool for safely monitoring the conditions of healthy patients (for preventive diagnosis) and unhealthy patients (such as acute and systemic inflammation, sepsis, tissue hypoxia, etc.) on a daily basis. Although part of the same imaging system for non-invasive imaging of oral mucosal microvessels, each main component is often presented individually below. Various oral mucosal sections include, but are not limited to, the buccal mucosa, lips (external and internal sections), floor of the mouth, tongue, gingiva, palate, tonsils, etc.

[0038] Embodiments of the oral mucosa stabilizer may - a mechanical holder with a function to correct the natural position of oral mucosal tissue by mechanical contact, - a mechanical holder with a function to maintain the corrected position of oral mucosal tissue, - a mechanical holder that exposes the microvessels of oral mucosal tissue, - a mechanical holder that enables the data collection device to preferentially access the microvessels of the oral mucosa, - a mechanical holder that corrects the natural position of oral mucosal tissue regardless of the presence or absence of an integrated optical system (e.g., a lens) or a light source (e.g., an LED), - a combination of mechanical holders that maintain the corrected position of oral mucosal tissue, and may include the features of.

[0039] Embodiments of the oral mucosa stabilizer may - Modifying the natural position of oral mucosal tissue in a non-invasive manner, - Exposing the microvessels of the oral mucosa, - Enabling privileged access to the data collection device, - Maintaining a stable position of the oral mucosal tissue during data acquisition, - Minimizing mechanical vibrations of the oral mucosal tissue during data acquisition, and can be used for any of the purposes.

[0040] The data collection device includes, but is not limited to, optical technologies such as single-photon, multi-photon, confocal, laser scan, systems with coherent light sources, systems with non-coherent light sources (e.g., LEDs, halogen lamps).

[0041] Referring to Figure 1A, one embodiment of the fixed bench-top tissue stabilizer 100 according to the present invention is shown as being disposed on the subject 50. The bench-top tissue stabilizer 100 includes a jaw holder 110 and a forehead holder 120 fixed to the subject 50 by a strap 122. The jaw holder 110 is provided with a vertical slide bar 131 and a horizontal slide bar 136 perpendicular to the base of the jaw holder 110 to adjust the vertical and horizontal positions of the adapter 140. The adapter 140 non-invasively modifies the natural position of the oral mucosal tissue 60 and maintains a stable position of the oral mucosal tissue 60 during data acquisition. The vertical bracket rod 137 connects the bench-top tissue stabilizer 100 to the bench-top. Figure 12A shows the method of using the bench-top tissue stabilizer 100, moving the adapter 140 towards the subject 50 (left), contacting the subject (center), and then the adapter 140 maintaining a stable position of the oral mucosal tissue 60 (right). Figure 13A shows the method of using the bench-top tissue stabilizer 100 in a system for imaging the microvessels of the tissue of the subject 50, and the stabilizer can be used with the subject sitting.

[0042] Referring to FIGS. 1B and 2, one exemplary embodiment of a portable tissue stabilizer 200 according to the present invention being worn on a human subject 50 is shown. The portable tissue stabilizer 200 includes a holder 210 for adjusting an adapter 240, which non-invasively modifies the natural position of the oral mucosal tissue 60 and maintains a stable position of the oral mucosal tissue 60 during data acquisition. An elastic strap 212 secures the holder 210 to the subject 50. The tissue stabilizer 200 is portable and can be carried by a subject (adult, infant, neonate). FIG. 12B shows how to use the portable tissue stabilizer, by moving the adapter 240 downward toward the subject 50 (left), contacting the subject (center), and then the adapter 240 maintaining a stable position of the oral mucosal tissue 60 (right). FIG. 13B shows how to use the portable tissue stabilizer on a human subject 50, where the stabilizer can be used with the subject lying on their side.

[0043] Referring to FIGS. 3A and 3B, one embodiment of a multi-component stabilizer tissue stabilizer 300 according to the present invention is shown. In FIG. 3A, the adapter 340 is shown as a dashed rectangle. In FIG. 3B, the adapter holder 341 is shown as a dashed rectangle. In FIG. 3C, the slide mechanism is shown as a dashed rectangle. In FIG. 3D, the jaw holder 310 and the forehead holder 320 are each shown as a dashed rectangle. The tissue stabilizer 300 includes a jaw holder 310 and a forehead holder 320 attached to a frame 324. The slide mechanism includes a vertical slide bar 331 to which a movable collar 332 is attached, which can adjust the vertical position of the adapter 340 and be fixed in place by a fixing screw 333. The adapter 340 is attached to the vertical slide bar 331 by an adapter holder 341. The vertical slide bar 331 is moved up and down and then fixed in place by the fixing screw 333 to adjust the vertical position of the adapter 340, non-invasively modify the natural position of the oral mucosal tissue 60, and maintain a stable position of the oral mucosal tissue 60 during data acquisition.

[0044] Referring to FIG. 4A, one exemplary embodiment is shown in which a jaw holder 400 of a tissue stabilizer according to the present invention is disposed on a human subject 50. The jaw holder 400 includes a base 410a having a horizontal wall 414a connected to a vertical wall 416a. The oral mucosal tissue 60 of the lip is spread and disposed against the vertical wall 416a of the jaw holder 400. To stabilize the imaging region, the oral mucosal tissue 60 can be (gently) pressed against the solid surface of the vertical wall 416a of the jaw holder 400.

[0045] Referring to FIG. 4B, one exemplary embodiment is shown in which a jaw holder 450 of a tissue stabilizer according to the present invention is disposed on a subject 50. The jaw holder 450 includes a base 410b having a horizontal wall 414b connected to a vertical wall 416b. The oral mucosal tissue 60 is pressed against the subject.

[0046] Referring to FIGS. 5, 6A, 6B, and 6C, one embodiment of a multi-component tissue stabilizer 500 according to the present invention is shown. FIGS. 8A and 8B show the multi-component tissue stabilizer 500 being used on a human subject 50. FIG. 8B is a detailed view of the oral mucosa tissue region, with the microvessels 70 exposed for inspection and imaging. The slide mechanism enables the precise positioning of the stabilizer and the optimization of the pressure applied to the oral mucosa tissue 60. The tissue stabilizer 500 includes a jaw holder 510. The slide mechanism includes a vertical slide bar 531 to which a movable collar 532 is attached, which can adjust the vertical position of the adapter 540 and be fixed in a predetermined position by a fixing screw 533. The adapter 540 is attached to the vertical slide bar 531 by an adapter holder 541. The vertical slide bar 531 can be moved up and down and then fixed in a predetermined position by the fixing screw 533 to adjust the vertical position of the adapter 540, non-invasively correct the natural position of the oral mucosa tissue 60, and maintain a stable position of the oral mucosa tissue 60 during data acquisition. The slide mechanism also includes a horizontal slide bar 536 to which a movable collar 537 is attached, which can be fixed in a predetermined position by a fixing screw 538. The horizontal slide bar 538 can be moved laterally and then fixed in a predetermined position by the fixing screw 538 to adjust the horizontal position of the adapter 540, non-invasively change the natural position of the oral mucosa tissue 60, and maintain a stable position of the oral mucosa tissue 60 during data acquisition.

[0047] Thus, the slide system (e.g., FIGS. 3A - 3B, 5, 6) enables the spatial positioning (linear or angular) of the adapter of the tissue stabilizer. This generates a normal force to maintain the altered position of the tissue and allows for the adjustment and optimization of the normal force applied to the tissue. It contributes to the stabilization of the tissue and enables the rapid adjustment and fixation (within 0 - 10 minutes) of the tissue stabilizer before and during imaging. The operation is usually manual, but the system can also be automated.

[0048] Referring to FIGS. 7A and 7B, one embodiment of an adapter 700 of a tissue stabilizer according to the present invention is shown. The adapter 700 includes an adapter holder 741 for attachment to a slide mechanism, similar to the tissue stabilizer 500. Opposing tabs 742 create lateral pressure points that enhance the stability of the oral mucosa tissue 60. This minimizes artifacts due to unwanted vibrations and movements, flattening the surface of the exposed tissue area. The tabs 742 help maintain focus (depth) during navigation (mechanical and optical) across the entire tissue. This optimizes tissue position and microvascular exposure before and during data acquisition. The lateral (XY) dimension of the pressure points typically varies in the range of about 0.1 to 5 mm.

[0049] FIG. 7B shows an adaptive curvature that conforms to the surface curvature of the oral mucosa tissue, thereby evenly distributing the mechanical pressure applied to the oral mucosa tissue 60 and minimizing discomfort to the subject. This contributes to the optimization of tissue position and microvascular exposure before and during data acquisition. The radius of curvature typically varies between about 2 and 100 mm. Referring to FIG. 7B, the angled design conforms to the surface curvature of the oral mucosa tissue and contributes to the even distribution of mechanical pressure applied to the oral mucosa tissue. This minimizes discomfort to the subject, optimizes tissue position during data acquisition, and exposes the microvessels. The angle of inclination typically varies between about 45 and 90 degrees.

[0050] Referring to FIG. 7A, the patterned surface finish 745 of the adapter 700 increases the contact area with the oral mucosal tissue 60, increasing the friction between the oral mucosal tissue and the stabilizer. This contributes to the stabilization of the oral mucosal tissue, optimizes the position of the tissue, and exposes the microvessels during data collection. The spacing between repeating features (such as extruded circles, squares, etc.) typically varies between about 1 and 15 mm. The dimensions of the repeating features typically vary between about 1 and 5 mm in each dimension (X, Y, Z). The vacuum generated by the adapter 700 creates air suction and maintains the position of the oral mucosal tissue. This optimizes the position of the tissue and exposes the microvessels during data collection. The vacuum pressure varies between about 0.1 and 5 SCFH.

[0051] Referring to FIG. 7C, another exemplary embodiment of the adapter 750 of the tissue stabilizer according to the present invention is shown. The adapter 750 includes an adapter holder 791 for attachment to a slide mechanism similar to the tissue stabilizer 500. Opposing tabs 792 create lateral pressure points that enhance the stabilization of the oral mucosal tissue. The adapter 750 has a patterned surface finish 795 that functions similarly to the patterned surface finish 745 of the adapter 700.

[0052] The adapters 700, 750 can have various dimensions. The dimensions of the section in contact with the oral mucosal region are X: 1 to 100 mm, Y: 1 to 100 mm, Z: 1 to 100 mm. The tabs 742, 792 can create lateral pressure points with dimensions such as X: 1 to 50 mm, Y: 1 to 50 mm, Z: 1 to 50 mm. The dimensions of the vacuum region are X: 1 to 100 mm, Y: 1 to 100 mm, Z: 1 to 100 mm. The dimensions of the patterned texture region 745 are X: 1 to 100 mm, Y: 1 to 100 mm, Z: 1 to 100 mm. The dimensions of the steps of the patterned texture region 745 are X: 0.1 to 20 mm, Y: 0.1 to 20 mm, Z: 0.1 to 20 mm.

[0053] Referring now to FIGS. 9A and 9B, another embodiment of the adapter 900 of the tissue stabilizer according to the present invention is shown. The adapter 900 includes a lower wall 940 and an adapter holder 941 for attachment to a slide mechanism, similar to the tissue stabilizer 500. Opposing tabs 942 create lateral pressure points that enhance the stabilization of the oral mucosa tissue 60. This minimizes artifacts due to unwanted vibrations and movements and flattens the surface of the exposed tissue area. The tabs 942 help maintain focus (depth) during navigation (mechanical and optical) across the tissue. This aids in optimizing tissue position and exposing microvessels before and during data acquisition. A transparent sheet 944 is attached between the tabs 942. Depending on the material of the adapter 900, the performance of the stabilizer can be optimized (e.g., glass sheet 944 + biocompatible photoresin). The number of materials is typically about 1 to 5. Coatings can be used to change the surface finish texture to enhance frictional contact, make the material biocompatible (sterilizable), or make the material hydrophobic (e.g., SigmaCote® silanization reagent).

[0054] Looking at FIGS. 10A and 10B, another exemplary embodiment of the adapter 1000 of the tissue stabilizer according to the present invention is shown. The adapter 1000 includes an adapter holder 1041 for attachment to a slide mechanism, similar to the tissue stabilizer 500. A rod 1047 is connected to the adapter holder 1041 using a connector 1048. FIG. 10B shows the adapter 1000 being used on a human subject 50, where microvessels are exposed for future investigation and imaging, and the rod 1047 is stabilizing the oral mucosa tissue 60 for imaging.

[0055] Referring to FIGS. 11A and 11B, another exemplary embodiment of the adapter 1100 of the tissue stabilizer according to the present invention is shown. The adapter 1100 includes an adapter holder 1148 for attachment to a slide mechanism, similar to the tissue stabilizer 500. A flexible elastic loop 1149 is connected to the adapter holder 1148. FIG. 11B shows the adapter 1100 being used on a human subject 50, where the microvessels are exposed for future investigation and imaging, and the loop 1149 is stabilizing the oral mucosal tissue 60 for imaging.

[0056] One method of using the tissue stabilizer according to the present invention is Step 1: Position the subject (see FIGS. 12A (left), 12B (left)), Step 2: Adjust the position of the stabilizer adjacent to the target oral mucosal tissue (see FIGS. 12A (center), 12B (center)), Step 3: Spread the target oral mucosal tissue to expose the microvessels (see FIGS. 12A (right), 12B (right)). and includes the steps of.

[0057] By the stabilizer applying a mechanical force to the non-image area of the tissue, precise positioning is possible to maintain the optimal position of the oral mucosal tissue. The amplitude of the force is optimized to avoid changes in blood flow in the microvascular system of the oral mucosal tissue. The parameters adjusted to maintain the position of the oral mucosal tissue without changing blood flow are - the spatial arrangement of the stabilizer with respect to the oral mucosal tissue (e.g., FIGS. 5, 6A - 6C), and - the strength of the vacuum suction system of the stabilizer, and - the material of the stabilizer, and - the mechanical design (geometric shape, curvature) of the stabilizer (FIGS. 3A - 3D), and - the surface finish / texture of the stabilizer area in contact with the oral mucosal tissue (e.g., patterned surface, coating, etc.) (FIGS. 7A - 7C). and include.

[0058] Referring to FIG. 14A, another exemplary embodiment of the jaw holder 1400 of the tissue stabilizer according to the present invention is shown. The jaw holder 1400 includes a vertical wall 1416 having a light output end 1446 of an optical fiber that transmits light from a light source of the lighting device. The integrated light source and optical system enable illumination of the oral mucosal tissue (such as oblique rear illumination, offset transmission illumination, etc.), enabling visual inspection of the tissue and illumination of the data acquisition device.

[0059] Referring to FIG. 14B, another exemplary embodiment of the adapter 1450 of the tissue stabilizer according to the present invention is shown. The adapter 1450 includes a patterned surface finish 1495 that increases the contact area with the oral mucosal tissue. The adapter 1450 includes a light output end 1496 of an optical fiber that transmits light from a light source of the lighting device. The adapter 1450 includes a line of an objective lens 1497 that collects scattered light from the tissue 60. The integrated light source and optical system enable illumination of the oral mucosal tissue (such as oblique rear illumination, offset transmission illumination, etc.), enabling visual inspection of the tissue and illumination of the data acquisition device.

[0060] Thus, the integrated light source and optical system (such as FIGS. 14A and 14B) enable illumination of the oral mucosal tissue (such as oblique rear illumination, offset transmission illumination, etc.). This enables visual inspection of the tissue and illumination of the data acquisition device.

[0061] Referring to FIGS. 15A, 15B, 16, and 17, components of an exemplary embodiment of the system 1500 according to the present invention are shown. This system has a bench-top tissue stabilizer 1501 disposed on a subject 50 and an imaging device 1550. The imaging device 1550 includes a housing 1552 and an imaging unit having a chip 1554. The housing can be miniaturized (for example, about 5 cm in length). The bench-top tissue stabilizer 1501 may be configured like the tissue stabilizer 100 described above. At least two methods can be used for focus adjustment. 1. Method 1: Mechanically move the camera or relay optics using actuator motor 1599 (see Figure 17). This includes a slide system for maintaining accurate alignment of the movable element (lens or camera / sensor). Or, 2. Method 2: Electric focus adjustment using an electrically adjustable lens (such as a liquid crystal lens, an electrowetting lens, etc.).

[0062] Figure 18A is an optical design of the light collection path of an exemplary embodiment of an imaging device of the system according to the present invention. The optical design includes a gradient index (GRIN) objective lens (NA≈0.75) 1870, a relay GRIN lens 1871, a double achromat lens 1872, and an image detector (e.g., a sensor / camera) 1890. Advantageous features of one embodiment of the imaging device include a high-speed CMOS sensor (up to 1000 fps) and a small aberration-corrected GRIN objective lens (NA≈0.75) for monitoring white blood cells moving rapidly within blood circulation with high resolution.

[0063] Figure 18B is another optical design of the light collection path of another exemplary embodiment of the imaging device of the system according to the present invention. This optical design includes a standard imaging GRIN objective lens (NA≈0.5) 1870a, a double achromat lens 1872a, and an image detector (e.g., a sensor / camera) 1890a.

[0064] The features of the optical designs of Figures 18A and 18B are light collection paths having three main components. 1. Objective lens. A lens that collects light from tissues. In particular, a special objective lens called a gradient-index (GRIN) lens is used. The GRIN lens is small (with a diameter of about 0.5 - 5 mm and a length of about 0.5 - 10 mm) and has a very short working distance (0 - 200 μm), so it is suitable for this application targeting shallow microvessels. Also, since the GRIN lens has a high numerical aperture (NA of about 0.5 - 1), the optical resolution of the system of the present invention is improved (see Fig. 18A). Furthermore, in the system of the present invention, a special GRIN lens assembly that is aberration-corrected, has a high NA of about 0.75, and generates a magnification of about 3 - 5 by itself can be used. 2. Relay optical system. The relay optical system transports light from the objective lens to the sensor. Aberration-corrected lenses (such as double achromat lenses) are advantageous. The relay optical system can be a single lens or multiple lenses (see Figs. 18A and 18B). 3. Sensor / Camera. When used as an image detector, the sensor / camera may be monochrome (black / white), binned 2×2 or 4×4, visible and near-infrared wavelengths, high acquisition rate (30 - 1000 fps), and compact size. See Figs. 18A and 18B.

[0065] Fig. 19A shows an example of the optical injection configuration and imaging modality of an imaging device 1900 including a GRIN objective lens 1970 and the optical output end 1962 of an optical fiber 1963 that transmits light from a light source. The illumination device uses oblique back-illumination microscopy (OBM).

[0066] Fig. 19B shows another exemplary optical injection configuration and imaging modality of an imaging device 1940 including a GRIN objective lens 1970 and the optical output end 1962 of an optical fiber 1963 that transmits light from a light source. The illumination device uses offset trans-illumination microscopy (OTM).

[0067] Therefore, the imaging modality is defined by the position of the illumination source (such as an optical fiber) with respect to the light collection optical system. Placing the illumination source at three different positions gives rise to three different non-limiting imaging modalities. (1) Oblique back illumination microscopy (OBM) (see FIG. 19A), (2) Offset transmission illumination microscopy (OTM) (see FIG. 19B).

[0068] Referring to FIGS. 20A, 20B, and 20C, another exemplary embodiment of the imaging device 2050 of the system 2000 according to the present invention is shown. The imaging device 2050 includes a housing 2052 and an imaging unit having a chip 2054. The system includes a controller 2080 that may include a memory and an electronic processor connected to the memory. The chip 2054 has an annular vacuum channel 2056, a perfusion channel 2058, an optical output end 2062 of an optical fiber that transmits light from a light source, and an objective lens 2070.

[0069] The imaging chip 2054 has advantageous characteristics. For example, the refractive index distribution type lens (objective lens 2070) has a cavity that stabilizes the GRIN objective lens. In particular, the diameter of the cavity is slightly smaller than the diameter of the GRIN lens (10 - 100 μm). When the GRIN objective lens 2070 is inserted, the cavity expands and a force perpendicular to the surface of the lens is generated. This normal force causes friction and holds the GRIN objective lens 2070 in the cavity without using additional mechanical support (such as screws). The illumination source cavity allows the illumination source (fiber or LED) to be inserted next to the objective lens. The vacuum channel 2056 enables the stabilization of tissue using air suction. In particular, the diameter of the vacuum channel 2056 is small and compact for a specific purpose. This allows the vacuum region to be quickly and efficiently filled without leaving a gap that reduces the suction efficiency. Large vacuum cavities are difficult to fill, especially in vivo because the tissue has an asymmetric and irregular profile. The perfusion channel 2058 hydrates the tissue and increases the numerical aperture of the imaging objective lens. Water or biocompatible oil can be considered as an immersion medium that is constantly updated in the imaging area. This can compensate for the amount of immersion medium removed by the vacuum system, absorption by the tissue, evaporation, etc. In addition to the immersion medium, pharmacological agents can also be delivered to the mucosal surface.

[0070] Next, referring to FIGS. 21A, 21B, 21C, and 22, another exemplary embodiment of the imaging device 2100 according to the present invention with a disposable chip is shown. The imaging device 2100 includes a housing 2152 and an imaging unit with a disposable chip 2154. The chip 2154 includes an annular vacuum channel 2156, a perfusion channel 2158, the optical output end 2162 of an optical fiber that transmits light from a light source, and an objective lens 2170. The disposable chip 2154 ensures that the part of the device that comes into contact with the patient is in a sterile state. The complex design can be realized by 3D printing.

[0071] Referring to FIGS. 23A and 23B, another exemplary embodiment of a portable system 2300 for imaging the microvasculature of tissue according to the present invention, which is worn on a human subject, is shown. The portable system 2300 includes a movable first arm 2302 and an opposing second arm 2304 that defines a space 2305. A hinge 2306 connects the first arm 2302 and the second arm 2304.

[0072] FIG. 23C shows one version of a portable system 2330 similar to the portable system 2300. The portable system 2330 includes a movable first arm 2302a and an opposing second arm 2304a that defines a space for receiving the oral mucosa 60. A light source 2364, an objective lens 2070, and an image detector (e.g., a sensor / camera) 2390 are disposed on the first arm 2302a, and the portable system 2330 is in an oblique rear illumination configuration, as shown in FIG. 23C.

[0073] Looking at FIG. 23D, another version of a portable system 2360 similar to the portable system 2300 is shown. The portable system 2360 includes a movable first arm 2302b and an opposing second arm 2304b that defines a space for receiving the oral mucosa 60. The light source 2364 is disposed on the second arm 2304b, and the objective lens 2070 and the image detector (e.g., a sensor / camera) 2390 are disposed on the first arm 2302b, and the portable system 2360 is in an oblique transmission illumination configuration, as shown in FIG. 23D.

[0074] Next, referring to FIGS. 24A, 24B, 25A, 25B, and 25C, another exemplary embodiment of a system 2500 for imaging the microvasculature of tissue according to the present invention, disposed on a human subject 50, is shown. This system includes a tissue stabilizer 2510 (similar to the tissue stabilizer 500 described above) and an imaging device 2550. The imaging device 2550 includes a housing 2552 having an imaging section with a chip 2554. The chip 2554 has an annular vacuum channel 2556, a perfusion channel 2558, an optical output end 2562 of an optical fiber for transmitting light from a light source, and an objective lens 2570. The imaging device 2550 includes a controller 2580 that may include a memory and an electronic processor connected to the memory.

[0075] Referring to FIG. 26, another embodiment of a system 2600 for imaging the microvasculature of tissue according to the present invention, disposed on a human subject 50, is shown. System 2600 uses a microscope objective lens. The system includes a tissue stabilizer 2601 with an adapter 2640 (similar to the tissue stabilizer 500 described above) and an imaging device 2650. The imaging device 2650 includes an LED light source 2664, an optical transmission fiber 2663 with an optical output end, a GRIN objective lens 2660, a microscope objective lens 2675, a double achromat lens 2672, a CMOS camera 2690, and an electric XYZ system 2695 for moving the imaging device 2650.

[0076] The tissue stabilizer can be composed of biocompatible, sterile, or sterilizable materials. One stabilizer can be composed of multiple materials (e.g., plastic + glass, metal + plastic, polymer resin + metal + glass, etc.). The materials are biocompatible and include metals (such as stainless steel, titanium, etc.), plastics, polymers, ceramic biomaterials, 3D printing materials (resins or photo resins), silicon, short-term and long-term implantable materials, etc. The materials can be any of sterile (materials sterilized by the manufacturer), single-use per patient, disposable, sterilizable (materials sterilizable by the user (certified researcher, clinician, etc.)), reusable, and materials sterilizable by chemical sterilants, radiation, moist heat, dry heat, etc.

[0077] The system of the present invention has many advantages. The miniaturized imaging device (see FIGS. 15 to 26) enables non-invasive, label-free visualization and recording of the contents of microvessels in in vivo oral mucosal tissue. To generate contrast between various types of biological tissues and cells, the device of the present invention combines the following two. 1. Phase gradient contrast. This is obtained by irradiating the tissue obliquely and enables visualization of the detailed fine morphology of tissues (such as membranes) and blood cells (such as membranes, granules, nucleosomes, etc.) that are inaccessible with existing commercially available devices (CytoCam, MicroScan, etc.). 2. Absorption contrast. This is made possible by using light in a specific wavelength range (530 + / - 20 nm) that is strongly absorbed by red blood cells (which carry hemoglobin) compared to white blood cells. This may also help to distinguish between oxygenated and non-oxygenated red blood cells.

[0078] The system of the present invention includes, but is not limited to, 1. Real-time imaging and characterization of the morphology of oral mucosal tissue. 2. Real-time imaging and characterization of the microvessels of the oral mucosa and their contents. 3. Quantification of blood contents (e.g., white blood cells, red blood cells, platelets, etc.). 4. Characterization of the morphology / shape of blood vessels and blood cells. 5. Characterization of the movement of blood cells. and has many purposes.

[0079] The oral mucosa stabilizer is not limited to small devices and can also be used in bench-top imaging configurations such as multiphoton systems, pulsed laser systems, and scanning systems.

[0080] Software operation method The system of the present invention includes image data processing. The software can identify, count, and track the contents of microvessels and cells from the records obtained from the image system. The operation can be performed completely automatically, semi-automatically, or manually. The code may be described in an existing programming language (such as Python, C++) as an executable script or program, and / or as a script and macro of existing commercial software or open-source software including ImageJ, Fiji, MATLAB®, Imaris, etc. The purpose of the software is to (1) report the number of cells per volume of different blood cell types in the microvessels and the type of cell movement, and (2) report the perfusion characteristics of the imaged blood vessels as a quality indicator of (1). For the pipeline adopting Methods 2 to 7, refer to Example 2 below.

[0081] Image analysis method #1: Intensity profile This method is based on the extraction of the temporal intensity profile within a selected region of interest (ROI) in the field of view. The temporal intensity records from single or multiple ROIs can be - for extracting blood flow and perfusion parameters, - for measuring the heart rate, - for detecting the presence of white blood cell (WBC) candidates on the circulation or blood vessel wall, - for confirming the detection of WBC candidates (using multiple ROIs at different locations of the blood vessel), - for reducing the background of the image, and can be used for these purposes.

[0082] Image analysis method #2: Selection of region of interest In this method, a volume within the video is specified for analyzing the contents of the microvessels. The specified volume does not include pixels other than the blood vessels.

[0083] The pixel attributes that can be used as the basis for the specification include, but are not limited to, intensity, change in intensity as a function of time, gradient of intensity along the edge, etc.

[0084] Any algorithm that measures and segments a specific image based on these attributes can be used. User input may be required to verify and correct the output of the automated algorithm.

[0085] The output of the region of interest selection includes, but is not limited to, a specified volume that can perform microvascular content analysis.

[0086] Image Analysis Method #3: Coordinate Transformation In this method, a transformed coordinate system that maintains the shape of the region of interest (ROI) of the microvessels is created. The main feature of such a transformed coordinate system is that two axes run parallel and perpendicular to the blood flow, respectively.

[0087] The axis x' parallel to the blood flow is defined using skeletonization. Skeletonization is a conventional image processing algorithm that reduces the foreground of a binary image to a "skeleton" that best captures the main attributes of the foreground, such as curvature, connectivity, and extent. In the case of the foreground sharing the shape of the blood vessels, the skeleton becomes a line that bends along the axis of the blood vessels according to the local curvature.

[0088] Additional grid lines along x' are drawn based on the skeleton, thereby also capturing the morphology of the microvascular ROI. Morphological operations of image processing (such as dilations and erosions) may be used to perform this step.

[0089] The axis y' perpendicular to the blood flow is defined by the normal of the skeleton. These normals intersect the skeleton respectively and have an inclination perpendicular to the inclination of the skeleton at the intersection points.

[0090] With the transformed coordinate system, the contents of the microvessels can be analyzed as a function of space (x’, y’) and time (t). The measurement units can be defined along the three axes. The software includes values that are pre-optimized but can be defined by the user for these distance and time units. The vascular block (unit along x’) is defined by the number of pixels along the skeleton. The y’ grid line index (unit along y’) refers to a specific y’ grid line where the measurement is taken. The time segment (unit along t) is specified by the number of video frames.

[0091] The output of the coordinate transformation is - the transformed coordinate system, - the units of microvessel content measurements defined using the transformed coordinate system, - the morphological parameters (length, width, cross-sectional area, volume, curvature, etc.) of the microvessel ROI, including but not limited to these.

[0092] Image Analysis Method #4: Digital Line Scan This method creates a spatio-temporal map using digital line scan and extracts perfusion parameters from it. The spatio-temporal map is usually defined as a 2D plot with time on one axis and distance along a dimension on the other axis, showing the position of an object as a function of time.

[0093] For each measurement unit, i.e., each vascular block, y’ grid line index, and time segment, a spatio-temporal map can be generated. The y-axis of these spatio-temporal maps represents the elapsed time since the start of the time segment, and the x-axis represents the length of the pixels along the skeleton of a specific vascular block.

[0094] Leukocytes and other high-intensity objects moving along the skeleton, or objects moving along the y’ grid lines running parallel to the skeleton, are displayed as bright sloped bands in the corresponding spatio-temporal maps. The slope of these bands corresponds to the velocity of the object.

[0095] The width of the band in the spatio-temporal diagram is related to the size of the object. A digital image filter may be applied to the diagram to extract the velocity values of objects within a specific size range.

[0096] The appearance time of an object within a measurement unit can be determined by taking the cumulative sum of pixel intensities along the slope of the band. Such a sum can be performed using sliding windows of different lengths and processed in combination with various image processing algorithms, ensuring that small white blood cells are detected evenly and allowing for deviations in the movement speed from the blood flow velocity values.

[0097] The output of the digital line scan includes - the spatio-temporal diagram for each measurement unit, - the velocity of bright objects (including candidate white blood cells) moving within the measurement unit, - the intensity profile as a function of x' and t within the measurement unit, but is not limited to these.

[0098] Image analysis method #5: Integration This method describes the process of assembling the output from multiple measurement units specified by the blood vessel block, y' grid line index, and time segment, and groups the output by source object.

[0099] Bright objects (such as candidate white blood cells) flowing within the vascular system are expected to be captured in multiple time segments with a time delay along x'. The direction of blood flow reflects the time required for an object to move downstream from one measurement unit to the next along x' (the blood vessel block). The flowing object may or may not span the entire width of the blood vessel. That is, it may be captured at a specific y' grid line but not at other grid lines. These elements are considered in the integration.

[0100] Integration is performed across all three analysis axes of (x’, y’) in the transformed coordinate system and time t. For integration along y’ by the y’ grid line index, a pixel-based rearrangement strategy is used to ensure that all fluid objects are captured regardless of their relative size compared to the overall width of the blood vessel. For integration along t, information from all time segments is combined to cover the entire length of the video. For integration along x’ by blood vessel blocks, information from different blood vessel blocks is aligned by matching fluid objects identified in one blood vessel block with those identified in another blood vessel block.

[0101] For matching between blood vessel blocks of the fluid, any image processing method that performs peak matching can be used, such as dynamic time warping (DTW) or cross-correlation. The output of the matching between blood vessel blocks is a set of intensity profiles that match one for each blood vessel block where the positions of the peaks coincide. The peak shift required for matching is related to the time delay that a fluid object moves from one blood vessel block to another, i.e., its velocity.

[0102] Each set of matching peaks between blood vessel blocks is assigned an index number. Each index number identifies one source object (such as a candidate white blood cell) that may contribute to the set of matching peaks. The number of these indexes is called the “gate index” and is an estimate of the number of candidate white blood cells.

[0103] The output of the integration includes, - the integration velocity as a function of x’ (the direction of blood flow) and time, - one integration intensity profile for each blood vessel block as a function of time, - a list of “gate indexes” each of which points to a candidate white blood cell, - an estimate of the number of candidate white blood cells, but is not limited to these.

[0104] Image analysis method #6: Gating This method specifies the spatial and temporal volume in which objects (such as candidate white blood cells) flowing within the video appear. One gate corresponds to one candidate white blood cell, and an index number ("gate index") is assigned during integration.

[0105] Gating is preparation for trimming the video into a collection of videos with shorter durations and smaller areas for the purpose of confirming the presence of white blood cells and the final white blood cell count. Each of these shorter and smaller videos approximates the appearance amount of candidate white blood cells. Since the white blood cell concentration in the blood is usually low (about 0.15% of the red blood cell concentration in healthy people), this trimming significantly speeds up the white blood cell confirmation and final counting process. As a gating technique, candidate white blood cell information from integration defined in the transformed coordinate system and time is acquired and converted into the (orthogonal) coordinate system and time of the video. The output is a volume ("gate") for each candidate white blood cell delimited by a bounding box within a continuous series of video frames that specifies the spatial and temporal positions of the appearance of cells within the video.

[0106] The output of gating is - Bounding box coordinates corresponding to the appearance amounts of candidate white blood cells in the video, sorted by gate index and includes, but is not limited to, these.

[0107] Image analysis method #7: Particle tracking by detection and linking The detection and counting of white blood cells can be performed using the detection and linking method for particle tracking. In this method, first, particles within individual frames of the video are detected. Then, the particles detected from consecutive frames are linked and their movement is tracked. The number of particles is determined by the number of tracks.

[0108] This method can be executed either alone or as part of a pipeline where the amount of candidate white blood cells has already been reported. In this case, this method helps to confirm the presence of the already identified candidate white blood cells and verify that their number is equal to the estimated value of the candidate white blood cell count. The bright bands in the spatio-temporal diagram are mainly white blood cells, but rarely, especially in very narrow microvascular ROIs, there may be regions of plasma where no blood cells are flowing. Furthermore, white blood cells may appear as multiplicities (doublets, triplets, etc.), in which case one gate index represents multiple white blood cells.

[0109] Therefore, the input to this method may be the complete video or a video cropped from the complete video using the spatial and temporal coordinates provided by gating.

[0110] (i) Detection The particles in this method are blood cells present in the microvasculature. The detection of these particles or cells can be performed manually, semi-automatically, or automatically using an algorithm that discriminates between foreground and background of different intensities.

[0111] Such algorithms may be algorithms from conventional image processing and include algorithms aimed at blob detection (Hessian determinant, Laplacian of Gaussian, etc.) and segmentation (thresholding, graph cut, WEKA®, etc.).

[0112] Such algorithms may be obtained from computer vision techniques. A convolutional neural network (CNN) is a deep learning algorithm that can learn features within an image. Various CNN layer architectures (including but not limited to various versions of SqueezeNet, ResNet, EfficientNet, MobileNet, R-CNN, YOLO) may be used. Modifications or combinations of this architecture may also be used. The CNN model is first trained to recognize and locate blood cells. The same model or another model can be trained to identify the subtypes (such as granulocytes, lymphocytes, etc.) of each cell.

[0113] The training data for the CNN model provided by software is a video of the microvessels of oral mucosal tissue recorded with acquisition parameters similar to the video to be analyzed. Videos of both healthy and diseased patients are used for training. Therefore, the weights and biases of the model are optimized according to the imaging system and application. Transfer learning can be used for the training of the CNN model.

[0114] The output of the detection is - the position of the detected cells, which can be reported as a list of pixels or as bounding boxes within each video frame, - the subtype of the cells, - the confidence score of the above prediction, and includes but is not limited to these.

[0115] (ii) Link The software provides estimated values of the motion parameters of each cell motion type (such as free-flowing, rolling, crawling, tethering, etc.) and ensures appropriate linkage. These parameters include but are not limited to speed, preferred position within the blood vessel lumen, directionality, etc.

[0116] To address the problem of cells being intermittently occluded by other cells within a video, the particle link algorithm can be executed regardless of the presence or absence of state estimation. The software also provides parameters optimized for such algorithms, such as a Kalman filter or a particle filter.

[0117] The output of the link is - The number of cells, with the option of counting subtypes (e.g., granulocytes, lymphocytes) within the input video, - The trajectory of the movement of each cell within the input video, - The type of movement of each cell within the input video (e.g., free flow, rotation, crawling, tethering), including, but not limited to, these.

[0118] Summary of information obtained from the software The summary report generated by the software is - Blood flow parameters (e.g., flow velocity, blood volume), - The number of white blood cells, - Subtypes of white blood cells (e.g., granulocytes, lymphocytes), - The type of movement of white blood cells (e.g., free flow, rotation, crawling, tethering) and movement parameters (e.g., speed, acceleration), - The morphology of blood vessels (endothelium, glycocalyx, etc.) - Heartbeat, and may include information of.

[0119] Examples The following examples are provided to demonstrate and further explain specific embodiments and aspects of the present invention and should not be construed as limiting the scope of the present invention. The descriptions provided in the examples are presented without being bound by theory.

[0120] Example 1 Summary of Example 1 A miniature oblique back-illumination microscope (mOBM) that enables real-time label-free phase-contrast imaging of rolling and adhesion of leukocytes, which are the initial stages of leukocyte mobilization characteristic of inflammation, in the microcirculation of the human oral mucosa. By imaging cell movement, new diagnostic information (such as changes over time in the progression of the disease, response to treatment, etc.) that cannot be obtained from conventional static diagnostic parameters such as cell count and morphology can be obtained.

[0121] The ability of white blood cells to migrate to various organs and tissues is a fundamental requirement for proper immune function [Reference 1]. The mobilization of white blood cells is initiated by a decrease in the rate at which circulating white blood cells become immobilized and rotate on the endothelial surface, followed by a multi-step process in which the white blood cells firmly adhere and extravasate into the tissue [Reference 2]. Although well-characterized in animal models using intravital microscopy, rotational and adhesion events (collectively referred to as leukocyte-endothelial cell interactions, or LEI [References 3, 4]) have rarely been observed in humans [Reference 5]. Conceptually, since conventional histopathology relies on the static examination of biopsy samples, imaging of cell motility as a potential source of diagnostic information has not yet been explored clinically. LEI has been reported to be significantly increased in the sublingual microvasculature of patients with systemic inflammation such as sepsis [Reference 6] and ischemia-reperfusion injury [Reference 7]. However, due to the lack of appropriate detection and analysis tools, it has been difficult to evaluate LEI in the clinical setting. Individual white blood cells cannot be resolved even using existing clinical devices such as CytoCam and MicroScan [Reference 8], and instead, their presence is inferred from the gaps and voids in blood vessels filled with red blood cells. The suboptimal image quality is further exacerbated by artifacts from intense movement and pressure, as well as the lack of appropriate analysis tools to quantify white blood cell motility [Reference 9]. Reflectance confocal microscopy (RCM) provides high-resolution cell imaging and is effectively used in dermatology clinics [Reference 10]. However, RCM requires laser scanning with a limited frame rate [Reference 11]. Nonlinear optical techniques such as third-harmonic generation (THG) microscopy [Reference 12] and two-photon-induced UV autofluorescence imaging [Reference 13] are also capable of label-free imaging of white blood cells, but these techniques require complex laser systems and scanning platforms, and there is no clear path for their introduction into the clinical setting.

[0122] To address these limitations, we developed a miniature oblique back-illumination microscope (mOBM) for non-invasively imaging the microvasculature of the human oral mucosa (Figs. 24A, 24B, 25A, 25B, 25C). The mOBM (Fig. 25B) includes an aberration-corrected refractive index distribution type (GRIN) objective lens with a diameter of 1 mm and a numerical aperture (NA) of 0.75, and a green light-emitting diode (LED) light source coupled to a large-core (0.1 mm) multimode optical fiber. The output end of the fiber is arranged such that photons enter the tissue from one side of the GRIN lens (Fig. 25C). The injected photons undergo multiple scattering in the deep tissue layer, and only a portion of the photons are collected by the GRIN lens, and back-illumination is performed at an oblique angle due to the offset shape of the optical fiber [Reference 14]. Asymmetric (oblique) back-illumination generates a phase-gradient contrast (PGC) image similar to a differential interference contrast (DIC) [Reference 15] or differential phase contrast (DPC) [Reference 16] microscope image, but there are important differences. Unlike DIC and DPC, OBM is designed to operate in thick, undamaged tissue [Reference 14] and is thus compatible with in vivo imaging. The use of a green LED further provides contrast based on the absorption of hemoglobin, with white blood cells standing out against dark red blood cells (Fig. 27, panel a, left).

[0123] To minimize artifacts due to movement and pressure, a custom tissue stabilizer (Figure 25A) was developed to gently expose the microvessels (Figure 24B) inside the lower lip and maintain their position during imaging. The selection of the inner lip is based on easy access, well-developed vascular beds, absence of skin pigmentation, absence of a highly scattering stratum corneum that degrades skin image quality, and subject comfort. We imaged the movement of leukocytes in the mucosal microvessels of subjects to verify the performance of the system. When imaging at a frame rate of 200 Hz, individual blood cells in the microcirculation are clearly depicted (Figure 27, panel a, left). These cells move at speeds of 1 mm / second or less, and when the exposure time for individual frames is set to less than 1 millisecond, their movement is effectively "frozen". Additionally, rotating cells are detected in a subset of blood vessels in healthy subjects (Figure 27, panel a, right). Judging from the diameter and flow rate, they are thought to be postcapillary venules.

[0124] We also imaged the inflamed sites caused by the presence of stomatitis (Figure 27, panels b - d) and confirmed that the average velocity of leukocytes was significantly reduced. Using leukocyte tracking for each automatic frame (see method below), the average rotational velocity in healthy tissue was found to be 58 ± 28 μm / s (Figure 27, panel e). This decreased to 4 ± 6 μm / s in inflamed tissue (Figure 27, panel e). Upon further examination, two populations were revealed: an adherent population with an average velocity of approximately 0.7 ± 0.7 μm and a slowly rotating population with an average velocity of 11 ± 6 μm (Figure 27, panel f). The adherent cells were limited to areas close to the stomatitis (Figure 27, panel d, dark arrowheads). On the other hand, the slowly moving leukocytes were detected in both the center and periphery of the inflamed area (Figure 27, panel d, light arrowheads). The displacement profile showed a characteristic stop - and - start movement pattern during inflammation (Figure 27, panel g).

[0125] To further miniaturize the device, the current focus adjustment mechanism (moving stage) can be replaced with an electrically adjustable lens [Reference 17]. Furthermore, a high-speed CMOS sensor (maximum frame rate of 1000 Hz) can also be replaced with a standard (30 Hz) video camera. Even in this case, it is possible to image rotating cells or adherent cells, but at this frame rate, it is not possible to resolve flowing cells. The high-speed imaging function gives rise to an attractive possibility of performing non-invasive white blood cell counting by resolving individual circulating cells with the help of machine learning and flagging white blood cells "on the fly". This is an actively researched theme in our laboratory and other laboratories [Reference 18]. We plan to install the mOBM device in several clinics due to its small size, low cost, and simple structure, and test its usefulness in the diagnosis of critically ill patients and premature infants at high risk of infection and sepsis. Furthermore, we envision that this device will be used even in resource-poor environments where there is a lack of expertise and infrastructure for collecting blood for standard laboratory analysis.

[0126] Method Features of the developed system. The optical design of the mOBM was performed using Zemax software (Figure 18A). It consists of a refractive index distribution type (GRIN) objective lens with a diameter of 1 mm, a relay optical system, and a CMOS camera. The imaging GRIN lens assembly (Grintech, GT-MO-080-032-ACR-VISNIR-08-20) is aberration-corrected and has a high numerical aperture (NA) of approximately 0.75 (in water). The relay optical system is composed of a GRIN lens (Edmund Optics, #64-519) with a pitch of approximately 1 / 4 and NA = 0.52, and an achromatic doublet lens (Edmund Optics, #49-772) with an effective focal length of 19.1 mm, which projects the enlarged image onto the CMOS sensor (Basler, daA1920-160um). Oblique illumination is provided by a 530 nm LED light source (Thorlabs, M530F2) coupled to a large-core optical fiber with a diameter of 1 mm, and its output (≒19 mW) end is placed on one side of the GRIN objective lens assembly (Figure 25C). The developed mOBM (Figure 25B) detects multiply scattered photons in a non-confocal manner and generates phase gradient contrast (PGC) images that reveal fine morphological details (such as cell membranes and granules). In addition to PGC, our device also benefits from absorption contrast generated using a wavelength of 530 nm, which is strongly absorbed by hemoglobin (inside red blood cells) rather than white blood cells. An important feature is the imaging chip (Figure 25C) composed of the following four channels: 1. A central channel containing a small GRIN objective lens. 2. An illumination channel containing a large-core multimode fiber. 3. A perfusion channel for maintaining the wet state of the oral mucosa tissue and maintaining the immersion medium in front of the objective lens. 4. A ring-shaped vacuum cavity surrounding the central channel to stabilize the imaging area. The magnification of our device is approximately 25, and as a result, ratios of approximately 7.2 pixels / μm (without binning) and approximately 3.6 pixels / μm (with 2×2 binning) were obtained. The field of view of our system is approximately 150×200 μm, and the working distance of the objective lens was adjusted by the motorized movement of the camera in the optical axis direction (20~100 μm). SolidWorks software was used for the mechanical housing, imaging chip, and design of the oral mucosa device of the mOBM (Figure 25A).The parts were printed using a 3D laser printer (Formlabs 3B). All parts that come into contact with human tissue were printed using a sterilizable biocompatible material (Formlabs, RS-F2-BMCL-01, RS-F2-BMBL-01).

[0127] Tissue stabilizer. The developed universal oral mucosa device (Figure 25A) was used to gently hold the lower lip tissue of the subject and expose the microvessels of the cheek without causing unwanted blood flow changes (Figure 24B). This device stabilized the tissue during the imaging session (about 20 - 60 minutes). An electric XYZ stage was used to position the imaging chip of the mOBM and identify the location of the target microvessels.

[0128] Imaging of healthy subjects. The subject first sits in front of the imaging system and gently fixes the head using a soft tissue strap. Next, the developed oral mucosa device is used to position the lower lip (Figure 25A). The subject had no additional requirements other than remaining seated during imaging. The imaging session is initiated by an operator who positions the imaging chip over the region of interest (ROI) using a motorized XYZ actuator with micrometer resolution. The imaging session usually lasts for 20 - 60 minutes.

[0129] Image acquisition parameters. For circulating white blood cells (Figure 27, panel a, left), the acquisition frame rate was fixed at 200 fps and the exposure time varied between 0.5 - 1 ms. When necessary, 2×2 binning was used to increase the signal-to-noise ratio. For rotating white blood cells (Figure 27, panel a (center), and Figure 27, panel b), the acquisition frame rate was fixed at 200 fps and the exposure time varied between 2.5 - 4 ms.

[0130] Image processing pipeline. Image processing was performed using ImageJ (open source software). Image processing pipeline for rotating and adherent white blood cells: 1. Registration (plugin: template matching). 2. Cropping of region of interest (ROI). 3. Smoothing of blood flow (plugin: Kalman filter). 4. Extraction of PGC (image - Gaussian blurred image (sigma radius: 20 - 40)). 5. Mean subtraction (PGC stack - mean image). 6. Tracking of white blood cells (plugin: TrackMate). White blood cells detached from the endothelial wall after a short contact were excluded from the analysis. To perform consistent LEI measurements, the rotational cell movement in straight and low - curvature blood vessels with a laminar blood flow profile was analyzed. Blood vessels with large curvatures and bifurcations that tend to have a turbulent (and thus unpredictable) blood flow profile were excluded from the analysis. Image processing pipeline for circulating white blood cells: 1. Registration (plugin: template matching). 2. Cropping of ROI. 3. Extraction of PGC (image - Gaussian blurred image (sigma radius: 20 - 40)).

[0131] Statistical analysis. In the statistical analysis (Figure 27, panel e), an unpaired t - test (also called Student's t - test) was used. For the rotating cell group (Figure 27, panel e, left column), n = 18, and for the inflammation group (Figure 27, panel e, right column), n = 39. P value < 0.0001.

[0132] Example 2 Referring to FIG. 33, an example of a process 3300 is shown that measures and reports perfusion parameters and white blood cell counts within a video of microvessels acquired by a non - invasive imaging system 3301. This system (non - invasive imaging system 3301) may be one of the systems described herein. In some embodiments, the process 3300 is implemented as computer - readable instructions on at least one memory and executed by at least one processor connected to the at least one memory.

[0133] The process 3300 has the following seven modules (M01 - M07). M01. Execution of motion artifact removal and flat field correction M02. Selection of a region of interest (ROI) suitable for evaluating the characteristics of microvascular contents M03. Definition of the coordinate system for the transformed ROI M04. Quantification of perfusion parameters within the ROI M05. Identification of the positions of candidate white blood cells and provision of an estimated value of their number M06. Detection of white blood cells and classification of subtypes by deep learning M07. Establishment of the final number of white blood cells per flow rate and reporting of the results

[0134] In process step 3301 that starts process 3300, a video of the microvessels is acquired by the non-invasive imaging system as described above. In process step 3302, as the output of 3301, a video of one or more blood vessels is acquired, and white blood cells show a difference in intensity compared to red blood cells.

[0135] M01. Execution of motion artifact removal and flat field correction In module 3303, image registration is performed to remove the motion artifacts generated in the video due to patient movement (such as breathing). The non-uniformity of illumination is also corrected by background subtraction of the video. In process step 3304, the output of module 3303 is a video without motion artifacts and with uniform illumination.

[0136] M02. Selection of a region of interest (ROI) suitable for evaluating the characteristics of microvascular contents Module 3305 specifies the volume for microvascular content analysis. The vessel boundaries are first estimated using the standard deviation time projection method (see Figure 34, panel b). In this method, pixels with constant intensity (tissue) over time and pixels that vary (vessels) are effectively separated. The user checks the separation (see Figure 34, panel c), modifies the resulting vessel boundaries as needed (e.g., to indicate overpasses or underpasses, see Figure 34, panel d), and finally draws a polygon (see Figure 34, panel f). The ROI is specified by pixels both within the vessel area (shaded) and within the polygon drawn by the user (bright color). In process step 3306, which is the output of module 3305, the video is cropped to the boundary volume of the ROI.

[0137] Definition of the transformed coordinate system for M03.ROI Module 3307 defines the transformed coordinate system of the ROI. The Skeletal Coordinate System (SCS) is an example of a transformed coordinate system. Its two axes x’ and y’ are parallel and perpendicular, respectively, to the direction of blood flow (see Figure 35, panel a).

[0138] The x’ axis is defined using skeletonization. The resulting skeleton coexists with the axis of the vessel and mimics its curvature (see Figure 35, panel b). The y’ grid lines running parallel to the x’ axis are created by replicating this skeleton to cover the full width of the vessel. The replication is performed by image dilation and edge detection using structuring elements of increasing size in steps (see Figure 35, panels c - d).

[0139] The y’ axis and the x’ grid lines of the SCS are defined by the normal to each pixel of the skeleton. These normals cross the skeleton and all its replications (see Figure 35, panels e - f).

[0140] As a transformed coordinate system, SCS is versatile for various microvascular ROI morphologies (see Fig. 35, panels g - i).

[0141] In process step 3308, the output of module 3307 is the transformed coordinate system. In process step 3309, another output of module 3307 is a set of vascular morphological parameters calculated by the module, such as the width, cross - sectional area, and curvature of the blood vessels.

[0142] M04. Quantification of perfusion parameters within the microvascular ROI In module 3310, the perfusion parameters of the microvascular ROI are measured for each measurement unit along the x’, y’, and t axes. These units are the blood vessel block, the y’ grid line index, and the time segment respectively. Each measurement unit has a corresponding spatio - temporal diagram (see Fig. 36, panels a - b).

[0143] For velocity measurement, only the y’ grid lines near the ROI axis are used. The spatio - temporal diagram is filtered in the Fourier space, suppressing the noise band and large white blood cells that may slow down blood flow (see Fig. 36, panel c). Next, the spatio - temporal diagram is returned to the image space and segmented (see Fig. 36, panels d - e). After removing lines with insufficient length, the angle of each inclined line is measured. These angle values are averaged, weighted by the aspect ratio, and converted into one flow velocity value of the spatio - temporal diagram.

[0144] The integration of velocity values from multiple spatio - temporal diagrams is performed according to the following rules. If multiple y’ grid lines are included in the measurement at a specific time point, their average velocity values are used. If time segments overlap, the average velocity values of the overlapping time segments are used for each reported time point. The result is a plot of the flow velocity as a function of time for each blood vessel block (see Fig. 36, panel f).

[0145] The blood flow rate (quantity / time) is calculated by multiplying the blood flow velocity by the cross-sectional area 3309 of the ROI. The blood flow volume is calculated by multiplying the blood flow rate by the duration of the video.

[0146] The process step 3311, which is the output of the module 3310, includes a plot of the velocity as a function of time for each blood vessel block and the blood flow volume.

[0147] M05. Identification of the positions of candidate white blood cells and provision of an estimated value of their number The module 3312 identifies candidate white blood cells and the approximate temporal and spatial coordinates of their appearance in the video.

[0148] The spatio-temporal diagram is used to identify the positions of candidate white blood cells (see Figure 37, panel a). Summing along the bands of these diagrams creates a one-dimensional vector in which the high-intensity peaks suggest the presence of these cells. In this example, a series of moving sums are performed using a sliding window width ranging from the minimum white blood cell diameter to the full width of the spatio-temporal diagram. The largest result among these sums is selected as the value of the output intensity vector (see Figure 37, panel b).

[0149] The integration of the intensity vectors collected from all spatio-temporal diagrams is performed as follows.

[0150] i. Integration along y' (by the y' grid line). The intensity vectors from different y' grid lines are first combined using a pixel-based sorting method. For each pixel along the y' grid line, the values from all y' grid lines are sorted in descending order. The result is a set of vectors, where the first vector contains the maximum value for each pixel of all y' grid lines, and the second vector contains the second-largest value for each pixel of all y' grid lines. The high-intensity peaks that appear on some y' grid lines indicate that there may be small white blood cells present and are expected to be captured in the first n s vectors. Here, n sis the number of y' grid lines spanning the minimum white blood cell diameter. These n s Vectors are averaged per pixel to create a new y' integrated intensity vector (see Figure 37, panel c).

[0151] ii. Integration along t (per time segment). The y' integration vectors are connected such that the output y't integration vector covers the entire time length of the video. Each blood vessel block has one such output (see Figure 37, panel d).

[0152] There is a time delay between the y't integration vectors of consecutive blood vessel blocks, which indicates the time required for cells to move downstream. The time delay is not consistent for each peak (see Figure 37, panel e).

[0153] iii. Integration along x' (by blood vessel block). Peak matching finds the peaks of each y't integration vector that are likely to belong to the same cell. The dynamic time warping (DTW) method is a peak matching algorithm that aligns time series to a similar set of peaks, but the individual peaks within the set do not need to share the same offset value. After the peak positions match, the y't integration intensity vectors of different blood vessel blocks are summed to complete the integration process by creating a y'tx' integration intensity vector that contains peak information from all y' grid lines, all time segments, and all blood vessel blocks (see Figure 37, panel f).

[0154] Gating is performed on the y'tx' integration intensity vector. A threshold is applied, and peaks above the threshold are characterized in terms of their height, and further, their position and width (i.e., the appearance time in the video). Using the position and width values of each peak, and the position shift between blood vessel blocks, the "gate" for the peak is defined (see Figure 37, panels f - h). This is defined as the approximate spatial and temporal volume in which candidate white blood cells appear.

[0155] All peaks, i.e., the volumes of all candidate white blood cells, are transformed from the transformed coordinate system to the orthogonal coordinate system of the video (see Figure 37, panel j). Process step 3314, which is the output of module 3312, are these volumes, each identified by a "gate index" corresponding to a candidate white blood cell. Another output of module 3312, process step 3313, is the number of such volumes, i.e., the "gate", which is also an estimated value of the candidate white blood cell count.

[0156] M06. Detection of White Blood Cells and Classification of Subtypes by Deep Learning Modules M06 3315 and M07 3317 confirm the presence of candidate white blood cells identified by module 3312 by the "detection and linking" method of particle tracking and ensure the accuracy of the final white blood cell count. This module 3315 is related to "detection".

[0157] Module 3315 first prepares an input image suitable for the deep learning model. Video 3302 is processed by two different background removal methods, the inputs and two outputs are merged into a 3-channel video, and cropped according to the gated volumes of candidate white blood cell appearance 3314. Other requirements of the deep learning model, such as tiling of the cropped images, are also performed.

[0158] The input image is then input into a deep learning model pre-trained with a white blood cell detection function. Process step 3316, which is the output of module 3315, includes the locations where white blood cells are detected in the video and the confidence scores of the detections. Depending on the model, it may also be possible to classify white blood cells according to subtypes (granulocytes, lymphocytes, etc.).

[0159] M07. Establishment of the Final White Blood Cell Count per Flow Rate, Reporting of Results Modules 3315 and 3317 confirm the presence of candidate white blood cells identified by module 3312 by means of the "detection and linking" method of particle tracking, ensuring the accuracy of the final white blood cell count. Module 3317 is related to "linking".

[0160] State estimation addresses scenarios where white blood cells are intermittently obscured by other cells within the video. The number of particle trajectories is reported as the final white blood cell count in step 3318, and dividing this number by the blood flow rate 3311 gives the final white blood cell count per blood flow rate. With this report, process 3300 ends in step 3319.

[0161] References 1. Germain, R.N., Robey, E.A. & Cahalan, M.D., "A Decade of Imaging Cellular Motility and Interaction Dynamics in the Immune System". Science 336, 1676 - 1681 (2012). 2. Vestweber, D., "How leukocytes cross the vascular endothelium". Nat. Rev. Immunol. 15, 692 - 704 (2015). 3. Springer, T.A., "Traffic signals for lymphocyte recirculation and leukocyte emigration: a multistep paradigm". Cell 76, 301 - 314 (1994). 4. von Andrian, U.H. et al., "Two-step model of leukocyte-endothelial cell interaction in inflammation: distinct roles for LECAM-1 and the leukocyte beta 2 integrins in vivo". Proc. Natl. Acad. Sci. 88, 7538-7542 (1991). 5. Sahu, A. et al., "In vivo quantification of rolling and adhered leukocytes in human sepsis". Nat. Commun. 13, 5312 (2022). 6. Fabian-Jessing, B.K. et al., "In vivo quantification of rolling and adhered leukocytes in human sepsis". Crit. Care 22, 240 (2018). 7. Uz, Z., Ince, C., Shen, L., Ergin, B. & van Gulik, T.M., "Real-time observation of microcirculatory leukocytes in patients undergoing major liver resection". Sci. Rep. 11, 4563 (2021). 8. Massey, M.J. & Shapiro, N.I., "A guide to human in vivo microcirculatory flow image analysis". Crit. Care 20, 35 (2016). 9. Ince, C. et al., "Second consensus on the assessment of sublingual microcirculation in critically ill patients: results from a task force of the European Society of Intensive Care Medicine". Intensive Care Med. 44, 281 - 299 (2018). 10. Ulrich, M., Lange - Asschenfeldt, S. & Gonzalez, S., "The use of reflectance confocal microscopy for monitoring response to therapy of skin malignancies". Dermatol. Pract. Concept. 2, 0202a10 (2012). 11. Rajadhyaksha, M., Gonzalez, S., Zavislan, J.M., Rox Anderson, R. & Webb, R.H., "In Vivo Confocal Scanning Laser Microscopy of Human Skin II: Advances in Instrumentation and Comparison With Histology". 11 The authors declare no competing interests. J. Invest. Dermatol. 113, 293 - 303 (1999). 12. Wu, C.-H. et al., "Imaging Cytometry of Human Leukocytes with Third Harmonic Generation Microscopy". Sci. Rep. 6, 37210 (2016). 13. Li, C. et al., "Imaging leukocyte trafficking in vivo with two - photon - excited endogenous tryptophan fluorescence". Opt. Express 18, 988 - 999 (2010). 14. Ford, T.N., Chu, K.K. & Mertz, J., "Phase - gradient microscopy in thick tissue with oblique back - illumination". Nat. Methods 9, 1195 - 1197 (2012). 15. Nomarski G., "Differential microinterferometer with polarized waves". J Phys Radium Paris 16, 9S (1955). 16. Hamilton, D.K. & Sheppard, C.J.R., "Differential phase contrast in scanning optical microscopy". J. Microsc. 133, 27 - 39 (1984). 17. Bagramyan, A. et al., "Focus - tunable microscope for imaging small neuronal processes in freely moving animals". Photonics Res. 9, 1300 - 1309 (2021). 18. Huang, L., McKay, G.N. & Durr, N.J., "A Deep Learning Bidirectional Temporal Tracking Algorithm for Automated Blood Cell Counting from Non - invasive Capillaroscopy Videos". Medical Image Computing and Computer Assisted Intervention - MICCAI 2021 (eds. de Bruijne, M. et al.) 415 - 424 (Springer International Publishing, 2021). doi:10.1007 / 978 - 3 - 030 - 87237 - 3_40. No citation of any document or reference should be construed as an admission that it is prior art with respect to the present invention.

[0162] Accordingly, the present invention provides a system and method for imaging microvessels of a subject's tissue, and more specifically, a system and method for non - invasively and stably imaging microvessels of the oral mucosa.

[0163] In light of the principles and exemplary embodiments described and illustrated herein, it will be recognized that the exemplary embodiments can be varied in arrangement and detail without departing from such principles. Also, while the foregoing description has focused on specific embodiments, other configurations are contemplated. In particular, expressions such as "in one embodiment", "in another embodiment", "in other embodiments", "in some embodiments", etc. are used herein, but these phrases are generally intended to refer to the possibility of embodiments and are not intended to limit the present invention to the configuration of a specific embodiment. As used herein, these terms can refer to the same or different embodiments that can be combined with other embodiments. In principle, any of the embodiments referred to herein can be freely combined with one or more of the other embodiments referred to herein, and any number of functions of different embodiments can be combined with each other.

[0164] The present invention has been described in considerable detail with reference to specific embodiments, but those skilled in the art will understand that the present invention can be used in other embodiments different from the described embodiments, which are presented for purposes of illustration and not limitation. Accordingly, the appended claims should not be limited to the description of the embodiments contained herein.

Claims

1. A system for imaging the microvessels of a subject's tissue, comprising: (a) a tissue stabilizer configured to contact the tissue of the subject and maintain the position of the region of the microvessels to be imaged; (b) an imaging device; wherein the imaging device comprises (i) a housing having an imaging section; (ii) an illumination device disposed within the imaging section of the housing, the illumination device having a light output end for irradiating the region of the microvessels with light, wherein the light output end is offset with respect to the optical axis of the imaging section; (iii) an objective lens disposed within the imaging section of the housing, the objective lens receiving at least a portion of the light scattered by the region of the microvessels; (iv) an image detector disposed within the imaging section of the housing, the image detector receiving the light redirected by the objective lens and detecting a microscopic image of the region of the microvessels; A system.

2. The tissue stabilizer comprises a base, a slide mechanism attached to the base, and an adapter for contacting the tissue, the adapter being attached to the slide mechanism, wherein the adapter is movable toward and away from the base. The system according to claim 1.

3. The base comprises a jaw holder, the tissue stabilizer further comprises a frame, and the jaw holder and the forehead holder are attached to the frame. The system according to claim 2.

4. The adapter includes a patterned surface finish for contacting the tissue. The system according to claim 2.

5. The adapter includes opposing tabs for contacting the tissue. The system according to claim 2.

6. The adapter applies mechanical pressure to the edge of the oral mucosa tissue at least 2 millimeters away from the imaging target area, minimizing the influence of the mechanical pressure on the imaging area. The system according to claim 2.

7. The adapter is bendable. The system according to claim 2.

8. The adapter is rigid. The system according to claim 2.

9. The adapter includes one or more light sources. The system according to claim 2.

10. The adapter is the system according to claim 2, including a vacuum system.

12. The adapter is the system according to claim 5, further including a transparent sheet attached between the opposing tabs.

13. The tissue stabilizer includes a base, a slide mechanism attached to the base, and an adapter for contacting the tissue, and the adapter is attached to the slide mechanism. The adapter is movable laterally with respect to the base. The system according to claim 1.

14. The adapter is sized to contact the oral mucosa of the subject. The system according to claim 1.

15. The adapter is sized to contact the lips of the subject. The system according to claim 1.

16. The adapter includes a rod attached between opposing connectors, and the rod is sized to contact the tissue. The system according to claim 2.

17. The adapter includes a flexible loop, and the rod is sized to contact the tissue. The system according to claim 1.

18. The objective lens is a microlens. The system according to claim 1.

19. The objective lens is a gradient index (GRIN) objective lens. The system according to claim 1.

20. The objective lens is a gradient index (GRIN) objective lens, and the system further includes a doublet achromat lens. The system according to claim 1.

21. The image detector is a camera. The system according to claim 1.

22. The image detector is a CMOS sensor. The system according to claim 1.

23. The image detector is movable with respect to the objective lens. The system according to claim 1.

24. The image detector detects a microscopic image using Oblique Back-illumination Microscopy (OBM). The system according to claim 1.

25. ​ The image detector detects microscopic images using Offset Trans-illumination Microscopy (OTM). The system according to claim 1.

26. The imaging unit further includes a vacuum device for stabilizing the tissue to be imaged. The system according to claim 1.

27. The imaging unit further includes a perfusion channel for supplying fluid to keep the tissue to be imaged in a wet state. The system according to claim 1.

28. The lighting device includes a light source and an optical fiber having the light output end. The system according to claim 1.

29. The imaging unit further includes a sterile disposable part. The system according to claim 1.

30. The sterile disposable part includes single or multiple optical elements, a light source, a perfusion channel, and a vacuum cavity. The system according to claim 29.

31. The imaging unit further includes an imaging chip including an objective lens, a vacuum device, a perfusion channel, and an illumination fiber of the lighting device. The objective lens is a microlens. The system according to claim 1.

32. The imaging unit further includes an imaging chip including an objective lens, a vacuum device, a perfusion channel, and an illumination fiber of the lighting device. The objective lens is a gradient index (GRIN) objective lens. The system according to claim 1.

33. The imaging chip is disposable. The system according to claim 32.

34. The microscopic image includes an image of the interaction between leukocytes and endothelial cells in the microvessels. The system according to claim 1.

35. The imaging is label-free imaging. The system according to claim 34.

36. The microscopic image is a phase gradient contrast image. The system according to claim 1.

37. The lighting device includes a light source and an optical fiber having the light output end. The light source includes a light-emitting diode. The system according to claim 1.

38. The imaging is performed at a frame rate of 1 Hz to 1000 Hz. The system according to claim 1.

39. The imaging is performed at a frame rate of 1 Hz to 300 Hz. The system according to claim 1.

40. The optical power to be injected is automatically adjusted by a controller to prevent saturation of the pixels of the data acquisition element. The system according to claim 1.

41. The scattered light collection time of the data acquisition element is automatically adjusted by software to prevent saturation of the pixels. The system according to claim 1.

42. The microscopic image includes an image of white blood cells in the microvessels. The system further includes a controller that communicates electrically with the illumination device and the image detector. The controller executes a program stored in the controller to (i) receive the microscopic image from the image detector, (ii) calculate the average rotational speed of the white blood cells in the microvessels using automated frame-by-frame white blood cell tracking. It is configured as follows. The system according to claim 1.

43. The controller executes a program stored in the controller to (iii) compare the average rotational speed of the white blood cells in the microvessels with the average rotational speed of white blood cells in healthy tissue. The system according to claim 41.

44. A system for imaging microvessels in a subject's tissue, comprising an imaging device including a housing having an imaging unit, a tissue stabilizer configured to contact the tissue of the subject and maintain the position of the region of the microvessels imaged by the imaging device, an illumination device disposed within the tissue stabilizer and having a light output end for irradiating the region of the microvessels with light, an objective lens disposed within the imaging unit of the housing, the objective lens receiving at least a portion of the light scattered by the region of the microvessels, an image detector disposed within the imaging unit of the housing, the image detector receiving the light redirected by the objective lens and detecting a microscopic image of the region of the microvessels. A system comprising.

45. The tissue stabilizer includes a base, a slide mechanism attached to the base, and an adapter for contacting the tissue, the adapter being attached to the slide mechanism. The adapter is movable toward and away from the base. The system according to claim 44.

46. The light output end of the illumination device is disposed within the adapter. The system according to claim 45.

47. The base includes a jaw holder, The light output end of the lighting device is disposed within the jaw holder, The system according to claim 45.

48. A system for imaging microvessels in a subject's tissue, A tissue stabilizer configured to contact the tissue of the subject and maintain the position of the region of the microvessels to be imaged, An illumination device disposed within the tissue stabilizer and having a light output end for irradiating the region of the microvessels with light, An objective lens disposed within the tissue stabilizer, the objective lens receiving at least a portion of the light scattered by the region of the microvessels, An image detector disposed within the tissue stabilizer, the image detector receiving the light redirected by the objective lens and detecting a microscopic image of the region of the microvessels, A system comprising.

49. The tissue stabilizer includes a first arm and an opposing second arm, the first arm and the second arm defining a space therebetween for receiving the tissue, The illumination device, the objective lens, and the image detector are disposed on the first arm such that the image detector detects a microscopic image using oblique back-illumination microscopy (OBM), The system according to claim 48.

50. The tissue stabilizer includes a first arm and an opposing second arm, the first arm and the second arm defining a space therebetween for receiving the tissue, The objective lens and the image detector are disposed on the first arm, the illumination device is disposed on the second arm, and the image detector detects a microscopic image using offset trans-illumination microscopy (OTM), The system according to claim 48.

51. The tissue stabilizer includes a first arm, an opposing second arm, and a hinge connecting the first arm and the second arm, a variable-size space for receiving the tissue being formed between the first arm and the second arm, The system according to claim 48.

52. A system for imaging microvessels in a subject's tissue, An imaging device capable of capturing an image, An electronic processor that communicates with the imaging device, comprising, by executing a program stored in the electronic processor, receiving the image from the imaging device, reducing a foreground of the image to a skeleton that captures one or more attributes of the foreground including at least one of curvature, connectivity, and extent, performing, the skeleton defines a transformed coordinate system for quantifying one or more perfusion parameters in the microvessels, a system.

53. The system according to claim 52, wherein the skeleton is a line that follows the axis of the blood vessel of the microvessel and bends according to the local curvature of the blood vessel. The system according to claim 52.

54. by executing the program stored in the electronic processor, the electronic processor creates a transformed coordinate system by generating a plurality of grid lines that cover the entire width of a region of interest (ROI) of the microvessels, performing, The system according to claim 52.

55. by executing the program stored in the electronic processor, the electronic processor creates the transformed coordinate system such that two axes are parallel and perpendicular to the blood flow respectively, performing, the axis parallel to the blood flow is defined by the skeleton, and the axis perpendicular to the blood flow is defined by the normal of the skeleton, The system according to claim 54.

56. by executing the program stored in the electronic processor, the electronic processor creates a collection of skeletons and lines created with reference to the skeleton that defines the x'-axis and y'-grid lines of the transformed coordinate system, performing, the y'-grid lines run in the direction of the blood flow in the microvessel ROI, The system according to claim 54.

57. The system according to claim 56, wherein the y'-grid lines of the transformed coordinate system have the same pixel length regardless of the curvature of the ROI. The system according to claim 56.

58. by executing the program stored in the electronic processor, the electronic processor draws a spatio-temporal diagram of each time segment and each y'-grid line of the blood vessel block, performing, the blood vessel block is defined as a unit of length along the axis of the microvessel ROI, The system according to claim 54.

59. The electronic processor, by executing the program stored in the electronic processor, integrates a plurality of spatio-temporal diagrams of individual y' grid lines, time segments, and blood vessel blocks to calculate a blood flow velocity and performs the following steps: The system according to claim 58.

60. The electronic processor, by executing the program stored in the electronic processor, calculates a blood flow rate by multiplying the blood flow velocity by the cross-sectional area of the ROI and performs the following steps: The system according to claim 59.

61. The electronic processor, by executing the program stored in the electronic processor, calculates the total number of white blood cells along the slope of the spatio-temporal diagram and generates an intensity profile and performs the following steps: The intensity profile is further integrated from a plurality of y' grid lines, time segments, and blood vessel blocks, and an estimated value of the number of white blood cells is obtained from the number of peaks of the integrated intensity profile. The system according to claim 58.

62. The integration is performed by a dynamic time warping method to match the peaks of the intensity profiles of the blood vessel blocks while allowing for variations in the time delay between candidate white blood cells. The system according to claim 59.

63. The electronic processor, by executing the program stored in the electronic processor, estimates the appearance time of candidate white blood cells by determining the peak positions of the intensity profiles and performs the following steps: The system according to claim 52.

64. The electronic processor, by executing the program stored in the electronic processor, uses the integrated intensity profile to gate the approximate space and time of the appearance of candidate white blood cells in the video and performs the following steps: The system according to claim 52.

65. A system for imaging the microvessels of a subject's tissue, comprising: an imaging device capable of capturing an image; and an electronic processor communicating with the imaging device, wherein the electronic processor, by executing a program stored in the electronic processor, receives the image from the imaging device; and accesses a deep learning model trained with learning data to detect perfusion and white blood cell feature data from the image input. ​ Applying the image to the machine learning model to quantify one or more perfusion parameters in the microvessels; A system that performs the above steps. **Claim 66** The deep learning model is a neural network. The system according to claim 65. **Claim 67** The neural network is a convolutional neural network. The system according to claim 66. **Claim 68** The machine learning model is applied to a gated spatial region and time including candidate white blood cells. The system according to claim 65. **Claim 69** By executing the program stored in the electronic processor, the electronic processor Detects the coordinates of the image where white blood cells are detected and the probability score of the detection. A system that performs the above steps The system according to claim 65. **Claim 70** A method for in vivo flow cytometry of a biological fluid of a subject, comprising: (a) contacting the tissue of the subject with a tissue stabilizer to maintain the position of the biological structure of the subject; (b) using an illumination device to provide light to a part of the region of the biological structure and continuously illuminating the region of the biological structure; (c) using an image detector to continuously detect a microscopic image from the region of the biological structure based on the light scattered by the biological structure of the subject, wherein the illumination is performed at an oblique angle by an offset shape of the illumination device; (d) analyzing the microscopic image to identify the characteristics of the biological fluid within the biological structure. A method comprising the above steps. **Claim 71** Detecting the microscopic image in step (c) includes generating an optical image by oblique back-illumination microscopy (OBM). The method according to claim 70. **Claim 72** Detecting the microscopic image in step (c) includes generating an optical image by offset trans-illumination microscopy (OTM). The method according to claim 70. **Claim 73** The biological structure is a microvessel of the subject. Step (d) includes quantifying one or more perfusion parameters in the microvessel. The method according to claim 70. **Claim 74** The biological structure is a microvessel of the subject, Step (d) includes quantifying the number of white blood cells in the microvessel, The method according to claim 70.

75. Step (c) includes detecting the microscopic image without a label, The method according to claim 70.

76. Step (c) includes detecting the microscopic image at a frame rate of 1 Hz to 1000 Hz, The method according to claim 70.

77. The biological structure is a microvessel of the subject, Step (d) includes calculating the average rotational speed of white blood cells in the microvessel using automated frame-by-frame white blood cell tracking, The method according to claim 70.

78. The biological structure is a microvessel of the subject, Step (d) includes reducing the foreground of each microscopic image to a skeleton that captures one or more attributes of the foreground including at least one of curvature, connectivity, and extent, The skeleton defines a transformed coordinate system for quantifying one or more perfusion parameters in the microvessel, The method according to claim 70.

79. The biological structure is a microvessel of the subject, Step (d) further includes creating a transformed coordinate system by generating a plurality of grid lines that cover the entire width of the region of interest (ROI) of the microvessel, The method according to claim 70.

80. The biological structure is a microvessel of the subject, Step (d) further includes creating the transformed coordinate system such that two axes are parallel and perpendicular to the blood flow respectively, wherein the axis parallel to the blood flow is defined by the skeleton and the axis perpendicular to the blood flow is defined by the normal of the skeleton, The method according to claim 79.

81. The biological structure is a microvessel of the subject, Step (d) includes creating a transformed coordinate system such that two axes are parallel and perpendicular to the blood flow respectively, The method according to claim 70.

82. Step (d) includes creating the transformed coordinate system wherein the axis parallel to the blood flow is defined by the skeleton and the axis perpendicular to the blood flow is defined by the normal of the skeleton, The method according to claim 81.

83. The biological structure is the microvessel of the subject, Step (d) includes drawing a spatio-temporal diagram of each time segment and each y'-grid line of the vascular block, where the vascular block is defined as a unit of length along the axis of the microvascular ROI. The method according to claim 79.

84. Step (d) includes integrating a plurality of spatio-temporal diagrams of individual y'-grid lines, time segments, and vascular blocks to calculate a blood flow velocity. The method according to claim 83.

85. The biological structure is the microvessel of the subject, Step (d) includes accessing a deep learning model trained with training data to detect perfusion and leukocyte feature data from the image input, and applying the image to the machine learning model to quantify one or more perfusion parameters in the microvessel. The method according to claim 77.

Images