CN116157829A

CN116157829A - Methods and systems for cloud-based automatic quantitative assessment of retinal microvasculature using optical coherence tomography angiography images

Info

Publication number: CN116157829A
Application number: CN202180056858.8A
Authority: CN
Inventors: 赵健; 陆秉文; 李一鸣; 沈伯松
Original assignee: University of Hong Kong HKU
Current assignee: University of Hong Kong HKU
Priority date: 2020-08-17
Filing date: 2021-08-16
Publication date: 2023-05-23
Also published as: WO2022037517A1

Abstract

Systems and methods for automatic quantitative assessment of retinal microvasculature and quantification of microvasculature properties are provided for diagnosis and treatment of retinal and extra-retinal diseases.

Description

Methods and systems for cloud-based automatic quantitative assessment of retinal microvasculature using optical coherence tomography angiography images

Technical Field

The present invention relates to methods and systems for cloud-based automatic quantitative assessment of retinal microvasculature using optical coherence tomography angiography images.

Background

Retinal diseases in humans can be manifestations of different physiological or pathological conditions, including diabetes resulting in Diabetic Retinopathy (DR), retinal Vein Occlusion (RVO), age-related macular degeneration (AMD), and the like. Recently, extended life and "resting", pressure-filled lifestyles have led to a rapid increase in the number of patients suffering from these vision-threatening conditions. There is an urgent need for large scale improvements in the manner in which these diseases can be screened, diagnosed and treated.

Retinal imaging is the only method to directly examine systemic blood vessels and their pathological changes. Retinal imaging can reflect not only retinal vascular disease, but also indicate the risk of systemic disease (including cardiovascular disorders and hypertension). Most retinal diseases cause abnormalities in the bed of microvasculature in the eye, including DR, RVO and AMD. In addition, retinal properties provide important information about the risk of developing vascular disorders. Systemic diseases such as cardiovascular disease, hypertension and atherosclerosis manifest themselves as retinal microvascular changes. Thus, measurement of retinal microvasculature plays a key role in the management of vascular disorders.

Optical Coherence Tomography Angiography (OCTA) has recently been introduced for imaging of microvascular networks in the human eye. Compared to traditional dye-based angiography, OCTA provides a safer, faster, more cost-effective, non-invasive method to diagnose and monitor retinal vascular abnormalities. OCTA has the ability to display superficial, deep and avascular retinal patterns and the microvasculature of the choriocapillaris and choroidal patterns.

Typically, the patient regularly visits an ophthalmologist or other eye care professional to make OCTA measurements. Professionals can examine images and intelligently compare those images to other retinas they have observed in the past. Such experience-based inspection, while effective in detecting gross abnormalities and large changes, may not be effective in observing minor changes in the patient's retina. The assessment is also subjective and its accuracy is heavily dependent on the clinician's previous experience, making the diagnosis prone to human error. Underexperienced ophthalmologists often consult with external specialists for interpretation of laboratory results and medical images to improve their diagnostic accuracy. On the other hand, due to the limited descriptive capabilities of the built-in software of the OCTA machine, most patients are not able to contact well-known ophthalmic clinicians and are also not able to obtain enough information directly from their OCTA report. When the patient is thus concerned about their eye-related health, the patient may have to seek additional advice from external specialists, which may be time consuming and expensive. Thus, current retinal analysis systems are subjective, time consuming, error prone, and require extensive manual input from medical personnel. Furthermore, there is a complete quantitative OCTA image-based biomarker deficiency for reliable monitoring of vascular disease.

Advances in computer-aided image processing and analysis techniques are critical to making imaging-based disease diagnosis scalable, cost-effective, and reproducible. This advancement will lead directly to an effective patient classification, leading to early diagnosis, timely treatment and improved quality of life.

Disclosure of Invention

Systems and methods for automated quantitative assessment of retinal microvasculature and generation of retinal digital maps are provided for diagnosis and treatment of retinal and extra-retinal diseases. The systems and methods of the present invention provide a Computational Retinal Microvascular Biomarker (CRMB) that can be used for quantification of retinal microvascular disease states. Advantageously, the system and method of the present invention may be integrated in a cloud computing platform for cloud-based analysis that provides extensive accessibility to patients and healthcare providers and enables the use of machine learning to identify new CRMBs.

Drawings

FIG. 1 shows a schematic diagram of an exemplary architecture of a cloud platform for a computerized automated Optical Coherence Tomography Angiography (OCTA) image analysis system.

Fig. 2 shows a schematic diagram of a flow of a process for extracting a Computational Retinal Microvascular Biomarker (CRMB) from an OCTA image in accordance with embodiments of the present technique.

Fig. 3A shows representative OCTA raw images on both deep and superficial capillary plexuses. Fig. 3B shows representative OCTA scan images on both deep and superficial capillary plexuses.

Fig. 4 shows a graphical representation of an algorithm for extracting and generating a Digital Vasculature Map (DVM) from a calibrated OCTA image.

Fig. 5A shows a graphical representation of a procedure for applying an diabetic retinopathy Early Treatment (ETDRS) grid to a DVM. Fig. 5B shows segments defined by ETDRS grids in the left and right eyes.

Image a of fig. 6 shows a Digital Vascular Map (DVM) generated using an OCTA image. Image B of fig. 6 shows a thresholded image of the DVM image. Image C of fig. 6 shows a thresholded sub-foveal image of the DVM image. Image D of fig. 6 shows a ridge filtered image of the DVM image. Image E of fig. 6 shows a thresholded ridge filtered image of the DVM image. Image F of fig. 6 shows image E of fig. 6 with small vessel objects removed. Image G of fig. 6 shows image F of fig. 6 depicting retinal blood vessels. Image H of fig. 6 shows image F of fig. 6 with skeletonized blood vessels. Image I of fig. 6 shows image H of fig. 6 with a separate main vessel segment.

Fig. 7 shows a graphical representation of a process for calculating a vascular dispersion index for a DVM.

FIG. 8 shows a graphical representation of an algorithm for detecting endpoints and branch points in a skeletonized DVM.

Fig. 9A shows an image of a DVM in which calculated vascular dispersion indexes of individual blood vessels are marked. Fig. 9B shows an image of a DVM in which the calculated vessel diameter index of each depicted vessel is marked. Fig. 9C shows an image of a DVM with calculated vessel tortuosity indices of the separate vessel segments marked therein.

Image a of fig. 10 shows a Digital Vascular Map (DVM) of the retina. Image B of fig. 10 shows a masked and denoised DVM. Image C of fig. 10 shows thresholding and closing DVM. Image D of fig. 10 shows a bridging DVM. Image E of fig. 10 shows a DVM with foveal avascular zone polygon endpoints. Image F of fig. 10 shows a DVM in which the foveal avascular zone is calibrated.

Fig. 11 shows a visualization of the distribution and pairwise relationship between different CRMBs in a sample of healthy subjects and retinal vein occlusion patients.

Detailed Description

Systems and methods are provided for automatic assessment of retinal microvasculature architecture, quantification of retinal microvasculature changes associated with retinal or extra-retinal diseases, treatment of diseases, and quantitative assessment of treatment success based on reversal of disease-related changes in retinal microvasculature. Using the systems and methods of the present invention, methods are also provided for establishing comprehensive Computational Retinal Microvasculature Biomarkers (CRMBs) through a knowledge-driven computerized automated analysis system based on fractal analysis using OCTA images. The systems and methods of the present invention employ a cloud computing network with a distributed input device and a cloud-based analysis platform for automated analysis of OCTA images, quantification of eye disorders based on CRMBs established using the systems and methods of the present invention, and treatment of corresponding eye disorders. Cloud computing refers to the ability to access computing resources over the internet for data storage, aggregation, synthesis, and retrieval purposes, as well as to utilize computing algorithms and software packages to act on data.

In a preferred embodiment, the present invention provides a computerized automated OCTA image analysis system for developing and quantifying Computational Retinal Microvasculature Biomarkers (CRMB) based on an OCTA image database and evaluating and quantifying retinal properties in a comprehensive and readily available manner. Using the system of the present invention, clinicians and patients can acquire and evaluate data reports in a few minutes, enabling comprehensive diagnosis and treatment in a clinical setting or out-patient. Rapid and comprehensive analysis may also enable seamless re-imaging in the event that conclusive results cannot be obtained using the initial OCTA report.

Advantageously, the system and quantitative retinal microvasculature biomarkers generated using the systems and methods of the invention can also be used for effective downstream research efforts.

In some embodiments, the cloud database of the system of the present invention stores images derived from thousands of patients, either uploaded by the patients themselves, or collected as part of a clinical study developed for disease or drug discovery, or generated by a conventional clinical workflow in which a group of patients are analyzed in a batch mode.

In a preferred embodiment, data obtained using the system of the present invention is provided to an e-closed deep learning system configured to evaluate and identify early biomarkers of retinal disease (including but not limited to retinal vascular disease) through machine learning techniques.

Using the computational microvascular biomarkers generated using the methods and systems of the present invention, treatment plans associated with the presence of retinal microvascular biomarkers and/or their underlying medical conditions can be designed, and patients can be treated and treatment efficiencies assessed using the CRMBs of the present invention.

In some embodiments, the quantified CRMB is used to quantify the extent of the extra-retinal disease condition. In some embodiments, the quantified CRMB is used to quantify the extent of a retinal disease condition. In some embodiments, the quantified CRMB is used to quantify both an extraretinal disease condition and a retinal disease condition.

In some embodiments, the system of the present invention is integrated into the interface of an OCTA machine, and the user can view the results of real-time disease biomarker quantification after capturing the OCTA images in an ophthalmic clinic.

In some embodiments, the system of the present invention provides machine learning techniques to develop quantization parameters for various extra-retinal disease conditions and retinal disease conditions. In a particular embodiment, machine learning techniques include, but are not limited to, deep learning and gradient boosting decision trees. Advantageously, the systems and methods of the present invention are provided for validating and evaluating novel objective microvasculature biomarkers for ocular disease, thereby facilitating diagnosis, monitoring, and treatment of corresponding ocular disease.

In a preferred embodiment, the systems and methods quantify retinal microvascular parameters including, but not limited to, fractal dimension index, foveal avascular area, and blood flow index. In further preferred embodiments, the systems and methods quantify retinal microvascular parameters, including retinal vessel geometry including, but not limited to, vessel dispersion, vessel diameter, and vessel tortuosity.

In further embodiments, machine learning techniques including, but not limited to, e-closed deep learning are provided to identify and quantify new CRMBs.

In certain embodiments, the system of the present invention acquires OCTA image data. For example, in some embodiments of the cloud-based operation of the system of the present invention, patient OCTA images are generated using, for example, the oeto VUE, cirrus 5000, and/or SPECTRALIS OCT OCTA machines at the medical center, including but not limited to two consecutive 3 x 3mm, frontal OCTA images of superficial and deep capillary plexuses. In other embodiments, a scanned image of the output from the OCTA machine is generated. Advantageously, data entry may be performed by both the ophthalmic doctor user and the patient, who may access the cloud platform and scan the OCTA report from the hospital visit for uploading into the system of the present invention. The OCTA image is uploaded to the cloud through an internet Application Programming Interface (API). In some embodiments, the OCTA images are uploaded as collected during research and development studies and/or clinical studies for disease and/or drug development. In some embodiments, the OCTA images are uploaded in batch mode or by the patient himself during a conventional clinical workflow.

In some embodiments, the uploaded OCTA image is anonymized by removing certain types of metadata and stored in the cloud with server-side encryption. Advantageously, the cloud infrastructure accommodates the management and analysis of large-scale imaging data using the system of the present invention.

In some embodiments, the system of the present invention is configured to calibrate the original OCTA image in order to input the image in a predetermined orientation. For example, when an OCTA original image is scanned from the printed output of the OCTA machine and the image is a portrait view, the system of the present invention rotates the image 90 degrees.

In some embodiments, the scanned image is converted from RGB (R, G, B) to Lab color space (L, a, B) and split equally into left and right halves.

In particular embodiments, pixels in each image half are analyzed, and the clockwise rotation includes more in the left half of the OCTA image

Is a picture of pixels of (a).

In other embodiments, the counter-clockwise rotation includes less in the left half of the OCTA image

Is a picture of pixels of (a). After a corresponding rotation, the scanned image is in the form of a cross-screen view and a predetermined orientation.

In a preferred embodiment, the image rotated to the predetermined orientation is further rotated by a small amount such that the reference lines are strictly horizontal and vertical, respectively.

In a preferred embodiment, the image rotated to the predetermined orientation is rotated further by the required degree (ranging from-90 degrees to 90 degrees) so that the reference lines are strictly horizontal and vertical, respectively.

In some embodiments, the scanned image is converted to grayscale, gaussian blurred to reduce noise, thresholded, and edge processed by a Canny edge detector.

In a preferred embodiment, the rotation angle is determined by detecting the orientation of a reference line in the edge-processed image using a hough line transform.

In further embodiments, a Digital Vascular Map (DVM) is generated from the OCTA image after the calibration procedure described above. For this purpose, the OCTA image is first converted into gray scale and thresholded using binarization of Otsu. All vertical and horizontal line segments in the resulting image are identified using a hough line transform and the detected line segments are combined into a new image. Identifying and classifying contours in the new image; the contour corresponds to a rectangle in the original OCTA image.

In a preferred embodiment, the upper left rectangle with a width/height greater than 0.27 times the width of the original OCTA image is extracted. In embodiments where the upper left rectangle does not have equal length and width, the rectangle is cropped to a square, where the cropping determinant is determined by performing harris angle detection on a filled, gaussian blurred, and thresholded version of the image. The square thus generated of the image is the DVM extracted from the calibrated OCTA image.

In some embodiments, the systems and methods of the present invention apply an early diabetic retinopathy treatment (ETDRS) grid to the generated DVM as described above. In particular embodiments, these systems and methods convert DVM to Lab color space (L, a, b) and replace with white pixels

So that only the color component is retained while removing most of the vessel signal. The resulting image is then converted to grayscale and skeletonized, and the reference circles in the ETDRS grid are identified by hough circle transformation. In particular embodiments, an inner reference circle C _{Inner part} Having a centre (x) _{Inner part} ,y _{Inner part} ) And r _{Inner part} And the outer reference circle C _{Outer part} Having a centre (x) _{Outer part} ,y _{Outer part} )＝(x _{Inner part} ,y _{Inner part} ) And r _{Outer part} Is set, and the radius of (a) is set. In a preferred embodiment, C _{Inner part} The inner zone generally corresponds to the foveal zone, while the secondary foveal zone is referred to as C _{Inner part} And C _{Outer part} An annular region therebetween.

In a specific embodiment, the secondary fovea is then divided into upper, lower, nasal and temporal segments. In some embodiments, the systems and methods calculate the zone boundaries for the right eye (OD) or left eye (OS), respectively, using the recording parameters of the reference circles in the ETDRS grid. In some embodiments, where the coordinates of the upper left corner of the DVM are (0, 0), the end points of the line segments that divide the upper and nasal segments of the right eye are:

And->

In some embodiments, the systems and methods of the present invention calculate fractal dimension index (FD) to quantify the complexity of retinal microvasculature in both the Superficial Capillary Plexus (SCP) and the Deep Capillary Plexus (DCP). The generated DVM is first binarized by adaptive thresholding and FD is calculated by applying a box counting algorithm to the resulting map (Lemmens et al 2020).

In some embodiments, the systems and methods automatically delineate retinal blood vessels in the SCP. In a preferred embodiment, these systems and methods binarize the original DVM by adaptive global thresholding, resulting in a thresholded DVM image. In a preferred embodiment, only foveal regions and sub-foveal regions are reserved, according to the ETDRS grid applied as described above. In some embodiments, the systems and methods of the present invention use a Sato filter to detect continuous ridges corresponding to blood vessels in the resulting DVM. The ridge filtered DVM image is then binarized by adaptive thresholding based on the median of the sub-foveal pixel intensities. In further embodiments, the systems and methods remove small objects from the binarized DVM image, the small objects corresponding to insignificant isolated vessel branches or noise. In further embodiments, the systems and methods identify contours in the resulting DVM, where each contour corresponds to a delineated retinal blood vessel.

In some embodiments, a capillary perfusion density index (PDC) is calculated for both DCP and SCP based on the generated DVM and its accompanying ETDRS grid. The map is first converted to gray scale, where the pixel intensity values range from 0 (black) to 255 (white). Next, PDC is calculated as the average intensity of all non-delineated vessel pixels in the considered section of the graph. Because brighter pixels (with larger intensity values) generally correspond to vascular structures in the OCTA image, when the map has a larger computational PDC, it has denser capillaries.

In some embodiments, a macrovascular perfusion density index (PDL) is calculated based on the generated DVM in the SCP and its ETDRS grid. For the considered section of the graph, PDL is calculated as the proportion of pixels belonging to the vessel or larger vessel depicted therein. When the map has a larger computed PDL, it has a denser large blood vessel.

In a preferred embodiment, these systems and methods calculate a vascular dispersion index (VDisp). Because in a healthy eye, the secondary foveal vessel is more centered towards the center of the fovea, VDisp refers to the degree of centering of the secondary foveal vessel. The larger the VDisp, the less centralised the vessel on average.

In some embodiments, these systems and methods fill and crop the thresholded, ridge-filtered DVM image after the small object is removed, such that the secondary fovea region is centered. In a further embodiment, for each delineated retinal blood vessel V _i (i=1, …, N, where N is the total number of vessels depicted), these systems and methods calculate a region V having a relationship with _i An ellipse E of the same second moment covered. In some embodiments, if V _i Is fully centripetal, it is equally oriented to connect E _i A line segment of the centroid of the center of the fovea.

According to the method of the present invention, VDisps are defined as E _i Main shaft and connection E of (2) _i Centroid sum C of (2) _{Inner part} An average of all angles between line segments of the center of (a), which corresponds to the fovea identified when the ETDRS grid is applied to the extracted and calibrated DVM image. In some embodiments, the systems and methods of the present invention calculate a plurality of vessel dispersion indices for each vessel of the DVM image. In some embodiments, these systems and methods calculate the overall VDisp of the DVM image, which is the average of all individual vessel dispersion indices.

In certain embodiments, these systems and methods separate primary and secondary retinal vessel segments. In particular embodiments, these systems and methods utilize the depicted retinal blood vessels to skeletonize the vasculature map. In further embodiments, the systems and methods detect endpoints and branch points in a skeletonized DVM image by an algorithm comprising the steps of: traversing each positive pixel in the DVM image and counting the number (n) of positive pixels in eight neighborhood regions of pixels; calculating the number of pixels that contain a single positive pixel in its eight neighborhood regions and applying an endpoint label to the pixels; calculating the number of pixels containing three or more positive pixels in eight neighborhood regions thereof, and applying a branch point mark to the pixels; pixels comprising branch point markers and which branch point markers are non-adjacent are calculated and the real branch point markers are applied to the pixels.

In some embodiments, the systems and methods mask the marked branch points from the DVM image and calculate all contours in the masked image, where each contour corresponds to a vessel segment. In further embodiments, the systems and methods represent the total number of branch points as N _bP . Then total number of vessel segments N _seg Equal to 1+2N _bP . In further embodiments, the systems and methods represent the geodesic length of a segment as L _seg And the geodesic length of the main vessel to which it belongs is denoted as L _ves 。

If it is

The segment is marked as a secondary vessel segment. Otherwise, it is marked as a main vessel segment.

In some embodiments, the systems and methods of the present invention calculate a vessel diameter index (VDiam). VDiam is defined as the average of the ratio between the pixel area and the geodesic length of all delineated retinal blood vessels. Representing S in the image as the side length of the DVM, N as the number of vessels depicted and V _i The formula for VDiam, representing the ith depicted vessel, is:

in some embodiments, the total VDiam of the DVM image is calculated and is the average of the individual vdiams of all the markers in the image divided by S.

In some embodiments, the systems and methods of the present invention calculate a vascular tortuosity index (VT). After calculating the end points and branch points, VT is quantized to an average of the ratio between the geodesic length and euclidean length of the separated vessel segments of the DVM image. Will M ₁ Expressed as the number of major vessel segments, M2 expressed as the number of minor vessel segments, S _i ^Primary Denoted as the ith main vessel segment and will be S _i ^{Secondary minor} Expressed as the i-th secondary vessel segment, the formula for VT is:

where a=0.8 is the weighting coefficient. In this VT calculation, the secondary vessel segment is of lower importance because it generally corresponds to a secondary vessel. If the vessels in the DVM image are on average more curved and distorted, the map will have a greater VT. In some embodiments, the individual VT's are displayed in a DVM image. In some embodiments, the total VT of the DVM image is displayed and is a weighted average of the individual VT of all markers in the DVM.

In some embodiments, these systems and methods demarcate the FAZ in the DVM image, which is defined as the hairless vascular zone within the innermost ring of the secondary fovea capillary plexus. In particular embodiments, the reference circles and lines representing the ETDRS grid in the DVM are masked with a median filtered version of their corresponding regions in the graph. In some embodiments, the systems and methods apply median filtering to suppress noise. In a preferred embodiment, the masked and denoised image is adaptively thresholded and blocked (bloated, then eroded) and small objects are removed. In further embodiments, the systems and methods apply bridging, i.e., if a 0-valued pixel has two non-zero adjacent pixels that are not connected, then the 0-valued pixel is set to 1 and the resulting graph is refined. In still further embodiments, the systems and methods remove small isolated objects from the image to improve the robustness of the FAZ identification method.

In some embodiments, these systems and methods calculate the largest polygon in the fovea avascular, i.e., the polygon that does not cover any of the positive vessel signals of the resulting image. In some embodiments, the indication (0, 0) corresponds to the upper left corner of the DVM, and the center of the foveal region is (x _{Inner part} ,y _{Inner part} ) And the side length of DVM is s. In some embodiments, for Θ=0, 3,6,9, …,357 (i.e., 120 Θ total), consider a factor from (x _{Inner part} ,y _{Inner part} ) To (x) _{Inner part} +s·sinθ，y _{Inner part} +s cos θ) on the directed line segment. In some embodiments, the first pixel with a positive signal (i.e., a portion of a blood vessel) is recorded as one of the endpoints of the FAZ. This particular Θ is skipped if the value of all pixels within the image on the directed line segment is zero. If fewer than three searches return a positive signal, an error is reported. In a particular embodiment, the polygons are defined by connecting all n=120 endpoints in the order in which they were found, being approximate FAZ. In other embodiments, polygons are defined by connecting fewer than 120 endpoints to scale the FAZ when fewer than all endpoints are used if some of the searches do not find any positive pixels.

In some embodiments, the systems and methods of the present invention calculate the FAZ Area Index (FAI). In some embodiments, the systems and methods scale vertices of FAZ polygons in a clockwise order as (a) ₁ ,b ₁ )、(a ₂ ,b ₂ )、……、(a _n ,b _n ). In particular embodiments, these systems and methods calculate the area of the polygon using a shoelace formula:

in further embodiments, the systems and methods define FAI as the FAZ polygon area divided by the pixel area of the DVM. Advantageously, the FAI is comparable between images of different sizes if the DVM is always focused on a 3 x 3 area.

In some embodiments, the systems and methods of the present invention calculate the FAZ Perimeter Index (FPI). In some embodiments, these systems and methods calculate the perimeter of the FAZ polygon using the following formula:

in some embodiments, the FPI is defined as the FAZ polygon perimeter divided by the side length (in pixels) of the DVM. Advantageously, if the DVM is always focused on a 3 x 3 region, the FPI is comparable between images of different sizes.

In some embodiments, the systems and methods of the present invention calculate FAZ non-circularity index (FACI). ). In a particular embodiment, the FACI quantifies irregularities in the shape of the calibrated FAZ and is calculated by the following formula:

Advantageously, when FACI is equal to one, FAZ is a perfect circle. In some embodiments, the FAZ has more irregular shapes when the FACI value is greater than one.

In a preferred embodiment, the system of the present invention employs a learning algorithm and uses training data generated by a cloud-based OCTA analysis system to derive model parameters to formulate a CRMB for a particular disease condition being quantified.

In some embodiments, the present invention provides a method for automatic quantitative assessment of retinal microvasculature and generation of retinal digital vascular maps, comprising:

recording Optical Coherence Tomography Angiography (OCTA) image data;

preprocessing OCTA image data; and

a Digital Vascular Map (DVM) of the retina is generated.

In some embodiments, the pre-treatment step of the method further comprises applying an early-treatment diabetic retinopathy study grid to the DVM; calculating fractal dimension index; retinal blood vessels were calibrated.

In further embodiments, the preprocessing step of the method further comprises calculating a vascular dispersion index; separating the primary vessel section and the secondary vessel section; calculating a blood vessel diameter index; calculating a blood vessel curvature index; calibrating a central concave avascular zone; calculating the area index of the central concave avascular zone; and calculating the non-roundness index of the central concave avascular zone.

In a preferred embodiment, the present invention provides a method of extracting a Computational Retinal Microvascular Biomarker (CRMB) from a DVM generated using the methods and systems of the present invention.

In some embodiments, the CRMB is a fractal dimension index, a vascular dispersion index, a capillary perfusion density index, a large vessel perfusion density index, a vessel diameter index, a vessel tortuosity index, a fovea avascular zone area index, a fovea avascular zone perimeter index, and/or a fovea avascular zone non-roundness index. In a preferred embodiment, the CRMB is a vascular dispersion index or a vascular tortuosity index.

In a specific embodiment, the invention provides a method for automatic quantification of a computational retinal microvascular biomarker of the retinal superficial capillary plexus from a subject suspected of having a retinal disease and for treating the subject if CRMB quantification indicates that the subject has a retinal disease, the method comprising:

recording Optical Coherence Tomography Angiography (OCTA) image data;

preprocessing OCTA image data;

generating a retinal Digital Vascular Map (DVM) of a superficial capillary plexus;

providing retinal DVM in a normal subject;

Quantifying at least one CRMB in the DVM of the subject and the DVM of the normal subject;

and treating a retinal disease in the subject if the amount of at least one CRMB in the subject exceeds the amount of said CRMB in a normal subject.

In further embodiments, the pre-processing step of the method comprises applying an early-treatment diabetic retinopathy study grid to the DVM, and quantifying at least one CRMB comprises quantifying a fractal dimension index (FD); quantifying a capillary perfusion density index; quantifying a great vessel perfusion density index; quantifying a vascular dispersion index; quantifying a vessel diameter index; quantifying a vascular tortuosity index; quantifying the area index of the fovea avascular zone; quantifying the foveal avascular zone perimeter index and/or quantifying the foveal avascular zone non-circularity index.

In some embodiments, the invention provides a method for automatic quantification of a computational retinal microvascular biomarker of the deep retinal capillary plexus from a subject suspected of having a retinal disease and treating the subject if CRMB quantification indicates that the subject has a retinal disease, the method comprising:

recording Optical Coherence Tomography Angiography (OCTA) image data;

Preprocessing OCTA image data;

generating a retinal Digital Vascular Map (DVM) of the deep capillary plexus;

retinal DVM providing a deep capillary plexus of a normal subject;

In some embodiments, the pre-processing step of the method comprises applying an early-treatment diabetic retinopathy study grid to the DVM, and quantifying at least one CRMB comprises quantifying a fractal dimension index; quantifying a capillary perfusion density index; quantifying a great vessel perfusion density index; quantifying a vascular dispersion index; quantifying a vessel diameter index; quantifying a vascular tortuosity index; quantifying the area index of the fovea avascular zone; quantifying the foveal avascular zone perimeter index and/or quantifying the foveal avascular zone non-circularity index.

In specific embodiments, the retinal disease is selected from the group consisting of diabetic retinopathy, retinal vein occlusion, and age-related macular degeneration.

In some embodiments, the treatment comprises administering an antibody to an angiogenic factor, an aptamer to an angiogenic factor, an anti-angiogenic steroid, an antioxidant supplement, laser photocoagulation therapy, transpupillary thermotherapy, and/or performing an ocular surgery.

In a preferred embodiment, the invention provides a method for automatic quantification of a computational retinal microvascular biomarker of the retinal superficial capillary plexus from a subject suspected of having an extra-retinal disease and for treating the subject if CRMB quantification indicates that the subject has an extra-retinal disease, the method comprising:

recording Optical Coherence Tomography Angiography (OCTA) image data;

preprocessing OCTA image data;

providing retinal DVM in a normal subject;

and treating the subject for an extra-retinal disease if the amount of at least one CRMB in the subject exceeds the amount of said CRMB in a normal subject.

In a specific embodiment, the pre-processing step of the method comprises applying an early-treatment diabetic retinopathy study grid to the DVM, and quantifying at least one CRMB comprises quantifying a fractal dimension index; quantifying a capillary perfusion density index; quantifying a great vessel perfusion density index; quantifying a vascular dispersion index; quantifying a vessel diameter index; quantifying a vascular tortuosity index; quantifying the area index of the fovea avascular zone; quantifying the foveal avascular zone perimeter index and/or quantifying the foveal avascular zone non-circularity index.

In some embodiments, the extraretinal disease is cardiovascular disease, systemic arterial hypertension, sickle cell anemia, preeclampsia, arteritis, aortic aneurysm, diabetes, and/or hypercoagulability.

In some embodiments, the treatment comprises administration of an antihypertensive drug, cholesterol-regulating drug, aspirin, a beta blocker, an alpha-2 receptor agonist, a central agonist, a peripheral adrenergic inhibitor, a calcium channel blocker, nitroglycerin, an angiotensin converting enzyme inhibitor, an angiotensin II receptor blocker, an anticoagulant, a blood-thinning drug, an antiplatelet agent, a digitalis formulation, a diuretic, a vasodilator, an iron supplement, a steroid, an antidiabetic drug, an intravascular stent procedure, and/or a vascular bypass procedure.

In some embodiments, the treatment comprises administration of metformin, sulfonylurea, meglitinide, thiazolidinedione, DPP-4 inhibitors, GLP-1 receptor agonists, SGLT2 inhibitors, insulin, and/or iron supplements.

In a preferred embodiment, the present invention provides a system for automatic quantitative assessment of retinal microvasculature and generation of retinal digital vascular maps, comprising:

A memory configured to store one or more executable instructions;

a processor configured to execute one or more executable instructions to perform the steps of:

recording Optical Coherence Tomography Angiography (OCTA) image data;

preprocessing OCTA image data; and

a Digital Vascular Map (DVM) of the retina is generated.

In some embodiments, the preprocessing comprises: applying an early-treatment diabetic retinopathy study grid to the DVM; calculating fractal dimension index; retinal blood vessels were calibrated.

In some embodiments, the preprocessing further comprises: calculating a vascular dispersion index; separating the primary vessel section and the secondary vessel section; calculating a blood vessel diameter index; calculating a blood vessel curvature index; calibrating a central concave avascular zone; calculating the area index of the central concave avascular zone; and calculating the non-roundness index of the central concave avascular zone.

In a preferred embodiment, the present invention provides a system for extracting a Computational Retinal Microvascular Biomarker (CRMB) from a DVM, comprising:

a memory configured to store one or more executable instructions;

a processor configured to execute one or more executable instructions to extract a Computational Retinal Microvascular Biomarker (CRMB) from the aforementioned DVM.

In some embodiments, the CRMB is selected from the group consisting of a blood fractal dimension index, a capillary perfusion density index, a large blood vessel perfusion density index, a blood vessel dispersion index, a blood vessel diameter index, a blood vessel tortuosity index, a fovea avascular zone area index, a fovea avascular zone perimeter index, and a fovea avascular zone non-roundness index.

In a preferred embodiment, the present invention provides a system for automatic quantification of a computational retinal microvascular biomarker of the retinal superficial capillary plexus from a subject suspected of having a retinal disease and for treating the subject if CRMB quantification indicates that the subject has a retinal disease, the system comprising:

a memory configured to store one or more executable instructions;

recording Optical Coherence Tomography Angiography (OCTA) image data;

preprocessing OCTA image data;

providing retinal DVM in a normal subject;

quantifying at least one CRMB in the DVM of the subject and the DVM of the normal subject; and

If the amount of at least one CRMB in the subject exceeds the amount of said CRMB in a normal subject, the retinal disease in the subject is treated.

In some embodiments, the preprocessing comprises: applying an early treatment diabetic retinopathy study grid to the DVM, and wherein quantifying at least one CRMB comprises quantifying a fractal dimension index; quantifying a capillary perfusion density index; quantifying a great vessel perfusion density index; quantifying a vascular dispersion index; quantifying a vessel diameter index; quantifying a vascular tortuosity index; quantifying the area index of the fovea avascular zone; quantifying the foveal avascular zone perimeter index and/or quantifying the foveal avascular zone non-circularity index.

In some embodiments, the retinal disease is selected from diabetic retinopathy, retinal vein occlusion, and age-related macular degeneration, and the treatment comprises administering antibodies to vascular growth factors, aptamers to vascular growth factors, anti-angiogenic steroids, antioxidant supplements, laser photocoagulation therapy, transpupillary thermotherapy, and/or performing ocular surgery.

In a preferred embodiment, the present invention provides a system for automatic quantification of a computational retinal microvascular biomarker of the deep retinal capillary plexus from a subject suspected of having a retinal disease and for treating the subject if CRMB quantification indicates that the subject has a retinal disease, the system comprising:

A memory configured to store one or more executable instructions;

recording Optical Coherence Tomography Angiography (OCTA) image data;

preprocessing OCTA image data;

generating a retinal Digital Vascular Map (DVM) of the deep capillary plexus;

retinal DVM providing a deep capillary plexus of a normal subject;

In some embodiments, the pre-processing comprises applying an early-treatment diabetic retinopathy study grid to the DVM, and quantifying the at least one CRMB comprises: quantifying fractal dimension index; quantifying a capillary perfusion density index; quantifying a great vessel perfusion density index; quantifying a vascular dispersion index; quantifying a vessel diameter index; quantifying a vascular tortuosity index; quantifying the area index of the fovea avascular zone; quantifying the foveal avascular zone perimeter index and/or quantifying the foveal avascular zone non-circularity index.

In a preferred embodiment, the present invention provides a system for automatic quantification of a computational retinal microvascular biomarker of the retinal superficial capillary plexus from a subject suspected of having an extra-retinal disease and for treating the subject if CRMB quantification indicates that the subject has an extra-retinal disease, the system comprising:

a memory configured to store one or more executable instructions;

recording Optical Coherence Tomography Angiography (OCTA) image data;

preprocessing OCTA image data;

providing retinal DVM in a normal subject;

if the amount of at least one CRMB in the subject exceeds the amount of said CRMB in a normal subject, treating the subject for an extra-retinal disease.

In some embodiments, the extra-retinal disease is selected from cardiovascular disease, systemic arterial hypertension, sickle cell anemia, preeclampsia, arteritis, aortic macroaneurysms, diabetes, and/or hypercoagulability, and the treatment comprises administration of antihypertensive agents, cholesterol-modulating agents, aspirin, beta blockers, calcium channel blockers, nitroglycerin, angiotensin converting enzyme inhibitors, angiotensin II receptor blockers, blood-thinning agents; iron supplements, steroids, antidiabetic agents, intravascular stent surgery and/or vascular bypass surgery.

In a preferred embodiment, the present invention provides a computer readable storage medium containing a computer program, and the computer program is executable by a processor to implement the steps of the method of the present invention.

Materials and methods

All patents, patent applications, provisional applications, and publications mentioned or cited herein are incorporated herein by reference in their entirety (including all figures and tables) to the extent not inconsistent with the explicit teachings of this specification.

The following is an example illustrating a procedure for practicing the present invention. These examples should not be construed as limiting. Unless otherwise indicated, all percentages are by weight and all solvent mixture ratios are by volume.

Example 1-System for computerized automatic OCTA image analysis

An automated analysis system (fig. 1) was designed for OCTA image analysis using a cloud platform. The system is designed for use by ophthalmologists and researchers at medical centers.

To this end, patient OCTA images were obtained from different OCTA machines available from medical centers containing OPTOVUE, cirrus and SPECTRALIS OCT2, which included two consecutive 3 x 3mm OCTA frontal images of superficial and deep capillary plexuses. In addition to the ophthalmologist, the patient can also access the cloud platform of the system and scan their OCTA report from the hospital visit. The OCTA image is uploaded to the cloud through an internet Application Programming Interface (API), anonymized by deleting certain types of metadata, and stored in the cloud with server-side encryption. The cloud infrastructure can house, manage, and analyze large-scale imaging data. Advantageously, the computerized automated OCTA image analysis system of the present invention determines the CRMB based on the image and returns a report to the user within a few minutes. Thus, during a patient out-patient visit, a clinician may discuss the OCTA results with the patient. Further, the system enables seamless re-imaging in the event that conclusive results cannot be obtained using the initial OCTA report. Moreover, the user in development can conduct an effective downstream statistical analysis of the quantitative evaluation of biomarkers. Advantageously, the cloud database can store images from thousands of patients, including images from data uploaded by the patients themselves, images collected as part of a clinical study of a disease or drug discovery study, or images collected during a conventional clinical workflow in which a group of patients are analyzed in a batch mode. The data obtained using the instant system is used to develop predictive features through machine learning techniques, including deep learning and gradient-driven decision trees.

Example 2-method for computerized automated OCTA image analysis of fractal dimension

A process of extracting different CRMBs in accordance with an embodiment of the present technique is generated (fig. 2, block [1 ]]). The process obtains OCTA image data (FIG. 2 block [2 ]]And fig. 3A to 3B). The process calibrates each original OCTA image so that it is in the correct orientation (FIG. 2, block [3 ]]). This is often required when the original image is a scanned version of the printout of the OCTA machine. For example, when the image is in a portrait viewIn the form of (c), it is rotated 90 degrees. Then, the scanned image is converted from RGB (R, G, B) to Lab color space (L, a, B) and equally divided into left and right halves. If the left half has more

The image is rotated clockwise. Otherwise, the image is rotated counterclockwise. After rotation, all scanned images are in the form of a cross-screen view and in the correct orientation.

The image thus oriented is further rotated by a small angle in order to make the reference line strictly horizontal/vertical. The scanned image is then converted to gray, blurred gaussian to reduce noise, thresholded and edge processed by a Canny edge detector. Then, the rotation angle is determined by detecting the orientation of the reference line in the edge-processed image using hough line transform.

The process extracts and generates a Digital Vasculature Map (DVM) from the calibrated OCTA image (fig. 2, block [4 ]). For this purpose, the OCTA image is first converted into gray scale and thresholded using binarization of Otsu. All vertical/horizontal line segments in the resulting image are identified by hough line transforms and the detected line segments are combined into a new image. Contours in the new image correspond to rectangles in the original OCTA image. The upper left rectangle with a width/height greater than 0.27 times the width of the original OCTA image is extracted. If the rectangle does not have equal length and width, it is cropped to a square, where the cropping coordinates are determined by performing Harris angle detection on the filled, gaussian blurred and thresholded version of the image. Thus, this square is the DVM extracted from the calibrated OCTA image. The DVM generation process is shown in fig. 4.

The process also applies an diabetic retinopathy Early Treatment (ETDRS) grid to the DVM generated as described above (fig. 2, block [5 ]]). To discard most of the vessel signal, the map is converted to Lab color space (L, a, b) and replaced with white pixels

Is used for the display of the display device,so as to retain only the color components. The resulting image is then converted to grayscale and skeletonized, and the reference circles in the ETDRS mesh are identified by Hough circle transform, where the inner reference circle C _{Inner part} Comprising a centre (x _{Inner part} ,y _{Inner part} ) And r _{Inner part} And the outer reference circle C _{Outer part} Comprising a centre (x _{Outer part} ,y _{Outer part} )＝(x _{Inner part} ,y _{Inner part} ) And r _{Outer part} Is set, and the radius of (a) is set. C (C) _{Inner part} The inner region generally corresponds to the foveal region, and the secondary foveal region refers to C _{Inner part} And C _{Outer part} An annular region therebetween. The process of applying the ETDRS grid to the DVM is shown in fig. 5A.

Next, the secondary fovea is divided into upper, lower, nasal and temporal segments by calculating segment boundaries from the recorded parameters of whether the eye under study is the right eye (OD) or the left eye (OS) using the reference circles in the ETDRS grid. For example, note that the coordinates of the upper left corner of the DVM are (0, 0), the end points of the line segments that divide the upper and nasal segments of the right eye are:

and->

Fig. 5B illustrates the sections defined by the ETDRS grid in the left and right eyes.

The process shown in fig. 2 determines a fractal dimension index (FD) at block [6] based on the DVM from block [4] to quantify the complexity of retinal microvascular structure in both the Superficial Capillary Plexus (SCP) and the Deep Capillary Plexus (DCP). The generated DVM is first binarized by adaptive thresholding and FD is calculated by applying a box counting algorithm to the resulting map.

This process automatically delineates retinal blood vessels in the superficial capillary plexus (fig. 2, box [7 ]). The process of retinal vessel delineation is shown in images a through G of fig. 6. The original vasculature map (image a of fig. 6) is first binarized by adaptive global thresholding, resulting in image B of fig. 6B. Depending on the ETDRS grid applied, only foveal and sub foveal areas are reserved (see fig. 2, block [5 ]). Next, successive ridges or blood vessels are detected in the resulting vasculature map using the filter of Sato (see image C of fig. 6). The ridge filtered image (see image D of fig. 6) is again binarized by adaptive thresholding based on the median of the sub-foveal pixel intensities, and small objects, which may be insignificant isolated vessel branches or noise, are removed from the binarized image (see image E of fig. 6) for better robustness. Finally, all contours in the resulting map see (image F of fig. 6) are identified. Each contour corresponds to a depicted retinal blood vessel, as shown in image G of fig. 6.

Example 3-method for computerized automated OCTA image analysis of pupil perfusion density

The process shown in fig. 2 determines a capillary perfusion density index (PDC) at block [8] based on the generated DVM and the ETDRS grid accompanying it. The map is first converted to gray scale, where the pixel intensity values range from 0 (black) to 255 (white). Next, PDC is calculated as the average intensity of all non-delineated vessel pixels in the considered section of the graph. Because brighter pixels (with larger intensity values) generally correspond to vascular structures in the OCTA image, when the map has a larger computational PDC, it has denser capillaries.

Example 4 computerized automated OCTA image analysis method for great vessel perfusion Density

The process shown in fig. 2 determines a macrovascular perfusion density index (PDL) based on the SCP and the generated DVM in its ETDRS grid at block [9 ]. For the considered section of the graph, PDL is calculated as the proportion of pixels belonging to the vessel or larger vessel depicted therein. When the map has a larger computed PDL, it has a denser large blood vessel.

Example 5-computerized automated OCTA image analysis method for vascular dispersion

The procedure described in example 2 was also used to determine the vascular dispersion index (VDisps) (FIG. 2, block [10 ]]And image G of fig. 6). In healthy eyes, the secondary foveal blood vessels are more concentric toward the center of the fovea. Vascular dispersity refers to the pairThe degree of centrality of the foveal blood vessel. The greater the vascular dispersion, the less centrality the vascular average. To determine VDisp, the image shown in image F of FIG. 6 is first filled and cropped such that the secondary foveal region is centered. Next, for each depicted retinal blood vessel V _i (i=1, …, N, where N is the total number of vessels depicted), determined to have a region V _i Ellipse E of the same second moment covered _i . If V is _i Is fully centripetal, then it is connected with E _i Is oriented the same as the line segment of the center of the fovea. Thus, VDisp is defined as E _i Main shaft and connection E of (2) _i Centroid sum C of (2) _{Inner part} The average of all angles between line segments of the center of (which corresponds to) the fovea, as identified when the ETDRS grid is applied to the DVM (as in fig. 2, block [5 ]]Shown), this process of VDisp calculation is shown in fig. 7. An example of the determined VDisps for the various blood vessels is shown in FIG. 9A. For example, as shown in FIG. 9A, the total VDisps for an example DVM is the average of all the individual VDisps calculated and displayed in the label.

The process also separates the primary and secondary retinal vessel segments (FIG. 2, block [11 ]]). The process of separating the large vessel segment and the small vessel segment is shown in image F of fig. 6 to image I of fig. 6. First, a vascular system map with retinal blood vessels depicted (image F of fig. 6 and block [7 ] of fig. 2]) Is skeletonized. The endpoints and branching points in the skeletonized image (see image H of fig. 6) are detected by the algorithm shown in fig. 8. The identified branch points are then masked from the image and all contours in the masked image are found. Each identified contour corresponds to a vessel segment. The total number of branch points is denoted as N _bP . Then total number of vessel segments N _seg Equal to 1+2N _bP . Further, the geodesic length of the line segment is denoted as F _seg And the geodesic length of the main vessel to which it belongs is denoted L _ves . If it is

The segment is marked as a secondary vessel segment. Otherwise, the segment is markedIs denoted as the main vessel segment.

Example 6-method for computerized automated OCTA image analysis of vessel diameter index

The described procedure is also used to determine a vessel diameter index (VDiam) (FIG. 2, block [12 ]]Image G of fig. 6 and image H of fig. 6). VDiam is defined as the average of the ratio between the pixel area and geodesic length of all delineated retinal blood vessels. Representing S as the side length of a Digital Vascular Map (DVM), N as the number of vessels depicted, and V _i The formula for VDiam, expressed as the ith depicted vessel in the image, is:

an example of the determined VDiam for each vessel is shown in fig. 9B. The total VDiam of a DVM is the average of the individual vdiams of all the markers in the image divided by S.

Example 7-method for computerized automated OCTA image analysis of vascular tortuosity

The above procedure is further used to determine a vascular tortuosity index (VT) (fig. 2, block [13 ]]And image I of fig. 6). Vascular Tortuosity (VT) is quantified as the average of the ratio between geodesic length and euclidean length of the separated vessel segments. Will M ₁ Expressed as the number, M, of major vessel segments ₂ Expressed as the number of secondary vessel segments, S _i ^Primary Denoted as the ith main vessel segment and will be S _i ^{Secondary minor} Expressed as the i-th secondary vessel segment, the formula for VT is:

where a=0.8 is the weighting coefficient. Because the secondary vessel segments generally correspond to secondary vessels, they are of less importance in calculating VT. The graph has a larger VT if the vessels in the graph are on average more curved and distorted. An example of calculated VT for each vessel segment is shown in fig. 9C. The total VT of the example DVM is a weighted average of the individual VT of all the markers in the graph.

Example 8-method for computerized automated OCTA image analysis of foveal avascular regions

The above procedure is further used to label Foveal Avascular Zones (FAZ) in DVM (FIG. 2, block [14 ]). FAZ is defined as the hairless vascular zone within the innermost ring of the secondary foveal capillary plexus. The reference circles and lines representing the ETDRS grid in the DVM are masked with their median filtered versions of the corresponding regions in the figure. Median filtering is then applied to suppress noise. The masked and denoised image (see, e.g., image B of fig. 10) is adaptively thresholded and closed (dilated, then eroded), followed by removal of small objects. Subsequently, bridging (setting 0-valued pixels to 1 if they have two non-zero adjacent pixels that are not connected) and refinement are applied to the resulting map (see, for example, image C of fig. 10). Again, small isolated objects are removed from the image to improve the robustness of the FAZ identification method.

The largest polygon in the foveal avascular zone, i.e. the polygon that does not cover any of the positive vessel signals, is obtained (see e.g. image D of fig. 10). The position (0, 0) corresponds to the upper left corner of the DVM. The center of the foveal region is denoted (x) _{Inner part} ,Y _{Inner part} ) (in FIG. 2, block [5 ]]Identified herein), and the side length of the DVM is denoted as s. For Θ=0, 3,6,9, …,357 (i.e. 120 Θ total), consider the following (x _{Inner part} ,y _{Inner part} ) To (x) _{Inner part} +s·sinθ，y _{Inner part} +s cos θ) on the directed line segment. The first pixel with a positive signal (i.e. a part of a blood vessel) is recorded as one of the endpoints of the FAZ. Half of these endpoints are shown as red dots in image E of fig. 10. If the values of all pixels in the image on the directed line segment are zero, then the specific Θ is skipped. If fewer than three searches return a positive signal, an error is reported. The polygon defined by connecting all n=120 (or less if some of the searches do not find any positive pixels) endpoints in the order in which they were found is an approximate FAZ (see, for example, image F of fig. 10). The overall process of calibrating FAZ is shown in FIG. 10。

Example 9-method for computerized automated OCTA image analysis of foveal avascular zone area index

The above procedure is used to determine the FAZ Area Index (FAI) (FIG. 2, block [15 ]]). In a clockwise order, the vertices of the FAZ polygon are represented and labeled as (a) as shown in FIG. 10 ₁ ,b ₁ )、(a ₂ ,b ₂ )、……、(a _n ,b _n ). The area of the polygon is calculated using the shoelace formula:

FAI is then defined as the FAZ polygon area divided by the pixel area of the DVM. Advantageously, the FAI is comparable between images of different sizes if the DVM is always focused on a 3 x 3 area.

Example 10-method for computerized automated OCTA image analysis of the circumference index of the foveal avascular zone

The above procedure is used to determine the FAZ Perimeter Index (FPI) (fig. 2, block [16 ]). According to the notation of block [15] of fig. 2, the perimeter of the FAZ polygon is calculated using the following formula:

the FPI is then defined as the FAZ polygon perimeter divided by the side length (in pixels) of the DVM. Advantageously, if the DVM is always focused on a 3 x 3 region, the FPI is comparable between images of different sizes.

Example 11-method for computerized automated OCTA image analysis of non-circularity index of foveal avascular zone

The described procedure is used to determine the FAZ non-circularity index (FACI) (fig. 2, block [17 ]). The FACI quantifies the irregularities in the shape of the calibrated FAZ and is calculated using the following formula:

when FACI equals one, FAZ is a perfect circle. A larger value of FACI indicates that the FAZ has more irregular shape.

Example 12 measurement of computational retinal microvascular biomarkers

Using the techniques discussed above, different Computational Retinal Microvascular Biomarkers (CRMBs) are automatically determined based on the OCTA images, including FD on the Superficial Capillary Plexus (SCP) and the Deep Capillary Plexus (DCP), PDC on the SCP and DCP, PDL, VDisp, VDiam and VT of the SCP, and FAI, FPI and FACI on both the SCP and DCP. To illustrate the usefulness of the proposed computerized automated analysis system, CRMB analysis was performed on 43 OCTA images from RVO patients and 30 OCTA images from healthy subjects. All OCTA images used in the study were 3mm x 3mm scans, taken from both SCP and DCP, and analyzed using the automated system of the present invention. Different CRMBs were determined, which represent the degree of retinal abnormalities from both the SCP and DCP layers in the macular region. Summary statistics are generated for the different CRMBs and student t-tests are performed to compare the magnitudes of CRMBs between the normal and RVO groups.

The fractal dimension index (FD) of both the SCP layer and DCP layer in eyes with RVO was significantly lower than those of control eyes (table 1 and fig. 11). Additionally, the FAI, FPI, and FACI of the two layers in the eyes with RVOs were significantly greater than the FAI, FPI, and FACI of the control eyes.

The vascular dispersion of the eyes with RVO is significantly greater than that of the control group. These results indicate that the proposed analysis system captures very well the clinically important properties of RVO. CRMB is also used to study other types of eye diseases to gain insight into the underlying clinical interpretation of these conditions.

TABLE 1 comparison of CRMB between Normal and RVO groups

Example 13 computational retinal microvascular biomarkers

To evaluate the effectiveness of the defined CRMB, pearson correlation tests were performed between ten secondary foveal LD phenotypes (SCP and DCP overall, temporal area, upper, nasal and lower) and the FAZ area provided by the machine-built-in OCTA software using the system of the present invention. As shown in fig. 11, all correlations are strongly positive (ranging from 0.70 to 0.88) and differ significantly from zero at the adjusted p-value cut-off point of 0.01. This demonstrates the reliability of the computerized automated analysis system of the present invention and its calculated CRMB.

It is understood that the examples and embodiments described herein are for illustrative purposes only and that various modifications or changes in light thereof will be suggested to persons skilled in the art and are to be included within the spirit and purview of this application. Furthermore, any element or limitation of any invention disclosed herein or embodiments thereof may be combined with any and/or all other elements or limitations disclosed herein (alone or in any combination) or any other invention or embodiments thereof, and all such combinations are contemplated as being within the scope of the present invention without limitation thereto.

Reference to the literature

Adhi, M., filho, M.A., louzada, R.N., kuehlewein, L., de Carlo, T.E., baumal, C.R., … Waheed, N.K, "analysis of retinal capillary networks and foveal avascular areas in venous occluded and contralateral eyes using optical coherence tomography angiography" (Retinal CapillaryNetwork and FovealAvascularZone in Eyes withVein Occlusion and Fellow Eyes Analyzed With Optical Coherence Tomography Angiography) "," ophthalmic research and vision (Invest Ophthalmol Vis Sci) ", 2016, 10 months 57 (9), 486-494.Doi:10.1167/iovs.15-18907

Aghamohamadian-Sharbaf, M., pourreza, H.R., banaee, T, "novel curvature-based algorithm for automatic grading of retinal vascular curvature (A Novel Curvature-Based Algorithm for Automatic Grading of Retinal Blood Vessel Tortuosity)", IEEE biomedical and health informatics (IEEE JBiomed Health Inform), 2016, 20 (2), 586-595.doi:10.1109/jbhi.2015.2396198

Arrigo, A., aragona, E., capone, L., pierro, L., romano, F., bandello, F., parodi, M.B, "advanced optical coherence tomography angiography of age-related macular degeneration concurrent unilateral choroidal neovascularization (Advanced Optical Coherence Tomography Angiography Analysis of Age-related Macular Degeneration Complicated by Onset ofUnilateral Choroidal Neovascularization)", "journal of eye science (Am JOphthalmol) 2018, 195,233-242.doi:10.1016/j.ajo.2018.08.001

Arrigo, A., romano, F., albertini, G., argona, E., bandelo, F., & BattagliaParodi, M., "vascular pattern of retinitis pigmentosa according to scanning source optical coherence tomography angiography (Vascular Patterns in Retinitis Pigmentosa on Swept-Source Optical Coherence TomographyAngiography)", "journal of clinical medicine (JClin Med)", 2019, 5 (9) doi:10.3390/jcm8091425

Brar, m., sharma, m., grewl, s.p.s., grewl, D.S, "quantification of retinal microvascular and neurodegenerative changes in retinal branch vein occlusion following maculocyst edema extinction by optical coherence tomography angiography (Quantification ofretinal microvasculature and neurodegeneration changes inbranch retinal vein occlusion after resolution of cystoid macular edema on optical coherence tomography angiography)", "journal of indian ophthalmology (Indian J Ophthalmol), 2019, 67 (11), 1864-1869.doi:10.4103/ijo.ijo_1554_18

Optical coherence tomography angiography of retinal vein occlusion of "Semi-automatic detection of foveal avascular zone in fluorescent angiograms in diabetes (Semi-automated detection of the foveal avascular zone in fluorescein angiograms in diabetes mellitus)", of conrate, j, valat, o, giorgi, r, adel, m, raccah, d, meyer, f, & ridins, B, ", clinical and experimental ophthalmology (Clin Exp Ophthalmol)," 2006, 34 (2), 119-123.doi:10.1111/j.1442-9071.2006.01138.Xcoscas, f, glaset-Bernard, a, miere, a, caillaux, v, uzzan, j, fupidi, m, … soided, E.H: evaluation of the superficial and deep capillary Conflict (Optical Coherence Tomography Angiography in Retinal Vein Occlusion: evaluation of Superficial and Deep Capillary Plexa) "," J.S. Ocular science 2016, 161,160-171.e. 161-162.doi:10.1016/j.ajo.2015.10.008

Two different optical coherence tomography angiography devices of Costanzo, e, parravano, m., gilradi, m., cavaleri, m., sacconi, r., araona, e., … queries, G "two different optical coherence tomography angiography devices analyze microvascular retinal and choroidal changes of retinal vein occlusion (Microvascular Retinal and Choroidal Changes in Retinal Vein Occlusion Analyzed by Two Different Optical Coherence Tomography Angiography Devices)", "journal of ophthalmology (ophtalmologica) 2019, 242 (1), 8-15.doi:10.1159/000496195

Quantitative analysis of retinal microvascular changes (Quantitative Analysis of Retinal Microvascular Changes after Conbercept Therapy in Branch Retinal Vein Occlusion Using Optical Coherence Tomography Angiography) following retinal branch vein occlusion kangbiepu treatment using optical coherence tomography angiography, "Deng, y, cai, x, zhang, s, su, l, chen, h, lin, y, … Chi, W," journal of ophthalmology, 2019, 242 (2), 69-80.doi:l 0.1159/000499608

Diaz, M., "automatic segmentation of foveal avascular regions of ophthalmic OCT-A images (Automatic segmentation of the foveal avascular zone in ophthalmological OCT-A images)", "public science library-complex (PLoS One)", 2019, 14 (2), e0212364.doi 10.1371/journ.fine.0212364, F., "Gomez-Ulla, F.," Penedo, M.G., "Ortega, M.

Dumitruscu, O.M., demaerschalk, B.M., valencia Sanchez, C., almader-Douglas, D., O' Carroll, C.B., aguilar, M.I., … Kumar, G, "retinal microvascular abnormalities" are used as surrogate markers for cerebral ischemic disease: meta Analysis (Retinal MicrovascularAbnormalities as Surrogate Markers ofCerebrovascular Ischemic Disease: AMeta-Analysis) "," journal of stroke and cerebrovascular diseases (JStroke CerebrovascDis), 2018, 27 (1), 1960-1968.Doi:10.1016/j. Jstrokeerebrovasdis.2018.02.041)

Grisan, E., foracchia, M., & Ruggeri, A, "novel method for automatic grading of retinal vascular curvature" (Anovel method forthe automatic grading ofretinal vessel tortuosity), "IEEE medical image Magazine (IEEE TransMedImaging)," 2008, 27 (3), 310-319. Doi:10.1109/tmi.2007.904657)

Hanada, n., iijima, h., sakurada, y., & imagawa, M, "recurrence of macular edema associated with retinal branch vein occlusion after intravitreal injection of bevacizumab (Recurrence of macular edema associated with branch retinal vein occlusion after intravitreal bevacizumab)", "journal of ophthalmology, jpn J Ophthalmol", 2012, 56 (2), 165-174.doi:10/s 10384-011-0113-4

"does the retinal vessel geometry vary with cardiac cycle" by Hao, h., sasong ko, m.b., wong, T.Y., cheAzemin, M.Z., alihmad, b., hodgson, l., … Kumar, D.K? (Does retinal vascular geometryvary with cardiac cycle

Hartnet, m.e., martiniak, d., byfield, g., geisen, p., zeng, g., & Bautch, V.L, "in ROP rat model, neutralizes VEGF to reduce arteriole and venous tortuosity and alters endothelial cell division direction: related to additional diseases (Neutralizing VEGF decreases tortuosity and alters endothelial cell division orientation in arterioles and veins in a rat model ofROP: relevance to plus disease) "," ophthalmic research and visual mechanics ", 2008, 49 (1), 3107-3114.Doi:10.1167/iovs.08-1780

Huang, D., jia, Y., gao, S.S., lumbroso, B., & Rispoli, M, "optical coherence tomography angiography (Optical Coherence Tomography AngiographyUsing the Optovue Device) using an Optovue device," development of ophthalmology (Dev Ophthalmol), "2016, 56,6-12.Doi:10.1159/000442770

Evaluation of "evaluation of retinal branch vein occlusion macular ischemia of eyes" by Kadomoto, s., muraoka, y., ooto, s., miwa, y., iida, y., suzuma, k., … Yoshimura, N: optical coherence tomography angiography study (EVALUATION OF MACULAR ISCHEMIA IN EYES WITH BRANCH RETINAL VEIN OCCLUSION: an Optical Coherence TomographyAngiography Study) "," Retina (Retina) "2018, 38 (2), 272-282.doi:10.1097/iae.0000000000001541

Quantitative measurements of vascular density and flow in patients with central retinal vein occlusion were performed using Optical Coherence Tomography Angiography (OCTA), khadabandh, a, shahraki, k, roohipore, r., riazi-Esfahani, h., yaser, m., faghihi, h., bazvand, F: is OCTA able to help differentiate between ischemic and non-ischemic groups? (Quantitative measurement of vascular density and flow using Optical Coherence Tomography Angiography (OCTA) in patients with central retinal vein occlusion: can OCTA help in distinguishing ischemic from non-ischemical type

Improved analysis of foveal avascular area by optical coherence tomography angiography of Kulikov, a.n., maltsev, d.s., and burns, M.A (Improved analysis of foveal avascular zone area with optical coherence tomography angiography), gray clinical and experimental ophthalmic literature (GraefesArch Clin Exp Ophthalmol) 2018, 256 (12), 2293-2299.Doi:10.1007/s00417-018-4139-x

Lee, H, fee, M., chung, H, & Kim, H.C, "quantitative retinal vascular tortuosity (QUANTIFICATION OF RETINAE VESSEF TORTUOSITY IN DIABETIC RETINOPATHY USING OPTICAF COHERENCE TOMOGRAPHY ANGIOGRAPHY) in diabetic retinopathy using optical coherence tomography angiography", "retina", 2018, 38 (5), 976-985.Doi:10.1097/iae.0000000000001618 "

Lee, H.E., wang, Y., fahed, A.E., and Fawzi, A.A, "study of retinal vein occlusion side branch versus vessel density using optical coherence tomography angiography" (Exploring the relationship between collaterals and vessel density in retinal vein occlusions using optical coherence tomography angiography), "public science library, complex" 2019, 14 (1), e0215790.Doi: 10.1371/journ. Fine. 0215790)

McIntosh, r.l., rogers, s.l., lim, l., cheung, n., wang, j.j., mitchell, p., … Wong, T.Y, "natural history of central retinal vein occlusion: evidence-based System review (Natural history of central retinal vein occlusion: an authentication-based systematic review) "," Ophthalmology (ophtalmology) 2010, 117 (6), 1113-1123.el 115.doi:10.1016/j.ophtha.2010.01.060

The physiological principle of "minimum work" of Murray, C.D: I. vascular system and blood volume depletion (The Physiological Principle of Minimum Work: I.the Vascular System and the Cost ofBlood Volume) "," Proc. Natl. Acad. Sci. USA (ProcNatlAcad Sci USA) 1926, 12 (3), 207-214.Doi:10.1073/pnas.12.3.207

The physiological principle of "minimum work" in Murray, C.D: oxygen exchange in capillaries II (The Physiological Principle of Minimum Work: II. Oxygen Exchange in Capillaries) ", proc. Natl. Acad. Sci. USA 1926, 12 (5), 299-304.Doi:10.1073/pnas.12.5.299

"overview of pediatric retinal microvascular changes as a predictor of cardiovascular disease (Review of paediatric retinal microvascular changes as a predictor of cardiovascular disease)", newman, a.r., andrew, n.h., casson, R.J, 45 (1), 33-44.doi:l0.1111/ceo.l2773, clinical and experimental ophthalmology 2017

Effect of anti-VEGF treatment of Papadia, m., miselli, m., jeannin, b., & Herbort, C.P on modern management of macular edema caused by BRVO and CRVO: longitudinal analysis of visual function, injection time interval and complications (The influence of anti-VEGF therapy on present day management of macular edema due to BRVO and CRVO: a longitudinal analysis on visual function, injection time interval and complications) "," International ophthalmology (Int ophtalmol), "2014, 34 (6), 1193-1201.Doi:10.1007/sl0792-014-0002-l

Qian, t., zhao, m., & Xu, X "anti-VEGF treatment compared to corticosteroid or laser treatment of macular edema secondary to retinal vein occlusion: meta analysis "(Comparison between anti-VEGF therapy and corticosteroid or laser therapy for macular oedema secondary to retinal vein occlusion: A meta analysis)," journal of clinical pharmacy and treatment (J Clin Pharm Ther), "2017, 42 (5), 519-529.Doi: 10.1111/jcpt.12551)

Rispoli, m., savostano, m.c. & Lumbroso, B, "capillary network abnormalities according to retinal branch vein occlusion (CAPILLARY NETWORK ANOMALIES IN BRANCH RETINAL VEIN OCCLUSION ON OPTICAL COHERENCE TOMOGRAPHYANGIOGRAPHY) of optical coherence tomography angiography," retina ", 2015, 35 (11), 2332-2338.Doi:10.1097/iae.0000000000000845

Rogers, s., mcIntosh, r.l., chemng, n., lim, l., wang, j.j., mitchell, p., … Wong, T.Y, "prevalence of retinal vein occlusion: summary data from U.S. European, asian and Australian population studies (The prevalence of retinal vein occlusion: pooled data from population studies from the United States, europe, asia, and Australia) "," ophthalmology, 2010, 117 (2), 313-319.e311.doi:10.1016/j.ophtha.2009.07.017 "

Rogers, S.L., mcIntosh, R.L., lim, l., mitchell, p., cheung, n., kowalski, j.w., … Wong, T.Y, "natural history of retinal branch vein occlusion: evidence-based System review "(Natural history of branch retinal vein occlusion: an evaluation-based systematic review), ophthalmology, 2010, 117 (6), 1094-1101.el095.Doi:10.1016/j.ophtha.2010.01.058

Sasongko, M.B., hodgson, L.A., wong, T.Y., kawasaki, R., cheung, C.Y., hsu, W., … Wang, J.J, "correlation and repeatability of retinal vessel geometry measurements for binocular stereoscopic retinal images (Correlation and reproducibility of retinal vascular geometric measurements for stereoscopic retinal images of the same eyes)", ophthalmic epidemiology (Ophthalmic Epidemiol) 2012, 19 (5), 322-327.Doi:10.3109/09286586.2012.702258

Sasongko, M.B., wang, J.J., donaghue, K.C., cheung, N., benitez-Aguirre, P., jenkins, A., … Wong, T.Y, "change of retinal microvascular geometry in young type 1 diabetics (Alterations in retinal microvascular geometry in young type Diabetes)", "Diabetes Care", 2010, 33 (6), 1331-1336.doi:10.2337/dcl0-0055

Seknazi, d., coscas, f., selam, a, rouimi, f, coscas, g., souied, e.h., and glaset-Bernard, a). Based on the correlation between the density of the macular vessels, vision and peripheral non-perfused areas of fluorescein angiography (OPTICAL COHERENCE TOMOGRAPHY ANGIOGRAPHY IN RETINAL VEIN OCCLUSION: correlations Between Macular Vascular Density, visual Acuity, and Peripheral Nonperfusion Area on Fluorescein Angiography) "," retina "2018, 35 (8), 156-1570.doi: 10.1097/iae.0000000000001737

Spaide, r.f. "microvascular blood flow abnormalities associated with retinal vasculitis: potential retinal injury mechanism (MICROVASCULAR FLOWABNORMALITIES ASSOCIATED WITH RETINAL VASCULITIIS: A Potential of Mechanism of Retinal Injury) "," retina "2017, 37 (6), 1034-1042.doi:10.1097/iae.0000000000001359

Spaide, R.F., klancnik, J.M., jr., & Cooney, M.J. "retinal vascular layer (Retinal vascular layers imaged by fluorescein angiography and optical coherence tomography angiography) imaged by fluorescein angiography and optical coherence tomography angiography", "JAMA ophthalmoscope (JAMA ophtalmol) 2015, 133 (1), 45-50.Doi: 10.1001/jamaophtalmol.2014.3616)

Micro vascular abnormalities according to optical coherence tomography angiography in macular edema associated with retinal branch vein occlusion (Micro vascular Abnormalities on Optical Coherence Tomography Angiography in Macular Edema Associated With Branch Retinal Vein Occlusion) in Suzuki, n., hirano, Y., yoshida, m., tomiyasu, t., uemura, a., yasukawa, t., ogura, Y, journal of ophthalmology, 2016, 161,126-132.el21.doi:10.1016/j.ajo.2015.09.038

Wingegarer, A., wakabayashi, T., fukushima, Y., sato, T., hara-Ueno, C., busch, C., … Nishida, K, "change in retinal microvascular and visual acuity after anti-vascular endothelial growth factor treatment for retinal vein occlusion (Changes in Retinal Microvasculature and Visual Acuity After Antivascular Endothelial Growth Factor Therapy in Retinal Vein Occlusion)", ophthalmic research and visual mechanics 2018, 59 (7), 2708-2716.doi:10.1167/iovs.17-23437

Witt, n., wong, t.y., hughes, a.d., chaturvedi, n., klein, b.e., evans, r., … Klein, R "retinal microvascular structural abnormalities and mortality risk of ischemic heart disease and stroke (Abnormalities of retinal microvascular structure and risk of mortality from ischemic heart disease and stroke)", "Hypertension (2006, 47 (5), 975-981.Doi:10.1161/01.Hyp.0000216717.72048.6c

Yamada, R., nishida, A., shimezo no, M., kameda, T., miyamoto, N., mandai, M., kurimoto, Y. "prediction factor for recurrence of macular edema after intravitreal injection of bevacizumab for retinal branch vein occlusion (Predictive factors for recurrence of macular edema after successful intravitreal bevacizumab therapy in branch retinal vein occlusion)", "J.J.J.

Yoo, s.j., kim, j.h, lee, t.g., kim, j.w., cho, s.w., and Han, J.I, "Natural Short-term course of disease of recurrent macular edema after retinal branch vein occlusion (Natural Short-term Course of Recurrent Macular Edema Following Intravitreal Bevacizumab Therapy in Branch Retinal Vein Occlusion)" by intravitreal injection of bevacizumab, "korean journal of ophthalmology (Korean J Ophthalmol), 2017, 31 (2), 95-101.doi:10.3341/kjo.2017.31.2.95

Youm, D.J., ha, M.M., chang, Y., & Song, S.J, "risk factor for retinal vascular diameter and retinal branch vein occlusion (Retinal vessel caliber and risk factors for branch retinal vein occlusion)", "current ophthalmic study (CurrEyeRes)", 2012, 37 (4), 334-338.Doi:10 3109/02713683.2011.629070

Lemmens, sophie et al, "System evaluation of the fractal dimension of the retinal vessels of patients with neurodegeneration and stroke: evaluation of potential biomarkers (Systematic review on fractal dimension of the retinal vasculature in neurodegeneration and stroke: assessment of a potential biomarker) "," neuroscience front (Frontiers inneuroscience), "14 (2020): 16)

Claims

1. A method for automatic quantitative assessment of retinal microvasculature and generation of retinal digital vascular maps, comprising:

Recording Optical Coherence Tomography Angiography (OCTA) image data;

preprocessing OCTA image data; and

a Digital Vascular Map (DVM) of the retina is generated.

2. The method of claim 1, wherein the preprocessing comprises:

applying an early treatment diabetic retinopathy study grid to the DVM;

calculating fractal dimension index; and

retinal blood vessels were calibrated.

3. The method of claim 2, wherein the preprocessing further comprises:

calculating a vascular dispersion index;

separating the primary vessel section and the secondary vessel section;

calculating a blood vessel diameter index;

calculating a blood vessel curvature index;

calibrating a central concave avascular zone;

calculating the area index of the central concave avascular zone; and

and calculating the non-roundness index of the central concave avascular zone.

4. A method of extracting a Computational Retinal Microvascular Biomarker (CRMB) from a DVM generated using the method of claim 1.

5. The method of claim 4, wherein the CRMB is selected from the group consisting of fractal dimension index, capillary perfusion density index, large vessel perfusion density index, vessel dispersion index, vessel diameter index, vessel tortuosity index, foveal avascular zone area index, foveal avascular zone perimeter index, and foveal avascular zone non-roundness index.

6. A method for automatic quantification of a computational retinal microvascular biomarker of the retinal superficial capillary plexus from a subject suspected of having a retinal disease and treating the subject if CRMB quantification indicates that the subject has a retinal disease, the method comprising:

recording Optical Coherence Tomography Angiography (OCTA) image data;

preprocessing OCTA image data;

generating a retinal Digital Vascular Map (DVM) of the superficial capillary plexus;

providing retinal DVM in a normal subject;

treating a retinal disease in the subject if the amount of at least one CRMB in the subject exceeds the amount of the CRMB in the normal subject.

7. The method of claim 6, wherein the preprocessing comprises: applying an early-treatment diabetic retinopathy study grid to the DVM, and wherein the quantifying at least one CRMB comprises:

quantifying fractal dimension index;

quantifying a capillary perfusion density index;

quantifying a great vessel perfusion density index;

quantifying a vascular dispersion index;

Quantifying a vessel diameter index;

quantifying a vascular tortuosity index;

quantifying the area index of the fovea avascular zone;

quantifying the perivascular index of the foveal avascular zone

And quantifying the non-roundness index of the foveal avascular zone.

8. The method of claim 7, wherein the retinal disease is selected from diabetic retinopathy, retinal vein occlusion, and age-related macular degeneration, and the treatment comprises administering antibodies to vascular growth factors, aptamers to vascular growth factors, anti-angiogenic steroids, antioxidant supplements, laser photocoagulation therapy, transpupillary thermotherapy, and/or performing ocular surgery.

9. A method for automatic quantification of a computational retinal microvascular biomarker of the deep retinal capillary plexus from a subject suspected of having a retinal disease and treating the subject if CRMB quantification indicates that the subject has a retinal disease, the method comprising:

recording Optical Coherence Tomography Angiography (OCTA) image data;

preprocessing OCTA image data;

generating a retinal Digital Vascular Map (DVM) of the deep capillary plexus;

retinal DVM providing a deep capillary plexus of a normal subject;

10. The method of claim 9, wherein the preprocessing comprises: applying an early-treatment diabetic retinopathy study grid to the DVM, and the quantifying at least one CRMB comprises:

quantifying fractal dimension index;

quantifying a capillary perfusion density index;

quantifying a great vessel perfusion density index;

quantifying a vascular dispersion index;

quantifying a vessel diameter index;

quantifying a vascular tortuosity index;

quantifying the area index of the fovea avascular zone;

quantifying the perivascular index of the foveal avascular zone

And quantifying the non-roundness index of the foveal avascular zone.

11. The method of claim 9, wherein the retinal disease is selected from diabetic retinopathy, retinal vein occlusion, and age-related macular degeneration, and the treatment comprises administering antibodies to vascular growth factors, aptamers to vascular growth factors, anti-angiogenic steroids, antioxidant supplements, laser photocoagulation therapy, transpupillary thermotherapy, and/or performing ocular surgery.

12. A method for automatic quantification of a computational retinal microvascular biomarker of the retinal superficial capillary plexus from a subject suspected of having an extra-retinal disease and treating the subject if CRMB quantification indicates that the subject has an extra-retinal disease, the method comprising:

recording Optical Coherence Tomography Angiography (OCTA) image data;

preprocessing OCTA image data;

providing retinal DVM in a normal subject;

treating a subretinal disease in the subject if the amount of at least one CRMB in the subject exceeds the amount of the CRMB in the normal subject.

13. The method of claim 12, wherein the preprocessing comprises: applying an early-treatment diabetic retinopathy study grid to the DVM, and the quantifying at least one CRMB comprises:

quantifying fractal dimension index;

quantifying a capillary perfusion density index;

quantifying a great vessel perfusion density index;

Quantifying a vascular dispersion index;

quantifying a vessel diameter index;

quantifying a vascular tortuosity index;

quantifying the area index of the fovea avascular zone;

quantifying the perivascular index of the foveal avascular zone

And quantifying the non-roundness index of the foveal avascular zone.

14. The method of claim 13, wherein the extra-retinal disease is selected from cardiovascular disease, systemic arterial hypertension, sickle cell anemia, preeclampsia, arteritis, aortic aneurysm, diabetes and/or hypercoagulability, and the treatment comprises administration of antihypertensive drugs, cholesterol-modulating drugs, aspirin, beta blockers, calcium channel blockers, nitroglycerin, angiotensin converting enzyme inhibitors, angiotensin II receptor blockers, blood-thinning drugs; iron supplements, steroids, antidiabetic agents, intravascular stent surgery and/or vascular bypass surgery.

15. An automated analysis system for Optical Coherence Tomography Angiography (OCTA) image analysis, comprising:

a memory configured to store one or more executable instructions;

a processor configured to implement the steps of the method according to any one of claims 1 to 14 by executing the one or more executable instructions.

16. A computer readable storage medium containing a computer program and executable by a processor to implement the steps of the method according to any one of claims 1 to 14.