Markers for use in identifying or predicting angiogenesis

Abstract

The present invention provides a labeled cellular stress marker for use in identifying or predicting angiogenesis in a subject. Also provided is a method of identifying or predicting angiogenesis in a subject, wherein the subject has been administered a labeled cellular stress marker. The method includes (a) generating an image from the subject using a first imaging device and (b) identifying labeled cellular stress marker-positive cells in the image. If labeled cellular stress marker-positive cells are identified in the image, angiogenesis is identified or predicted.

Description

Technical Field

[0001] The present invention relates to labeled cell stress markers for use in identifying or predicting angiogenesis in a subject. The present invention also relates to a method for identifying or predicting angiogenesis in a subject.

Background Art

[0002] Angiogenesis is an important pathological condition associated with cancer, allergic diseases, autoimmune diseases, and eye diseases such as age-related macular degeneration (AMD), particularly wet AMD. Wet AMD is a major cause of visual impairment and affects 2.5% of people over 65 years old in the UK. This figure rises to 6.3% in people over 80 years old in the UK.

[0003] Currently, the diagnosis of wet AMD requires taking a photograph of the retina to evaluate the occurrence of abnormal blood vessels. However, since wet AMD progresses rapidly without symptoms, it is not uncommon for diagnosis and treatment to be too late to be effective, which can lead to serious diseases.

[0004] It is necessary to identify subjects at risk of developing wet AMD so that treatment can be carried out at an early stage. The present invention aims to address such problems and other problems.

Summary of the Invention

[0005] According to a first aspect of the present invention, there is provided a labeled cell stress marker for use in identifying or predicting angiogenesis in a subject.

[0006] As used herein, the term “cellular stress marker” refers to a marker that identifies cells undergoing cellular stress. Therefore, a marker may be known as a compound or molecule that specifically binds to stressed cells. Cellular stress may predict cells that are ultimately undergoing apoptosis. In some embodiments, cellular stress markers may also label apoptotic cells (i.e., markers may label not only apoptotic cells but also cells undergoing cellular stress). Therefore, in embodiments, cellular stress markers may be referred to herein as “apoptotic markers.”

[0007] As used herein, the term "apoptotic marker" refers to a marker that can distinguish apoptotic cells from living cells. This marker may also distinguish apoptotic cells from necrotic cells. Therefore, the marker may be known as a compound or molecule that specifically binds to apoptotic cells.

[0008] Cellular stress markers may include or consist of markers that identify stressed cell membranes.

[0009] The term "angiogenesis" is a known term in the art, referring to the growth of new blood vessels from the existing vascular system. Conveniently, the inventors have discovered that labeled cellular stress markers can be used to identify stressed cells at the single-cell level and predict or identify angiogenesis. In some embodiments, the angiogenesis is in its early stages. By identifying angiogenesis in its early stages or predicting its development, it may be possible to clinically target the suppression or promotion of angiogenesis in its early stages, ensuring the greatest possible effect for the patient.

[0010] "Predicting angiogenesis" can be understood as predicting the occurrence of angiogenesis before it actually happens. This allows for unexpected early clinical intervention, which in turn can prevent the rapid progression of symptoms such as vision loss in wet AMD.

[0011] It was completely unexpected that cellular stress markers could identify or predict angiogenesis, because angiogenesis is related to the proliferation of new cells, not to cellular stress and / or apoptosis.

[0012] In some embodiments, the cell stress marker is annexin or a fragment or variant thereof. The annexin family of proteins is known as apoptosis markers. Annexins are proteins that reversibly bind to the cell membrane in the presence of cations. The annexins useful in this invention may be natural or recombinant.

[0013] A “fragment” would be understood to refer to a cleaved version of annexin that substantially retains the functional activity of the whole annexin. “Substantially retains” would be understood to refer to at least 80%, at least 85%, at least 90%, at least 95%, at least 97%, at least 98%, or at least 99% of the functional activity of the whole annexin. An annexin variant would be understood to refer to a protein that has a different amino acid sequence from wild-type annexin so as to have at least 80%, at least 85%, at least 90%, at least 95%, at least 97%, at least 98%, or at least 99% sequence identity with wild-type annexin. An annexin variant would substantially retain the functional activity of wild-type annexin.

[0014] Annexin may be annexin 5, 11, 2, or 6. In some embodiments, annexin is annexin 5. In another embodiment, annexin is annexin 128 (Tait et al., 2005). Annexin 128 is a variant of annexin 5 and differs from wild-type annexin 5 by two single amino acid mutations. Annexin 128 contains an exposed thiol group at the N-terminus, which improves the binding efficiency of molecular tags such as fluorescent labels.

[0015] Without being constrained by theory, the inventors hypothesize that phosphatidylserine, to which annexin can bind, is exposed in stressed endothelial cells. Therefore, for example, the identification of phosphatidylserine by annexin binding may indicate cellular stress. Surprisingly, the inventors discovered that the exposure of phosphatidylserine on the cell membrane may indicate the subsequent progression of early-stage angiogenesis.

[0016] For example, other apoptosis markers known in the art include the C2A domain of synaptotagmin-I, non-peptide isatinsulfonamide analogs such as duramycin and WC-11-89, and ApoSense such as NST-732, DDC, and ML-10 (Saint Hubert et al., 2009). Such apoptosis markers are suitable for use as cell stress markers in the present invention.

[0017] The labeling of cellular stress markers may be visible. Exemplary visible labels may include, but are not limited to, quantum dots, nanospheres, and / or nanorods. Other visible labels will be known to those skilled in the art.

[0018] The label may be fluorescent. A fluorescent label refers to a compound or molecule (such as a fluorescent dye molecule) that emits light in response to excitation and may be selected for use because it improves the signal-to-noise ratio, thereby improving image resolution and sensitivity, while adhering to exposure safety standards to avoid the effects of phototoxicity. It is preferable that the fluorescent label causes little to no inflammation when administered. The fluorescent label may have wavelengths in the infrared or near-infrared spectrum. The fluorescent label may have emission wavelengths of about 400 nm to about 1000 nm, preferably about 500 nm to about 900 nm, and more preferably about 700 nm to about 900 nm.

[0019] Suitable fluorescent labels include one or more of the following: sodium fluorescein, indocyanine green (ICG), curcumin, IRDye700, Dy-776 (sometimes referred to as D-776), Dy-488, and Dy-781. In some embodiments, the fluorescent label is D-776.

[0020] Labeled cell stress markers may be prepared using standard techniques for conjugating labels, such as fluorescent labels, to marker compounds. Such labels may be available from well-known sources such as Dyomics. Appropriate techniques for conjugating labels to markers are known in the art and may be provided by label manufacturers.

[0021] It will be understood that angiogenesis (optionally, early-stage angiogenesis) can be associated with pathological conditions such as disease, as with normal growth and development. Therefore, angiogenesis may be physiological, therapeutic, or pathological. In the context of this invention, physiological angiogenesis will be understood to refer to angiogenesis associated with normal growth and healing in the subject. For example, it has been previously shown that angiogenesis can be induced by exercise, particularly endurance exercise at high altitudes, such as mountain climbing. Therefore, physiological angiogenesis may be associated with prior exercise by the subject. Therapeutic angiogenesis may be associated with wound healing and / or the integration of tissue engineering scaffolds.

[0022] In some embodiments, angiogenesis is pathological angiogenesis. Pathological angiogenesis would be understood as angiogenesis related to or associated with a disease. Pathological angiogenesis may be related to the disease, the stage of the disease, the severity of the disease, and / or prediction of disease onset.

[0023] Cellular stress markers may be used for the identification or prediction of pathological angiogenesis. Identification or prediction of pathological angiogenesis by cellular stress markers may predict or identify a specific disease, disease stage, and / or disease severity. Advantageously, the ability to predict a specific disease, disease stage, and / or disease severity, or to identify a specific disease in its early stages, allows for the selection, administration, and optional monitoring of the most appropriate course of treatment.

[0024] In the context of disease, the term “predict” as used herein may be understood to mean predicting the disease, the progression of the disease stage and / or the severity of the disease, pre-symptomatically, or in the early stages of the disease when the subject is asymptomatic and / or the disease is not otherwise identifiable. In embodiments in which early-stage angiogenesis is identified, this may be used to identify the disease in the early stages. For example, early-stage angiogenesis may be identified when the subject has no other symptoms of the disease, or has only limited or mild symptoms of the disease, and is in the early stages of the disease where diagnosis is difficult or impossible by other means. This is useful in enabling clinicians to select and administer treatments in the early stages, thereby potentially avoiding unnecessary progression of the disease.

[0025] The disease may include any disease related to angiogenesis. For example, the disease may include any disease in which anti-VEGF can be administered as treatment. In some embodiments, the disease includes cancer, autoimmune diseases, cardiovascular diseases, allergic diseases, sickle cell disease, and / or eye diseases. It will be understood that some diseases may be, for example, eye diseases and autoimmune diseases (such as diabetic retinopathy).

[0026] In some embodiments, angiogenesis is associated with cardiovascular disease. In other embodiments, angiogenesis is associated with effective treatment of cardiovascular disease (i.e., therapeutic angiogenesis). Examples of cardiovascular diseases associated with angiogenesis include venous occlusion, such as stroke.

[0027] Various autoimmune diseases are well known to those skilled in the art. They include, but are not limited to, sarcoidosis, rheumatoid arthritis, multiple sclerosis (MS), psoriasis, diabetes (especially type I diabetes), ulcerative colitis, systemic lupus erythematosus (lupus), Guillain-Barré syndrome, chronic inflammatory demyelinating polyneuropathy, Graves' disease, Hashimoto's thyroiditis, myasthenia gravis, autoimmune encephalitis, neuromyelitis optica, and myelin oligodendrocyte glycoprotein antibody disease (MOG).

[0028] In some embodiments, the autoimmune disease is a central nervous system (CNS) autoimmune disease. Exemplary CNS autoimmune diseases include, but are not limited to, neurosarcoidosis, MS, neuromyelitis optica, and myelin oligodendrocyte glycoprotein antibody disease (MOG).

[0029] Allergic diseases may include or consist of rhinitis, asthma, and / or hay fever. Typically, allergic diseases are chronic allergic diseases. It has been found that chronic allergic diseases can be associated with angiogenesis. "Chronic allergic disease" will be understood to refer to a long-term and recurrent allergic disease. Long-term may refer to a condition that persists and / or recurs for at least 6 months, at least 1 year, at least 2 years, at least 5 years, at least 10 years, or at least 20 years.

[0030] In some embodiments, the disease includes ocular diseases. Ocular diseases include, but are not limited to, retinal vein occlusion, diabetic retinopathy, age-related macular degeneration (AMD), optic neuritis, and ocular vascular diseases. In some embodiments, the disease includes ocular neurodegenerative diseases. The term “ocular neurodegenerative disease” is well known to those skilled in the art and refers to diseases caused by the gradual progression and loss of ocular nerve cells. These include, but are not limited to, AMD, optic neuritis, and diabetic retinopathy. Neurodegenerative diseases include, for example, Parkinson’s disease, Alzheimer’s disease, Huntington’s disease, and Friedreich’s ataxia.

[0031] In some embodiments, the eye disease includes or comprises diabetic retina or AMD.

[0032] In some embodiments, the eye disease includes or comprises AMD. Optionally, the eye disease includes or comprises wet AMD. Surprisingly, the inventors have found that angiogenesis in a subject can be identified or predicted using labeled cellular stress markers, and that the identification or prediction of angiogenesis is a predictive marker for wet AMD. This allows clinicians to predict whether a subject will develop wet AMD. The symptoms of wet AMD progress very rapidly and can lead to permanent vision loss. Therefore, such prediction is particularly beneficial because it allows clinicians to treat the subject in an early stage, thereby mitigating vision loss and, in some embodiments, avoiding vision loss.

[0033] In some embodiments, the subject is a mammal. For example, the subject may be a rat, mouse, human, rabbit, horse, dog, cat, cow, or sheep. In some embodiments, the subject is selected from mice, rats, humans, and rabbits. In some embodiments, the subject is a human.

[0034] A second aspect of the present invention provides the use of labeled cellular stress markers for identifying or predicting angiogenesis in a subject. The angiogenesis may be physiological angiogenesis or therapeutic angiogenesis, and is preferably physiological angiogenesis as described herein.

[0035] A third aspect of the present invention provides a method for identifying or predicting angiogenesis in a subject, the subject being administered a labeled cellular stress marker, comprising the following steps: (a) A step of generating a first image from the object using a first imaging device; (b) Identifying labeled cell stress marker-positive cells in the image; If labeled cells positive for cellular stress markers are identified in the image, angiogenesis is identified or predicted.

[0036] In a further embodiment of the present invention, a method for identifying or predicting angiogenesis in a subject is provided, comprising the following steps: (a) A step of administering a labeled cell stress marker to the target; (b) A step of generating an image from the object using an imaging device; (c) Identifying labeled cell stress marker-positive cells in the image; If labeled cells positive for cellular stress markers are identified in the image, angiogenesis is identified or predicted.

[0037] The inventors have discovered that cellular stress can be labeled at the single-cell level using labeled cellular stress markers. Identifying cellular stress at the single-cell level is advantageous because it allows for the prediction of subsequent angiogenesis and / or detection in the early stages of angiogenesis, thereby enabling disease prediction or identification before other or further symptoms develop.

[0038] The administration of labeled cell stress markers may be intravenous, intranasal, topical, or intravitreous. Topical administration is understood to refer to application to the surface of the target, such as the skin or the surface of the eye. In some embodiments, the labeled cell stress markers are administered intravenously. In other embodiments, the labeled cell stress markers are administered intravitreously.

[0039] In some embodiments, the method further includes administering an appropriate treatment for a disease, disease stage, disease severity, or predicted disease identified based on angiogenesis or predicted angiogenesis. The appropriate treatment may include or consist of anti-angiogenic agents (i.e., treatments that suppress angiogenesis), such as anti-VEGF compounds. Exemplary anti-angiogenic agents may include, but are not limited to, axitinib, bevacizumab, ranibizumab, cabozantinib, everolimus, lenalidomide, pazopanib, ramucirumab, regorafenib, sorafenib, sunitinib, salomid, vandetanib, and aflibercept. The appropriate treatment may be administered intravenously, intranasally, topically, or intravitreally.

[0040] In some embodiments, identifying labeled cellular stress marker-positive cells involves counting the number of labeled cellular stress marker-positive cells in an image. Advantageously, this step does not need to be repeated. In other words, the inventors have surprisingly found that only one identification step is required to identify angiogenesis or predict its subsequent development, and therefore the step of identifying labeled cellular stress marker-positive cells needs to be performed only once. Thus, the method of the present invention is particularly rapid, accurate, and cost-effective.

[0041] In some embodiments, identifying labeled cellular stress marker-positive cells includes determining the location of labeled cellular stress marker-positive cells in a subject. This may include determining the anatomical location of labeled cellular stress marker-positive cells shown in an image. As described above, the cellular stress markers of the present invention may label both stressed and apoptotic cells. Therefore, it may be desirable to distinguish between apoptotic cells and stressed cells. The inventors have found that one way to achieve this is to examine the location of labeled cells and consider whether they are in a place where angiogenesis is likely to occur, e.g., near blood vessels, or in a place where angiogenesis is unlikely to occur, e.g., in the photoreceptor layer of the retina. Therefore, the method of the present invention may further include a step of determining the location of cellular stress marker-positive cells, the location of which indicates whether the cell is involved in angiogenesis or apoptosis. This ensures that the identification and / or prediction of angiogenesis is particularly accurate, as physicians can reduce false positives from apoptotic-positive cells. The step of determining the location of cellular stress marker-positive cells may include identifying the location of cells relative to anatomical structures or tissue types, such as blood vessels.

[0042] The positioning step may include locating the cell in a first generated image, or locating the cell in a second image, particularly an image showing the cell from a different perspective, or an image showing tissue or other structures not visible in the first image. Optionally, the method includes determining the position of labeled cell stress marker-positive cells in multiple planes. For example, the position of labeled cell stress marker-positive cells may be determined across the x and y axes in a 2D image, for example. The position of labeled cell stress marker-positive cells may be determined across the x, y, and z axes. In some embodiments, the cell position may be determined from multiple images. For example, the cell position may be determined using at least a second and a third image. Each of the multiple images may be taken using similar or identical reference frames, and after the multiple images are obtained, they may be superimposed to form the x, y, and z axes. In other words, a 3D image is formed. The 3D image may be used to locate the cell.

[0043] This method may include, for example, the step of generating a second image of the target using the same or a different imaging device, and optionally the step of comparing that image with the first image. If this method includes comparing one or more images, it is preferable to overlay one image on the other. This method may include a step of doing so. It may also include an image registration step, i.e., an image alignment step, in order to correctly match the cell stress marker-positive cells shown in each image.

[0044] If this method involves the use of two or more images, those images may be obtained using the same or different imaging devices. For example, the first imaging device may include or consist of a confocal scanner laser microscope. An example of a confocal scanner laser microscope is a confocal scanning laser ophthalmoscope (cSLO).

[0045] In some embodiments, first and second imaging devices are used to acquire multiple images. The second imaging device may include or consist of a computed tomography (CT) scanner, a magnetic resonance imaging (MRI) scanner, a positron emission tomography (PET) scanner, a polarizer, a reflective image scanner, a fundus camera, and / or an optical coherence tomography (OCT) scanner. Preferably, the second imaging device is different from the first imaging device.

[0046] In some embodiments, the second imaging device includes or comprises a computed tomography (CT) scanner, a magnetic resonance imaging (MRI) scanner, a positron emission tomography (PET) scanner, and / or an optical coherence tomography (OCT) scanner.

[0047] In some embodiments, the second imaging device includes or comprises an OCT scanner.

[0048] In some embodiments, first, second, and third imaging devices are used to acquire multiple images. The third imaging device may include or consist of a computed tomography (CT) scanner, a magnetic resonance imaging (MRI) scanner, a positron emission tomography (PET) scanner, a polarizer, a reflective image scanner, a fundus camera, and / or an optical coherence tomography (OCT) scanner. Preferably, the third imaging device is different from the first and second imaging devices.

[0049] The first imaging device may be used to acquire 2D or 3D images. Optionally, the first imaging device is used to acquire 2D images. The second imaging device may be used to acquire 2D or 3D images. The second imaging device is used to acquire 3D images. In some embodiments, the second or third imaging device is used to provide a mask image. A mask image will be understood to mean a “background” image, i.e., an image that can be used to set the background value of an image used to locate cells.

[0050] In some embodiments, the third imaging device includes or comprises a reflective image scanner.

[0051] In one embodiment, the method includes determining the location of one or more blood vessels. The location of the blood vessels may be determined by angiography, a method known in the art. For example, the location of the blood vessels may be determined using fluorescein angiography (FA), indocyanine green angiography (ICGA) and / or optical coherence tomography (OCT), and optionally optical coherence tomography angiography (OCTA).

[0052] In some embodiments, determining the location of labeled cellular stress marker-positive cells in an image may include determining the location of the cells relative to the location of blood vessels. For example, if labeled cellular stress marker-positive cells are identified as being adjacent to or on a blood vessel, this may be used to identify cellular stress rather than apoptosis. This may be used as a positive indicator of early-stage angiogenesis or predicted subsequent angiogenesis.

[0053] For example, in embodiments where the identification or prediction of angiogenesis is used to predict whether a subject is likely to develop an ophthalmic neurodegenerative disease, the determination of the location of labeled cellular stress marker-positive cells (optionally further including determining their location relative to blood vessels) may be used to predict whether a subject will develop wet AMD rather than glaucoma. This is because wet AMD is represented by cellular stress, while glaucoma is associated with apoptosis. This is particularly beneficial because both wet AMD and glaucoma are serious, rapidly progressing eye diseases requiring different treatments. If the two diseases can be distinguished before symptoms appear or in the early stages of the disease, the correct treatment can be selected and implemented to avoid unnecessary vision loss.

[0054] The identification of labeled cellular stress marker-positive cells, and optionally their location in the image, may be performed by computer. This may include generating training data for use in training a discriminative model, such as an artificial neural network. The artificial neural network may include a convolutional neural network (CNN). The training data may include control data manually classified as containing or not containing labeled cellular stress marker-positive cells from subjects already known to have (or not have) a disease, such as wet AMD.

[0055] In some embodiments, the discriminative model may be trained using training data. This may be in addition to, or instead of, the step of generating training data for use in training the discriminative model. Once trained, the discriminative model may be manually validated.

[0056] A trained discrimination model (e.g., a convolutional neural network) may be used to identify the locations of labeled cellular stress marker-positive cells and optionally labeled cellular stress marker-positive cells.

[0057] Alternatively, the identification of labeled cell stress marker-positive cells, and optionally the location of labeled cell stress marker-positive cells in the image, may be performed manually.

[0058] In some embodiments, the identification of at least 5, at least 10, at least 15, at least 20, at least 30, or at least 40 labeled cellular stress marker-positive cells in an image predicts that the subject will develop the disease. In some embodiments, the identification of 5 or more, 10 or more, 15 or more, 20 or more, 30 or more, or 40 or more labeled cellular stress marker-positive cells in an image predicts that the subject will develop the disease. For example, the identification of at least 5, at least 10, at least 15, at least 20, at least 30, or at least 40 labeled cellular stress marker-positive cells in an image of a subject's eye may predict that the subject will develop wet-type AMD. In some embodiments, the identification of at least 5, optionally 5 or more labeled cellular stress marker-positive cells in an image of a subject's eye predicts that the subject will develop wet-type AMD.

[0059] Advantageously, the inventors have also found that an increase in the number of labeled cellular stress marker-positive cells identified correlates with the severity of the disease that subsequently develops; that is, the more labeled cellular stress marker-positive cells identified in the image, the more severe the disease that subsequently develops. Therefore, in some embodiments, the identification of at least 10, at least 15, at least 20, at least 30, or at least 40 labeled cellular stress marker-positive cells in the image predicts an increase in the severity of the disease in question. This can be advantageously used to determine whether the subject requires higher doses of treatment, more frequent treatment, or entirely different treatment.

[0060] In some embodiments, the step of generating an image is repeated to obtain second, third, fourth, fifth, or more additional images. The images may be compared with one or more prior images of the same region of the subject, for example, the same eye of the subject. The repetition of image generation may be monthly, every two months, every six months, annually, or biennially (every two years). The method may also include overlaying images on one, two, three, four, or more prior images. Comparing images of the same region of the subject over time may be used to monitor disease progression. Alternatively, in embodiments where appropriate treatment has been administered, comparing images of the same region of the subject over time may be used to monitor the effectiveness of the administered treatment. For example, after generating the initial image, the subject may be given appropriate treatment. The images obtained thereafter may be used to monitor the effectiveness of the treatment.

[0061] In some embodiments, the method may be performed on subjects to whom another method has already been performed in order to diagnose or predict the likelihood of the subject developing a particular disease. Thus, the method may be used as a “second opinion” to test and / or improve the accuracy of the other (first) method and to reliably identify false negatives by the first method.

[0062] In another aspect of the present invention, a labeled apoptosis marker is provided for use in identifying apoptosis in a body part of an object other than the eye. In one embodiment, the body part is not part of the central nervous system. The body part may be all or part of a limb (e.g., an arm or a leg), a liver, a spleen, a heart, one or more lungs, a stomach, a small intestine, a large intestine, a kidney, skin, or a bladder.

[0063] Throughout this specification and in the claims, the words “comprise” and “contain,” and their variations, such as “comprising” and “comprises,” mean “comprise but not limited to,” and do not exclude other components, integers, or steps. Furthermore, the singular form includes the plural form unless otherwise specified in the context; in particular, where the indefinite article is used, unless otherwise specified in the context, this specification is understood to assume both singular and plural forms.

[0064] Preferred features of each aspect of the present invention may be as described in relation to any of the other aspects. Within the scope of this application, the various aspects, embodiments, examples and alternatives, in particular their individual features, described in the preceding paragraphs, claims and / or the following description and drawings are expressly intended to be taken independently or in any combination. That is, all embodiments and / or features of any embodiment may be combined in any way and / or in any combination, provided that such features are not incompatible.

[0065] Next, one or more embodiments of the present invention will be described only illustratively with reference to the accompanying drawings. [Brief explanation of the drawing]

[0066] [Figure 1]This study demonstrates that a DARC CNN system can predict new subretinal fluid (wet AMD) in OCT. Alignment Application: For each eye, a reflective OCT localization map (A) was aligned to baseline autofluorescence images (B) and OCT scans (C) acquired with a confocal scanning laser ophthalmoscope (cSLO). A custom application was used to manually place automated landmarks (filled white circles) and fiducial markers (yellow and blue crosses, A and B) defining the OCT region in a series of scans. The DARC image (D) was acquired 240 minutes after intravenous injection of Anx776 with cSLO, automatically aligned to the baseline autofluorescence image (B), and also aligned to the OCT images (A and C). Points corresponding to the DARC image can be identified within the OCT scan (C) using cyan vertical lines, allowing identification of all cross-sectional structures within the OCT image at that vertical position. Therefore, in this example, at baseline, the location of the yellow DARC spot (D) surrounded by cyan can be recorded in x,y coordinates across all retinal layers of the baseline OCT scan (C). This can be done for all DARC spots, and the corresponding points are placed in all follow-up OCT scans. DARC spot localization: DARC spots are automatically detected using the CNN-assisted algorithm already described (Lindekleiv et al. (2013)). The yellow DARC spot (D) surrounded by cyan is located on the baseline (C) and aligned to the follow-up OCT scan (E) at 36 months. The presence of new subretinal fluid (SRF) is automatically identified by the SRF CNN and outlined in the annotated image at 36 months (purple region, F). This indicates that the yellow DARC spot (D) surrounded by cyan predicts new SRF at 36 months (E, F). Unique DARC spot prediction: Individual DARC spots observed at baseline can be evaluated in relation to a series of OCT scans performed during follow-up.Unique DARC spots associated with SRF(D) are surrounded by different colors representing different time points: 12–18 months (red, D, G), 18–24 months (light green, D, L), 24–30 months (purple, D, Q), and 30–36 months (yellow, D). These are shown in the OCT scans corresponding to follow-up (H, M, R) and baseline (J, O, T), and the SRF CNN annotates newly formed SRF regions (I, N, S) with the same color coding compared to baseline (K, P, U). The cyan lines in each OCT scan highlight the vertical cross-sections passing through the indicated DARC spots (white arrows; G, L, Q). [Figure 2] This study shows that a CNN DARC count > 5 is associated with an increased risk of developing SRF (wet AMD) in OCT at 36 months or longer. (A) This plot shows the percentage of eyes that did not develop SRF (did not develop wet AMD) during 36 months, grouped by CNN DARC count > 5 (red) or ≤ 5 (blue). Eyes with a CNN DARC count > 5 had a significantly higher incidence of SRF compared to eyes with a CNN DARC count ≤ 5 (Wilcoxon rank, p=0.0156). (B) This table shows the breakdown of converted eyes by baseline diagnosis. All 7 eyes that developed new SRF had a CNN DARC count > 5, and 16 of the 17 eyes (n=10) that had active choroidal neovascularization (CNV) at baseline or the 17 eyes (n=7) that converted during follow-up also had a CNN DARC count > 5. [Figure 3]This study shows that CNN DARC counts significantly increased in eyes with larger SRF accumulation areas. The violin plots show the data distribution of unique DARC spots covering SRF on OCT and the total SRF accumulation area on OCT at each time point. To detect CNV lesions on OCT, SRF regions were identified using a CNN trained on data from The Retinal OCT Fluid Challenge (RETOUCH). The average SRF area was calculated for each interval, taking into account the number of OCT scans and interocular variability at 6-month intervals. Eyes were divided into "high SRF" and "low SRF" using a 3000 μm threshold. At all time points, a DARC count of 5 or higher in all eyes was associated with large SRF accumulation. All p-values ​​represent the Mann-Whitney significance level. Horizontal lines show the median and interquartile range, minimum and maximum values. [Figure 4]This study demonstrates that DARC identifies the earliest changes in endothelial activity in a rabbit angiogenesis model. An angiogenesis model was created by intravitreal administration of 1 μg of human VEGF (hVEGF165-Peprotech, London UK) in 50 μl to only the left eye of three rabbits. ANX776 (0.2 mg) and 1% fluorescein sodium were administered intravenously at 2 and 4 days. Both eyes of the rabbits were imaged using cSLO ICGA and FFA settings. DARC spots (identified as white dots positively labeled with ANX776) were clearly visible in the hVEGF-treated left eye (AC) compared to the untreated right eye (DF). DARC counts were obtained by three masked manual observers. (G) Individual dots were manually observed separately in the left and right eyes. All hVEGF-treated eyes were positive for DARC staining compared to the untreated contralateral eye. There was a significant difference in DARC counts between the treated eye and the control eye (p=0.0203). Inset (AF) shows fluorescein angiography 4 days after hVEGF administration. All treated eyes showed fluorescein leakage at 4 days compared to the control (inset DF) (inset AC). Local administration of ranibizumab reduced angiogenic DARC labeling (HJ). In AF, there was a significant difference in DARC counts between the treated eye and the control eye, but this significant difference disappeared with ranibizumab treatment (N). Local administration of ranibizumab significantly reduced DARC counts compared to the vehicle control (O, p=0.0262). [Figure 5] This study demonstrates that anti-angiogenic therapy reduced vascular leakage in a rabbit angiogenesis model. A rabbit angiogenesis model was created in the same manner as in Figure 4. The anti-angiogenic drug ranibizumab was administered locally. All control eyes in rabbits treated with hVEGF showed fluorescein leakage on day 4 compared to ranibizumab eyes (GI) (AC). [Modes for carrying out the invention] [Examples]

[0067] background Age-related macular degeneration (AMD) is the leading cause of blindness in developed countries, and its incidence is projected to rise with the increasing aging population. AMD has two subtypes, classified as either having choroidal neovascularization (CNV) (wet type) or not (dry type), and is characterized by the formation of subretinal or intraretinal fluid (SRF or IRF, respectively) from leaking blood vessels. Optical coherence tomography (OCT) is currently considered the gold standard for the management of AMD, particularly for evaluating treatments targeting anti-vascular endothelial growth factor (anti-VEGF).

[0068] Detection of Apoptotic Retinal Cells (DARC), a technique for monitoring the rate of cell death in the retina, recently completed a Phase I clinical trial for glaucoma diagnosis (Cordeiro et al., 2017). This technique currently involves intravenous administration of a novel fluorescent agent called Anx776, a modified endogenous protein annexin A5 conjugated with the near-red fluorescent agent Dy-776. Anx776-positive labeled cells are visualized as "white dots" on the retina using the ICGA setting of a commonly used confocal scanning laser ophthalmoscope (cSLO). DARC leverages the unique optical properties of the eye, enabling direct observation of apoptosis in single neurons of patients using a fluorescently labeled human annex V derivative. We decided to study DARC as a technique for predicting the onset of disease, particularly wet-type AMD (i.e., before diagnosis is possible due to insufficient symptoms).

[0069] Materials and methods The binding of D776 to Annexin 5 Recombinant bacterial-expressed annexin 5 (5 mg in 1 ml of isotonic buffer) was dialyzed over 3 hours in the "labeled buffer," with two exchanges (250 ml each). The labeled buffer was prepared by dissolving 21 g (250 mmol) of sodium bicarbonate in 400 ml of distilled water. 1 g (0.2%) of sodium azide was added. The pH was adjusted to pH 9.0 with a concentrated aqueous solution of sodium hydroxide. Before use, the buffer was diluted by adding 9 parts water to 1 part of the concentrated stock solution (v / v). Subsequently, the annexin 5 protein was ready for the labeling reaction. The ratio of dye to protein was kept constant in each labeling reaction. Therefore, 0.5 mg of D776-NHS-ester was used to label 1 mg of annexin 5. This resulted in a molar ratio of approximately 500 of dye to annexin 5 protein. The NHS-ester was dissolved in dimethylformamide by vortexing, and the dye was added to the annexin 5 protein in the labeled buffer to initiate the reaction. The final concentration of annexin 5 in the labeling reaction was 5 mg / ml. The Eppendorf tubes were placed in a shaker and labeling was performed over 2 hours.

[0070] After the reaction was complete, the colored labeling solution was carefully pipetteed into 5 or 10 ml of Sephadex G-25 column and permeated into the gel. Next, 100 mM PBS buffer pH 7.4 was slowly added to the column to elute the conjugate. The labeled protein flowed forward as a relatively sharp band, while the free dye flowed slowly backward. Once the bound annexin 5 reached the bottom of the column, it was collected in an Eppendorf tube and ready for immediate use.

[0071] Clinical trial design A single-center, open-label phase 2 clinical trial of DARC was conducted at The Western Eye Hospital, Imperial College Healthcare NHS Trust. All participants received a single intravenous injection of fluorescent annexin 5 (ANX776, 0.4 mg) between February 15, 2017, and June 30, 2017. The trial recruited AMD patients already receiving treatment at the Department of Retina at Western Eye Hospital, and informed consent was obtained in accordance with the Declaration of Helsinki after approval by the Brent Research Ethics Committee (ISRCTN10751859). All participants underwent DARC only once at baseline and underwent optical coherence tomography (OCT) scans at each follow-up appointment in accordance with standard care for up to 36 months after DARC.

[0072] Patients were considered to participate in this study if they met the inclusion and exclusion criteria shown in Table 1. Specifically, to be included in the AMD cohort, patients needed to show changes consistent with CNV (wet AMD), "early AMD" (drusen, retinal pigment epithelial (RPE) pigment changes), or "late AMD" (geographic atrophy of the RPE (dry AMD)). In addition, the presence of ocular diseases that increase the risk of choroidal neovascularization (CNVM), and the current or past 30 days or more of use of chloroquine, hydroxychloroquine, chlorpromazine, thioridazine, quinine sulfate, clofazimine, cisplatin, carmustine, (BCNU), diferoxamine, amiodarone, isoletinoin, or gold were excluded. Eyes were classified at baseline into active CNV, history of CNV, early AMD, and late AMD based on clinical examination and retinal imaging findings.

[0073] Of the 19 AMD subjects, 31 eyes were analyzed, and images at baseline and 240 minutes were evaluated.

[0074] DARC image acquisition and blinding of trials Retinal images were acquired using cSLO (HRA+OCT Spectralis, Heidelberg Engineering GmbH, Heidelberg, Germany), as previously described by Normando et al. (2020). Baseline infrared autofluorescence images were acquired after pupil dilation (1% tropicamide and 2.5% phenylephrine) using high-resolution ICGA mode fluorescence settings (786nm diode laser excitation; photodetector with 800nm ​​barrier filter). All participants underwent DARC only once at baseline. Each patient received an intravenous injection of ANX776 (single dose of 0.4 mg), and images were taken during and after ANX776 injection at 15, 120, and 240 minutes, with an average of 100 frames recorded at each time point. All images were anonymized before analysis, and only the 240-minute images were analyzed in this study.

[0075] Furthermore, a pilot study was conducted using three New Zealand white rabbits to evaluate DARC using an established retinal neovascularization model. All experiments were designed and conducted according to protocols approved by the UK Home Office, in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Briefly, 2.5 kg rabbits were anesthetized by intramuscular injection of 50 mg / kg ketamine hydrochloride and 10 mg / kg xylazine, respectively. The pupils were dilated with 2.5% phenylephrine and 1% tropicamide. 50 μl containing 1 μg of human VEGF165 (Sigma UK) dissolved in 0.1% bovine serum albumin (Sigma UK) was intravitreal injected into the left eye only of each animal. 48 hours later, under general anesthesia, the rabbits were intravenously administered Anx776 (0.2 mg) and 1% fluorescein sodium, and DARC and 448 imaging were performed 40 minutes later using high-resolution ICGA and FFA modes of HRA-Spectralis. Three masked observers performed manual counting of DARC images. In the anti-angiogenic experiment, three rabbits were administered ranibizumab eye drops. Three rabbits served as vehicle controls.

[0076] OCT scan acquisition OCT scans were recorded at each follow-up examination for each patient, in accordance with standard treatment, up to 36 months after DARC. DARC images and OCT scans were evaluated in post-hoc analysis.

[0077] DARC Spot Detection A complete method for spot detection is described by Normando et al. (2020). Briefly, it involved training a central neural network (CNN) to classify spots from a control dataset identified using template matching and manually classified by five observers. Two of the five observers' matches were used to train the CNN, and it was found to have an accuracy of 97%, sensitivity of 91.1%, and specificity of 97.1%.

[0078] The reported CNN model and weights are used in this work without modification.

[0079] Alignment of CLSO and OCT The DARC images for each eye were aligned with a corresponding series of OCT scans, allowing for the identification of the DARC spot locations on the corresponding OCT scans during subsequent follow-up. This alignment also made it possible to match the location of the DARC spot identified in the CLSO image with the location of subretinal fluid (SRF) in the OCT image. To register the DARC CLSO images with the OCT image, baseline autofluorescence images and DARC images (240 minutes) were automatically aligned using affine transformation followed by non-rigid B-spline transformation. This was completed using multi-resolution alignment and gradient descent optimization within SimpleITK(NumFOCUS,US). Next, fiducial markers (white dots, Figure 1) were manually placed on the CLSO autofluorescence and OCT reflectance images, and the reflectance images of consecutive Spectralis OCT scans were aligned with the registered DARC images by calculating and applying affine transformations. Once all images were aligned, a location on the 240-minute DARC image could be selected, and the same location could be identified in the corresponding OCT image slice.

[0080] Identification of SRF by CNNs To detect choroidal neovascularization (CNV) lesions on OCT, a CNN was used to identify the region of submucosal resonant fluid (SRF). The CNN was trained using data from The Retinal OCT Fluid Challenge (RETOUCH), described by Bogunovi et al. (2019).

[0081] A custom UNET described by Ronneberger et al., 2015, was adapted. A general training mechanism for conventional UNETs had been previously implemented and was trained using data from The Retinal OCT Fluid Challenge (RETOUCH), described by Bogunovi et al. (2019), where intraretinal fluid (IRF), subretinal fluid (SRF), and detached pigment epithelial (PED) areas were manually annotated by multiple observers.

[0082] The 3-channel label mask for each OCT scan image was created so that each annotation type was represented in a separate channel of the 3-channel RGB image. The model was trained for 100 epochs using UNET with a Windows PC equipped with an 8Gb NVIDIA GTX 1080 graphics card, with a fixed image size of 512x512 pixels, 32 filters, and a dropout of 0.2, and randomly augmented only horizontal inversion. The training dataset was divided into approximately 2 / 3 (N=4,624) for training and approximately 1 / 3 (N=2,312) for validation, with each SRF scan image having a size of 497x495 pixels.

[0083] After the model training was complete, the model was tested on a 1 / 3 validation set. Comparing the DICE coefficients obtained from the trained networks (as described in Normando et al., 2020), SRF scored 0.83 compared to the best system reported in RETOUCH, while Spectralis scored 0.75. The reported results are from a challenge dataset combining Cirrus, Spectralis, and Topcon, and cannot be directly compared to results from a partitioned subset of the training dataset.

[0084] To ensure that the results obtained using the RETOUCH dataset are representative of our DARC study, 735 anonymized OCT images were manually annotated by senior retinal specialists using VGG Image Annotator (VIA), Visual Geometry Group, Oxford University, and https: / / gitlab.com / vgg / via / blob / master / Contributors.md. The DICE coefficient was calculated using the aforementioned SRF CNN and was 0.84. Next, SRF CNN was applied to 20,630 OCT images from a dataset of 31 eyes.

[0085] DARC spots are defined as areas where any SRF is detected in a normal position on the OCT scan surface, intersecting with the SRF region.

[0086] Calculation of SRF accumulation Eyes were classified based on the presence or absence of SRF at 6-month intervals. The average SRF area was calculated for each interval, taking into account the number of OCT scans and interocular variability at 6-month intervals. A new SRF was defined as an eye in which SRF was detected by OCT and which had been classified as SRF-negative in the preceding period.

[0087] The SRF amount for each eye over a 6-month period was defined as the normalized sum of SRF pixels detected in all OCT scans at a specific point in time for the entire eye. The accumulated SRF is the average of the SRF amount per OCT scan over the 6-month period.

[0088] statistical analysis Statistical analysis was performed using GraphPad Prism (version 8.01), Sklearn (version 0.0), SPSS (IBM SPSS version 25), and Python (version 3.6). Precision-recall curves were constructed using the precision, recall, F1 score, and area under the curve (AUC) generated for the CNN test of SRF. Agreement between the clinical definition of SRF and the CNN definition was evaluated using Kappa statistics. Eyes with CNN DARC counts above 5 and those below 5 were compared, and graphs were created based on eyes with residual SRF. These were compared using Wilcoxon rank pairs and simple linear regression statistics at each time point. A confusion matrix was constructed to evaluate the predictive value of eyes with new SRF using only DARC spots covering SRF, from which positive predictive value (PPV), specificity, and sensitivity were calculated. Spearman's correlation coefficient was used to determine the correlation between unique DARC spots covering SRF on OCT and the total SRF accumulation area on OCT at each time point (p<0.05). The DARC spots covering the SRF at each OCT time point were evaluated using a violin plot, and the areas were divided into those with low and high SRF accumulation. The high and low SRF groups were compared using the Mann-Whitney test. Manual DARC counts were analyzed using a paired t-test comparing the left eye treated with human VEGF with the untreated right eye of rabbits.

[0089] Example 1: DARC predicts new subretinal fluid (wet AMD) on OCT at 36 months. We decided to investigate DARC as a predictive technique for the onset of disease, particularly wet AMD (i.e., before a diagnosis is made due to insufficient symptoms).

[0090] A post-hoc analysis of DARC outcomes was conducted from a cohort of 19 AMD patients who participated in the Phase 2 DARC clinical trial at the Western Eye Hospital in ICNHT, London. Patients were regularly followed up with OCT scans for up to 36 months after DARC. Patient demographics are shown in Tables 2 and 3. Of the 31 eyes evaluated by DARC, 29 eyes were eligible for regular OCT scan follow-up up to 36 months after DARC. The average number of OCT scans every six months was 1.70 at 30–36 months and 2.59 at 12–18 months. As patients continued treatment according to standard care, the average number of anti-VEGF intravitreal injections varied from 2.0 at 30–36 months to 3.61 at 0–6 months.

[0091] Analysis at each time point was performed to determine whether DARC spots from baseline predicted the occurrence of wet AMD by intersecting with arbitrary regions detected by the new SRF and CNN on OCT, as shown in Figure 1. A new SRF was defined using an SRF CNN, and at each time point, eyes were divided into a group with new wet AMD and a group without wet AMD.

[0092] The incidence of new SRFs at each time point is shown in Table 4. Using only the definition of OCT CNN SRF, the conversion rates at 6 months, 12 months, 18 months, 24 months, 30 months, and 36 months were 7.4%, 11.5%, 26.1%, 29.2%, and 33.3%, respectively. Eyes were classified as "converted" only if they showed a new SRF for the first time. The CNN DARC count was 5 or higher in all eyes except for one eye associated with SRF formation.

[0093] Both autofluorescence and DARC spots covering the SRF on baseline OCT were excluded from further analysis, where only the first appearance of individual DARC spots covering new SRF (referred to as "unique DARC spots") was included to avoid repeated counting. At a given time point, if an eye showed a unique DARC spot covering new SRF on OCT, it was classified as true positive; an eye with SRF but no unique DARC spot was false negative; an eye with no SRF but a unique DARC spot was false positive; and an eye with neither SRF nor a DARC spot was classified as true negative. Confusion matrices (each confusion matrix determining the predictive judgment of DARC CNN by eye at each time interval) were constructed for 6-month follow-up periods from 6 months to 36 months after baseline. Table 4 summarizes the results of the confusion matrices, showing the ability of unique DARC spots to predict eyes with new SRF activity. Thus, at 6 months, this system had a specificity of 90%, sensitivity of 83%, PPV of 71%, and NPV of 95%. The PPV rate for the DARC DL system was over 70% at all time points, peaking at 86% at 30 months. Specificity was 79–90%, and sensitivity was over 80%.

[0094] Figure 2A shows the proportion of eyes that did not develop SRF (wet AMD) at each 6-month mark, grouped by a CNN DARC count threshold of 5. Intergroup comparisons showed that eyes with a CNN DARC count > 5 had a significantly higher rate of developing SRF compared to eyes with a CNN DARC count ≤ 5 (Wilcoxon rank, p=0.0156). Linear regression of the same data was also performed, and the slopes of the regression estimates were statistically significantly different (p<0.0001, R). 2 =0.92 DARC CNN > 5 vs R 2=1.0 (DARC count ≤ 5). This indicates a significant difference from the "time * group" effect, i.e., the effect of time progression on the CNN DARC count > 5 group and the CNN DARC count ≤ 5 group. Observing the fitted lines and data points, the hazard is large in the first 16 months, and the hazard is relatively flat thereafter, while CNN DARC count ≤ 5 is significantly associated with an extended survival period without residual SRF.

[0095] Figure 2B shows the breakdown of converted eyes based on baseline diagnosis. All seven eyes that developed new SRF had a CNN DARC count > 5. Of the 17 eyes (n=7) that had active CNV (n=10) at baseline or converted during follow-up, 16 had a CNN DARC count > 5.

[0096] This indicates that DARC can predict the onset of wet AMD and lead to early diagnosis.

[0097] Example 2: DARC improves prediction accuracy compared to other prediction methods. Next, the results of the DARC+OCT model in Example 1 were compared with the results reported by Yim et al. (2020). Yim explained that using deep learning (DL) with OCT improves the prediction of the development of choroidal neovascularization (CNV) or wet AMD at 6 months by analyzing OCT features. However, the predictive values ​​were low, specifically, a sensitivity of 80% and a specificity of 55%, a specificity of 90% and a sensitivity of 34%, and positive predictive values ​​(PPVs) of 7.3% and 13.2%, respectively.

[0098] Table 5 shows a comparison of the results obtained using the DARC DL model in Example 1 with the results reported by Yim et al. (2020). The PPV of the DARC DL system is much higher, at 100% at 6 months, compared to 7% or 13% for the Yim system. This is also evident from the fact that DARC DL consistently shows high specificity and sensitivity, at 100% and 50% at 6 months, and 95% and 83% at 36 months. The conversion rates in Example 1 are similar to those of Yim et al. at 6 months, at 14.8% and 13.5%, respectively, despite the difference in sample size.

[0099] Example 3: Increase in DARC count correlates with the observed magnitude of SRF accumulation. We hypothesized that the DARC DL system would perform better than the Yim system because it can evaluate single-cell activity. To further investigate this, we evaluated the relationship between the amount of SRF accumulation and the DARC count in each eye at each time point. Eyes were divided into "high SRF" and "low SRF" at each time point using a threshold of 2000 pixels. Figure 3 shows that in eyes with a large amount of SRF accumulation at each time point, the DARC count covering the SRF increased significantly (p=0.002, <0.0001, 0.0019, 0.0016, 0.0026, and 0.0118 at 6, 12, 18, 24, 30, and 36 months, respectively). Therefore, we found that the larger the SRF, the higher the DARC count, and that eyes with greater SRF accumulation were unique in that they had a DARC count greater than 5. In fact, the DARC spots covering the SRF at each time point were positively correlated with the OCT SRF area at 6 months (Spearman's r=0.65, p=0.0001), 12 months (r=0.72, p<0.0001), 18 months (r=0.60, p=0.0006), 24 months (r=0.57, p=0.0012), 30 months (r=0.67, p<0.0001), and 36 months (r=0.59, p=0.0008). This is supported by a single DARC spot predicting new areas of the SRF, as shown in Figure 1.

[0100] This confirms that a higher DARC count is associated with a higher severity of wet AMD.

[0101] Example 4: Identifying early-stage angiogenesis in rabbits using DARC. A rabbit angiogenesis model was created by intravitreal administration of 1 μg of human VEGF (hVEGF165-Peprotech, London UK) in 50 μl. 48 hours later, ANX776 (0.2 mg) was administered intravenously, and the rabbits were imaged using cSLO ICGA settings.

[0102] All rabbits treated with hVEGF showed positive DARC staining in their eyes two days after administration compared to the untreated contralateral eye. Figure 4 shows DARC images of both eyes of each animal. In the hVEGF-treated left eye (AC), DARC spots were clearly visible compared to the untreated right eye (DF), showing a significant increase in treated eyes (p=0.0203, G). All treated eyes began to show fluorescein leakage on day 4, as shown in inset AC, compared to the control (inset DF).

[0103] This further confirmed the predictive and identifying value of DARC for early-stage angiogenesis.

[0104] Example 5: DARC can be used to monitor the effectiveness of anti-angiogenic therapy. The rabbit model of angiogenesis described in Example 4 was locally treated with the anti-angiogenic drug ranibizumab. Angiogenic DARC labeling decreased with local administration of ranibizumab (Figure 4H-J). Local administration of ranibizumab significantly reduced DARC counts compared to the vehicle control (Figure 4O, p=0.0262).

[0105] Figure 5 shows the efficacy of local treatment with ranibizumab. All control rabbits treated with hVEGF showed fluorescein leakage on day 4 compared to ranibizumab-treated eyes (GI) (Figure 5A-C). Images were collected 8 minutes (DF) and 20 minutes (AC, GI) after administration of 1% fluorescein sodium. This figure confirms that local administration of ranibizumab reduced vascular leakage in vivo. This confirms that DARC can be used to monitor the therapeutic effect.

[0106] [Table 1]

[0107] [Table 2]

[0108] [Table 3]

[0109] [Table 4]

[0110] [Table 5]

[0111] References Bogunovi, H. et al. RETOUCH - The Retinal OCT Fluid Detection and Segmentation Benchmark and Challenge. XX, (2019). Cordeiro, M.F. et al. Real-time imaging of single neuronal cell apoptosis in patients with glaucoma. Brain, 274, 61 (2017). Normando, E. M. et al. A CNN-aided method to predict glaucoma progression using DARC (Detection of Apoptosing Retinal Cells). Expert Rev. Mol. Diagn. 00, 1-12 (2020). Lindekleiv, H, Erke MG. Projected prevalence of age-related macular degeneration in Scandinavia 2021-2040. Acta Ophthalmol. 91 (4):307-311 (2013). Ronneberger, O., Fischer, P. & Brox, T. U-Net: Convolutional Networks for Biomedical Image Segmentation. 1-8 (2015). Saint-Hubert et al. Molecular imaging of cell death. Methods. 48(2), 178-87 (2009). Tait et al. Structural requirements for in vivo detection of cell death with 99mTc-annexin V. J Nucl Med. 46(5), 807-815 (2005). Yim, J. et al. Predicting conversion to wet age-related macular degeneration using deep learning. Nat. Med (2020). Doi:https: / / doi.org / 10.1038 / s41591-020-0867-7

Claims

1. A labeled cell stress marker for use in identifying or predicting angiogenesis in a subject.

2. The labeled cellular stress marker for use according to claim 1 , wherein the cellular stress marker is an annexin or a fragment or variant thereof.

3. The labeled cellular stress marker for use according to claim 2, wherein the cellular stress marker is annexin 5, 11, 2 or 6.

4. A labeled cellular stress marker for use according to any one of claims 1 to 3, wherein the label is fluorescent.

5. The labeled cellular stress marker for use according to claim 4, wherein the fluorescent label has a wavelength in the infrared spectrum.

6. The labeled cell stress marker for use according to claim 4 or 5, wherein the fluorescent label is D-776.

7. The labeled cell stress marker for use according to any one of claims 1 to 6, wherein said angiogenesis is physiological, therapeutic or pathological angiogenesis.

8. The labeled cellular stress marker for use according to claim 7, wherein said physiological angiogenesis is associated with previous exercise by said subject.

9. The labeled cell stress marker for use according to claim 7, wherein said therapeutic angiogenesis is associated with wound healing and / or the integration of tissue engineering scaffolds.

10. 8. The labeled cell stress marker for use according to claim 7, wherein said pathological angiogenesis is associated with disease, disease stage, disease severity or predicts disease onset.

11. The labeled cell stress marker for use according to claim 10, wherein said pathological angiogenesis predicts the onset of disease.

12. 12. The labeled cellular stress marker for use according to claim 10 or 11, wherein the disease comprises cancer, an autoimmune disease, a cardiovascular disease, an allergic disease, a sickle cell disease and / or an eye disease.

13. The labeled cellular stress marker for use according to claim 12, wherein the disease comprises an eye disease.

14. The labeled cell stress marker for use according to claim 13, wherein the eye disease comprises age-related macular degeneration (AMD), optionally wet AMD.

15. 1. A method for identifying or predicting angiogenesis in a subject, the subject being administered a labeled cell stress marker, the method comprising the steps of: (a) generating an image from the object using a first imaging device; (b) identifying cells positive for the labeled cell stress marker within the image; If the labeled cell stress marker positive cells are identified in the image, angiogenesis is identified or predicted.

16. The method of claim 15, wherein the cell stress marker is an annexin or a fragment or variant thereof.

17. The method of claim 16, wherein the cell stress marker is annexin 5, 11, 2 or 6.

18. The method of any one of claims 15 to 17, wherein the label is fluorescent.

19. 20. The method of claim 18, wherein the fluorescent label has a wavelength in the infrared spectrum.

20. The method of claim 18 or 19, wherein the fluorescent label is D-776.

21. The method according to any one of claims 15 to 20, wherein said angiogenesis is physiological, therapeutic or pathological angiogenesis.

22. 22. The method of claim 21, wherein the physiological angiogenesis is associated with prior exercise by the subject.

23. 22. The method of claim 21, wherein said therapeutic angiogenesis is associated with wound healing and / or tissue engineering scaffold integration.

24. The method of any one of claims 15 to 23, wherein the labeled cell stress marker is administered intravenously, intranasally or topically.

25. The method of any one of claims 15 to 24, wherein identifying the labeled cell stress marker positive cells comprises counting the number of the labeled cell stress marker positive cells in the image.

26. 26. The method of claim 25, wherein the identification of cells positive for the labeled cell stress marker is performed in silico.

27. 27. The method of claim 25 or 26, further comprising determining the location of the labeled cell stress marker positive cells in the subject.

28. 28. The method of claim 27, wherein the location is determined using another image obtained using a second imaging device.

29. 29. The method of claim 28, wherein the second imaging device comprises a computed tomography (CT) scanner, a magnetic resonance imaging (MRI) scanner, a positron emission tomography (PET) scanner, a polarizer, a reflective image scanner, a fundus camera, and / or an OCT scanner.

30. 30. The method of claim 29, wherein the second imaging device comprises an OCT scanner.

31. The method of any one of claims 15 to 30, wherein the first imaging device comprises a confocal scanner laser microscope, optionally a confocal scanning laser ophthalmoscope (cSLO).

32. A labeled apoptotic marker for use in identifying apoptosis in non-ocular body parts of a subject.

33. 33. The labeled apoptosis marker of claim 32, wherein the body part is not the central nervous system.