Methods and systems for automating opthalmic diagnosis and treatment

A computer-implemented method for automated detection of glaucoma using camera-independent machine vision, comprising:

• acquiring one or more images and/or videos of a subject's eye using an imaging device selected from one or more of a camera, a slit lamp biomicroscope, a binocular indirect ophthalmoscope, an anterior segment camera, a fundus camera, or combinations thereof;
• inputting the acquired one or more images and/or videos into a processing unit, the processing unit configured to:
• analyze the one or more images and/or videos with a trained neural network to identify biomarkers indicative of glaucoma, the trained neural network trained with a dataset of retinal images and/or videos annotated with glaucoma severity levels;
• outputting a risk assessment score for glaucoma based on the identified biomarkers;
• and
• providing a diagnostic report including the risk assessment score and recommended clinical actions, therapeutic options, or both.

A system for automatically monitoring a progression of an ocular disease or pathology, the system comprising:

• an imaging device for acquiring images and/or videos of a subject's eye at two different points in time, wherein the imaging device is selected from one or more of a camera, a slit lamp biomicroscope, a binocular indirect ophthalmoscope, an anterior segment camera, a fundus camera, or combinations thereof;
• a processing unit configured to:
• at least two of registering, aligning, or comparing, images and/or videos from a first point in time of the at least two different points in time and images and/or videos from a second point in time of the at least two different points in time, to detect changes in one or more ocular structure from the first point of time to the second point in time;
• and
• evaluate a progression rate of the ocular disease or pathology;
• an output module configured to one or more of:
• generate one or more reports indicating potential disease or pathology;
• generate one or more trend analysis reports indicating disease or pathology progression;
• or
• provide one or more therapeutical options, suggest one or more treatment, suggest one or more treatment plan, suggest one or more modifications to a treatment plan, or combinations thereof.

A system for automated ophthalmic diagnostic classification using video ophthalmoscopy, comprising:

• an imaging device configured to acquire video frames and/or static images of a subject's optic disc during an ophthalmoscopic examination;
• an input module for receiving said video frames and/or static images;
• a neural network operatively connected to the input module, the neural network configured to:
• analyze the acquired video frames and/or static images using a machine learning model trained to segment optic nerve head features across varying image qualities and patient populations;
• measure a radial width of a neuroretinal rim by determining a shortest distance between an outer optic disc margin and an inner cup margin from the segmented optic nerve head features, wherein the measurement:
• substantially or completely excludes blood vessels extending beyond rim tissue;
• accounts for a circumferential extent of complete rim absence, if present;
• and
• is taken along a radial line extending from a center of the optic disc to a location where the rim is thinnest;
• calculate a Disc Damage Likelihood Scale (DDLS) score based on the measured neuroretinal rim width and an optic disc size;
• adjust the calculated DDLS score based on the optic disc size;
• and
• an output module configured to:
• provide real-time or substantially real-time DDLS classification during the ophthalmoscopic examination;
• generate a report indicating glaucoma risk assessment based on the DDLS score;
• and
• suggest clinical recommendations based on the risk assessment.