Machine-learning techniques for prediction of future visual acuity
Methods and systems disclosed herein relate generally to systems and methods for predicting a future visual acuity of a subject by using machine-learning models. An image of at least part of a retina of a subject can be processed by one or more first machine-learning models to detect a set of retina-related segments. Segment-specific metrics that characterize a retina-related segment of the set of retina-related segments can be generated. The segment-specific metrics can be processed by using a second machine-learning model to generate a result corresponding to a prediction corresponding to a future visual acuity of the subject.
1. A computer-implemented method comprising:
processing, using one or more image-segmenting models, an image of at least part of a retina of a subject to detect a set of retina-related segments in the image, each of the set of retina-related segments including a retinal structure or a type of retinal fluid;
generating a set of segment-specific metrics, each of the set of segment-specific metrics characterizing a retina-related segment of the set of retina-related segments, wherein at least one segment-specific metric of the set of segment-specific metrics identifies an amount of fluid volume corresponding to the type of retinal fluid, and wherein the type of retinal fluid is (i) an intra-retinal fluid, and a decrease of a value corresponding to the at least one segment-specific metric indicates an increase of a likelihood that a future visual acuity of the subject at a particular future time point will exceed a predetermined acuity threshold, or (ii) a sub-retinal fluid, and an increase of a value corresponding to the at least one segment-specific metric indicates an increases a likelihood that the future visual acuity of the subject at a particular future time point will exceed a predetermined acuity threshold;
processing the set of segment-specific metrics using a metric-processing model to determine that the amount of the fluid volume corresponding to the type of retinal fluid is a primary indicator for determining a degree of a future enhancement of a visual acuity of the subject;
generating a result corresponding to a prediction corresponding to the future visual acuity of the subject; and
outputting the result.
2. The computer-implemented method of claim 1 , wherein an image-segmenting model of the one or more image-segmenting models includes a deep-convolutional neural network.
3. The computer-implemented method of claim 1 , wherein an image-segmenting model of the one or more image-segmenting models uses an intra-retinal-layer segmentation algorithm.
4. The computer-implemented method of claim 1 , wherein:
an image-segmenting model of the one or more image-segmenting models is used to detect a segment corresponding to the type of retinal fluid; and
another model of the one or more image-segmenting models is used to detect a segment corresponding to the retinal structure.
5. The computer-implemented method of claim 1 , wherein the metric-processing model includes a trained gradient-boosting machine.
6. The computer-implemented method of claim 1 , wherein the image is an optical-coherence-tomography (OCT) image.
7. The computer-implemented method of claim 1 , wherein a retina-related segment of the set of retina-related segments includes a particular retinal structure that indicates one or more retinal layers of the retina.
8. The computer-implemented method of claim 7 , wherein the particular retinal structure indicates one or more parts of the retina located beneath a retinal pigment epithelium, the one or more parts of the retina including a Bruch's membrane, a choroid, and a sclera.
9. The computer-implemented method of claim 1 , wherein a retina-related segment of the set of retina-related segments indicates a pigment epithelial detachment of the retinal structure.
10. The computer-implemented method of claim 1 , wherein a retina-related segment of the set of retina-related segments includes a particular type of retinal fluid that includes a sub-retinal fluid (SRF) or intra-retinal fluid (IRF).
11. The computer-implemented method of claim 1 , wherein a retina-related segment of the set of retina-related segments indicates one or more deformities present in the retina, the one or more deformities including a macular hole, a macular pucker, and a deteriorated macula.
12. The computer-implemented method of claim 1 , wherein the result indicates a prediction that the future visual acuity of the subject at a particular future time point will exceed a predetermined acuity threshold.
13. A system comprising:
one or more data processors; and
a non-transitory computer readable storage medium containing instructions which, when executed on the one or more data processors, cause the one or more data processors to perform one or more operations comprising:
processing, using one or more image-segmenting models, an image of at least part of a retina of a subject to detect a set of retina-related segments in the image, each of the set of retina-related segments including a retinal structure or a type of retinal fluid;
generating a set of segment-specific metrics, each of the set of segment-specific metrics characterizing a retina-related segment of the set of retina-related segments, wherein at least one segment-specific metric of the set of segment-specific metrics identifies an amount of fluid volume corresponding to the type of retinal fluid, and wherein the type of retinal fluid is (i) an intra-retinal fluid, and a decrease of a value corresponding to the at least one segment-specific metric indicates an increase of a likelihood that a future visual acuity of the subject at a particular future time point will exceed a predetermined acuity threshold, or (ii) a sub-retinal fluid, and an increase of a value corresponding to the at least one segment-specific metric indicates an increases a likelihood that the future visual acuity of the subject at a particular future time point will exceed a predetermined acuity threshold;
processing the set of segment-specific metrics using a metric-processing model to determine that the amount of the fluid volume corresponding to the type of retinal fluid is a primary indicator for determining a degree of a future enhancement of a visual acuity of the subject;
generating a result corresponding to a prediction corresponding to the future visual acuity of the subject; and
outputting the result.
14. The system of claim 13 , wherein an image-segmenting model of the one or more image-segmenting models includes a deep-convolutional neural network.
15. The system of claim 13 , wherein an image-segmenting model of the one or more image-segmenting models uses an intra-retinal-layer segmentation algorithm.
16. The system of claim 13 , wherein:
an image-segmenting model of the one or more image-segmenting models is used to detect a segment corresponding to the type of retinal fluid; and
another model of the one or more image-segmenting models is used to detect a segment corresponding to the retinal structure.
17. The system of claim 13 , wherein a retina-related segment of the set of retina-related segments includes a particular retinal structure that indicates one or more retinal layers of the retina.
18. A computer-program product tangibly embodied in a non-transitory machine-readable storage medium, including instructions configured to cause one or more data processors to perform one or more operations comprising:
processing, using one or more image-segmenting models, an image of at least part of a retina of a subject to detect a set of retina-related segments in the image, each of the set of retina-related segments including a retinal structure or a type of retinal fluid;
generating a set of segment-specific metrics, each of the set of segment-specific metrics characterizing a retina-related segment of the set of retina-related segments, wherein at least one segment-specific metric of the set of segment-specific metrics identifies an amount of fluid volume corresponding to the type of retinal fluid, and wherein the type of retinal fluid is (i) an intra-retinal fluid, and a decrease of a value corresponding to the at least one segment-specific metric indicates an increase of a likelihood that a future visual acuity of the subject at a particular future time point will exceed a predetermined acuity threshold, or (ii) a sub-retinal fluid, and an increase of a value corresponding to the at least one segment-specific metric indicates an increases a likelihood that the future visual acuity of the subject at a particular future time point will exceed a predetermined acuity threshold;
processing the set of segment-specific metrics using a metric-processing model to determine that the amount of the fluid volume corresponding to the type of retinal fluid is a primary indicator for determining a degree of a future enhancement of a visual acuity of the subject;
generating a result corresponding to a prediction corresponding to the future visual acuity of the subject; and
outputting the result.