Traitement en cours

Veuillez attendre...

Paramétrages

Paramétrages

Aller à Demande

1. WO2021067699 - PRÉDICTION DE PARAMÈTRES CLINIQUES ASSOCIÉS AU GLAUCOME À PARTIR DE MOTIFS DE CHAMP VISUEL CENTRAL

Note: Texte fondé sur des processus automatiques de reconnaissance optique de caractères. Seule la version PDF a une valeur juridique

[ EN ]

Having described the invention, we claim:

1 . A system comprising:

a processor;

an output device; and

a non-transitory computer readable medium storing instructions executable by the processor to provide:

a pattern decomposition component that receives a set of visual field data for a patient representing, for each of a plurality of locations in the central region of an eye of the patient, a deviation in sensitivity to a visual stimulus from an age-adjusted normal value and decomposes the set of visual field data into a linear combination of a set of patterns defined via archetypal analysis over a corpus of visual field data to provide a set of decomposition coefficients;

a machine learning model that determines a clinical parameter for the patient from at least the set of decomposition coefficients; and

a user interface that provides the determined clinical parameter to a user at the output device.

2. The system of claim 1 , further comprising a preclassifier that classifies the patient into one of a plurality of categories based on at least one metric associated with the eye and selects the set of patterns from a plurality of sets of patterns defined via archetypal analysis according to the category into which the patient is classified.

3. The system of claim 1 , wherein the set of patterns defined via archetypal analysis represents patterns seen in end-stage glaucoma and comprises a first subset of patterns representing temporal-sparing, a second subset of patterns representing hemifield loss, a third subset of patterns representing a central island of intact vision, a fourth subset of patterns representing nasal loss, and patterns representing nearly total loss, total loss, inferonasal quadrant sparing, and nearly intact central vision.

4. The system of claim 1 , wherein the set of patterns defined via archetypal analysis represents patterns seen in severe glaucoma and comprises a first subset of patterns representing loss of visual acuity in the superior portion of the visual field, a second subset of patterns representing loss of visual acuity in the inferior portion of the visual field, a third subset of patterns representing diffuse patterns of loss of visual acuity, each with an island of relatively low loss, and patterns representing an intact visual field, loss across the entire field, loss in the temporal region of the visual field, and loss in the nasal region of the visual field.

5. The system of claim 1 , wherein the set of patterns defined via archetypal analysis represents patterns seen in one of mild glaucoma and moderate glaucoma and comprises a first subset of patterns representing loss of visual acuity in the superior portion of the visual field, a second subset of patterns representing loss of visual acuity in the inferior portion of the visual field, a third subset of patterns representing patterns of loss of visual acuity that are present in the superior and the inferior regions.

6. The system of claim 1 , wherein the clinical parameter represents an expected rate of change in a measure of visual acuity for the eye.

7. The system of claim 1 , wherein the machine learning model determines the clinical parameter from the set of decomposition coefficients and an additional feature representing the patient.

8. The system of claim 7, wherein the additional feature representing the patient is one of a mean deviation from the set of visual field data, a pattern standard deviation from the set of visual field data, and an intraocular pressure of the eye.

9. The system of claim 7, wherein the additional feature representing the patient is one of a blood pressure, blood glucose level, age, and sex of the patient.

10. The system of claim 7, wherein the additional feature representing the patient is mean absolute difference between total deviation values at each of the plurality of locations in visual field pattern and a reconstructed baseline visual field generated as the sum of the set of patterns defined via archetypal analysis weighed by the set of coefficients.

11. A method comprising:

obtaining a set of visual field data for a patient representing, for each of a plurality of locations in the central region of the eye, a deviation in sensitivity to a visual stimulus from an age-adjusted normal value;

decomposing the set of visual field data into a linear combination of a set of patterns defined via archetypal analysis over a corpus of visual field data to provide a set of decomposition coefficients;

determining a clinical parameter for the patient at a machine learning model from at least the set of decomposition coefficients; and

providing the determined clinical parameter to a user at a display.

12. The method of claim 11 , wherein the clinical parameter represents a subtype of early central visual field loss.

13. The method of claim 11 , wherein the set of patterns defined via archetypal analysis is a selected set of patterns defined via archetypal analysis from a plurality of sets of patterns defined via archetypal analysis, the method further comprising: classifying the patient into one of a plurality of categories based on at least one metric associated with the eye; and

selecting the selected set of patterns defined via archetypal analysis according to the category into which the patient is classified.

14. The method of claim 11 , wherein determining the clinical parameter for the patient at the machine learning model comprises determining the clinical parameter at the machine learning model from the set of decomposition coefficients and an additional feature representing the patient, the additional feature being one of a biometric parameter of the patient, a measured characteristic of the eye, and a global metric derived from the set of visual field data.

15. The method of claim 11 , wherein the clinical parameter is a parameter representing an expected visual acuity for the eye after a predetermined period of time.

16. A system comprising:

a processor;

an output device; and

a non-transitory computer readable medium storing instructions executable by the processor to provide:

a preclassifier that classifies the patient into one of a plurality of categories representing the severity of glaucoma for the patient and selects the set of patterns from a plurality of sets of patterns defined via archetypal analysis according to the category into which the patient is classified;

a pattern decomposition component that receives a set of visual field data for a patient representing, for each of a plurality of locations in the central region of an eye of the patient eye, a deviation in sensitivity to a visual stimulus from an age-adjusted normal value and decomposes the set of visual field data into a linear combination of the selected set of patterns defined via archetypal analysis to provide a set of decomposition coefficients;

a machine learning model that determines a clinical parameter for the patient from at least the set of decomposition coefficients; and

a user interface that provides the determined clinical parameter to a user at the output device.

17. The system of claim 16, wherein determining the at least one clinical parameter for the patient at the machine learning model comprises determining the at least one clinical parameter at the machine learning model from the set of decomposition coefficients and an additional feature representing the patient selected as one of mean deviation from the set of visual field data, a pattern

standard deviation from the set of visual field data, an intraocular pressure of the eye, a blood pressure of the patient, blood glucose level of the patient, an age of the patient, a sex of the patient, and a mean absolute difference between total deviation values at each of the plurality of locations in visual field pattern and a reconstructed baseline visual field generated as the sum of the set of patterns defined via archetypal analysis weighed by the set of coefficients.

18. The system of claim 16, wherein the plurality of sets of patterns comprises:

a first set of patterns defined via archetypal analysis, representing patterns seen in end-stage glaucoma and comprising a first subset of patterns representing temporal-sparing, a second subset of patterns representing hemifield loss, a third subset of patterns representing a central island of intact vision, a fourth subset of patterns representing nasal loss, and patterns representing nearly total loss, total loss, inferonasal quadrant sparing, and nearly intact central vision;

a second set of patterns defined via archetypal analysis represents patterns seen in severe glaucoma and comprises a first set of patterns representing loss of visual acuity in the superior portion of the visual field, a second set of patterns representing loss of visual acuity in the inferior portion of the visual field, a third set of patterns representing diffuse patterns of loss of visual acuity, each with an island of relatively low loss, and patterns representing an intact visual field, loss across the entire field, loss in the temporal region of the visual field, and loss in the nasal region of the visual field; and

a third set of patterns defined via archetypal analysis represents patterns seen in one of mild glaucoma and moderate glaucoma and comprises a first set of patterns representing loss of visual acuity in the superior portion of the visual field, a second set of patterns representing loss of visual acuity in the inferior portion of the visual field, a third set of patterns representing patterns of loss of visual acuity that are present in the superior and the inferior regions.

19. The system of claim 16, wherein the preclassifier receives the set of visual field data, determines a mean deviation across the plurality of locations, and classifies the patient according to the determined mean deviation.

20. The system of claim 16, wherein the set of visual field data is a first set of visual field data and the preclassifier receives a second set of visual field data representing the entire eye and classifies the patient according to a metric determined from the second set of visual field data.