SYSTEMS AND METHODS FOR AUTOMATED CLASSIFICATION OF ABNORMALITIES IN OPTICAL COHERENCE TOMOGRAPHY IMAGES OF THE EYE

• US 20120274898A1
• Filed: 04/26/2012
• Published: 11/01/2012
• Est. Priority Date: 04/29/2011
• Status: Active Grant

• Current Assignee: Carl Zeiss Meditec Incorporated (Carl-Zeiss-Stiftung), Doheny Eye Institute, University of Southern California
• Original Assignee: Carl Zeiss Meditec Incorporated (Carl-Zeiss-Stiftung), Doheny Eye Institute, University of Southern California
• Inventors: Sadda, SriniVas R., Stetson, Paul F.
• Granted Patent: US 9,226,654 B2
• Time in Patent Office: Days
• US Class Current: 351/206
• CPC Class Codes: A61B 3/0025 characterised by electronic signal processing, e.g. eye models
• CPC Class Codes: A61B 3/102 for optical coherence tomography [OCT]
• CPC Class Codes: G01B 9/02044 Imaging in the frequency domain, e.g. by using a spectrometer
• CPC Class Codes: G01B 9/02083 characterised by particular signal processing and presentation
• CPC Class Codes: G01B 9/02091 Tomographic interferometers, e.g. based on optical coherence
• CPC Class Codes: G06T 2207/10101 Optical tomography; Optical coherence tomography [OCT]
• CPC Class Codes: G06T 2207/20081 Training; Learning
• CPC Class Codes: G06T 2207/30041 Eye; Retina; Ophthalmic
• CPC Class Codes: G06T 2207/30104 Vascular flow; Blood flow; Perfusion
• CPC Class Codes: G06T 7/0012 Biomedical image inspection
• CPC Class Codes: G06T 7/12 Edge-based segmentation

Abstract

Systems and methods for classifying abnormalities within optical coherence tomography images of the eye are presented. One embodiment of the present invention is the classification of pigment epithelial detachments (PEDs) based on characteristics of their internal reflectivity, size and shape. The classification can be based on selected subsets of the data located within or surrounding the abnormalities. Training data can be used to generate the classification scheme and the classification can be weighted to highlight specific classes of particular clinical interest.

Claims

1. A method of analyzing an abnormality in the retinal layers of the eye, said method comprising:

collecting three dimensional optical coherence tomography (OCT) intensity data of the retinal layers of the eye including the abnormality;

segmenting the OCT intensity data to identify the boundaries of the abnormality;

determining one or more representative values of the abnormality;

classifying the abnormality based on the one or more representative values using predetermined criteria;

displaying or storing the classification.

2. A method as recited in claim 1, wherein the one or more abnormalities is a pigment epithelial detachment (PED).

3. A method as recited in claim 2, wherein the determining and classifying are carried out on subsets of the intensity data within the abnormality boundaries.

4. A method as recited in claim 3, wherein the subsets are A-scans contained within the identified PED boundaries.

5. A method as recited in claim 3, wherein the mean of the intensity for the portion of the A-scan within the PED boundary and the standard deviation of the intensity for the portion of the A-scan within the PED boundary are determined as the representative values.

6. A method as recited in claim 3, wherein the mean of the intensity for the portion of the A-scan within the PED boundary and the curvature of the RPE are determined as the representative values.

7. A method as recited in claim 4, further comprising calculating a single index for each PED by combining the classifications for the A-scans within the identified PED boundaries.

8. A method as recited in claim 7, wherein the classifications are displayed based on the location of the A-scans.

9. A method as recited in claim 3, further comprising calculating a confidence index for the classification.

10. A method as recited in claim 3, further comprising calculating a fibrovascular PED risk index based on the percentage of the subsets within a PED that are classified as fibrovascular.

11. A method as recited in claim 1, wherein the predetermined criteria are developed using a collection of training data.

12. A method as recited in claim 2, wherein the PED is classified as one of:

serous, fibrovascular, and drusenoid.

13. A method as recited in claim 2, wherein the predetermined criteria have a higher sensitivity for fibrovascular PEDs.

14. A method as recited in claim 2, wherein the one or more representative values is selected from:

mean of intensity, median of intensity, standard deviation of intensity, RPE curvature, size, choroidal thickness, location relative to the fovea, location relative to optic nerve head, and presence or absence of blood flow.

15. A method as recited in claim 1, further comprising normalizing the data prior to the classifying.

16. A method of characterizing regions of pigment epithelial detachment (PED) within the eye of a patient based on image data obtained from an optical coherence tomography (OCT) system, said method comprising the steps of:

identifying tissue regions within the image that correspond to PEDs;

determining a first intensity value associated with each identified region, said intensity value being selected from a mean, median, and average intensity;

determining a second value associated with each identified region, said second representative value being selected from one of uniformity of the intensity of the region or a geometric attribute of the region;

classifying the type of PED in the identified region based on the first and second values; and

displaying or storing the results of the classification.

17. A method as recited in claim 16, wherein the second value is based on the standard deviation of the intensity of the identified region.

18. A method as recited in claim 16, wherein the second value is based on the curvature of the identified region.

19. A method as recited in claim 16, wherein the display is arranged to identify both the classification of the identified region as well as the position within the eye where the region is located.

20. A method as recited in claim 16, wherein the identified region is classified as one of:

serous, fibrovascular, and drusenoid.

21. An optical coherence tomography (OCT) system for characterizing regions of pigment epithelial detachment (PED) within the eye of a patient, said OCT system comprising:

a light source arranged to generate a beam of radiation

a beam divider for separating the beam along a sample arm and a reference arm;

optics for scanning the beam in the sample arm over a set of transverse locations on the eye;

a detector for measuring radiation returning from both the sample arm and the reference arm, the detector generating output signals in response thereto; and

a processor for converting the output signals into image data, said processor identifying tissue regions within the image that correspond to PEDs, said processor determining a first intensity value associated with each identified region, said intensity value being selected from a mean, median, and average intensity, said processor determining a second value associated with each identified region, said second representative value being selected from one of uniformity of the intensity of the region or a geometric attribute of the region, and wherein said processor classifies the type of PED in the identified region based on the first and second values.

22. A system as recited in claim 21, wherein the second intensity value is based on the standard deviation of the intensity of the identified region.

23. A system as recited in claim 21, wherein the second value is based on the curvature of the identified region.

24. A system as recited in claim 21, further comprising a display for displaying the resulting classification wherein the display is arranged to identify both the classification of the identified region as well as the position within the eye where the region is located.

25. A system as recited in claim 21, wherein the identified region is classified as one of:

serous, fibrovascular, and drusenoid.