Identifying suspicious areas in ophthalmic data
An ophthalmic image diagnostic tool and method submits a test image to a neural network trained to identify abnormal regions of an ophthalmic image, to distinguish between multiple types of abnormalities, and to associate an abnormality type with each identified potentially abnormal region. Each potentially abnormal region in the test image is highlighted, and in response to a user-selection of a highlighted region, a previously diagnosed sample (e.g., from a library of samples) of the abnormality type associated with the selected highlighted region is displayed.
1. An ophthalmic image diagnostic method, comprising:
submitting a test ophthalmic image to a neural network trained to identify potentially abnormal regions of an ophthalmic image, the neural network distinguishing between one or more abnormality types, wherein at least one abnormality type is associated with each identified potentially abnormal region;
displaying the test ophthalmic image on a display and highlighting on the display a region of the test ophthalmic image identified as a potentially abnormal region; and
displaying on the display at least one diagnosed ophthalmic sample associated with the abnormality type corresponding to the highlighted potentially abnormal region, wherein the test ophthalmic image is of a first patient and the displayed diagnosed ophthalmic sample is of different patient.
2. An ophthalmic image diagnostic method, comprising:
submitting a test ophthalmic image to a neural network trained to identify potentially abnormal regions of an ophthalmic image, the neural network distinguishing between a plurality of abnormality types, wherein the neural network associates at least one abnormality type with each identified potentially abnormal region;
displaying the test ophthalmic image on a display and highlighting on the display any region of the test ophthalmic image identified as a potentially abnormal region by the neural network;
providing access to a diagnostic library, the diagnostic library having at least one diagnosed ophthalmic sample associated with each of the plurality of abnormality types; and
responding to a user-selection of a highlighted region of the displayed test ophthalmic image by displaying at least one diagnosed ophthalmic sample, from the diagnostic library, associated with the abnormality type corresponding to the selected highlighted region.
3. The method of claim 2 , wherein the displayed, at least one diagnosed ophthalmic sample includes at least one link to another diagnosed ophthalmic sample, from the library, associated with the abnormality type corresponding to the selected highlighted region.
4. The method of claim 1 , wherein each highlighted region on the displayed test ophthalmic image is color-coded according to its associated abnormality type.
5. The method of claim 1 , wherein the displayed, at least one diagnosed ophthalmic sample includes a sample tissue image representative of the associated abnormality type.
6. The method of claim 1 , wherein:
the displayed, at least one diagnosed ophthalmic sample includes a sample tissue image; and
a portion of the sample tissue image that is most similar to the highlighted region of the displayed test ophthalmic image is highlighted.
7. The method of claim 6 , further including:
responding to a user-selection of the highlighted portion of the sample tissue image by displaying a second diagnosed ophthalmic sample.
8. The method of claim 1 , wherein:
the test ophthalmic image is of a first patient;
the displayed at least one diagnosed ophthalmic sample is of said first patient; and
the displayed at least one diagnosed ophthalmic sample includes a tissue image with a highlighted portion corresponding to the highlighted region of the displayed test ophthalmic image.
9. The method of claim 1 , wherein the displayed, at least one diagnosed ophthalmic sample includes a textual description of the associated abnormality type.
10. The method of claim 1 , wherein the displayed, at least one diagnosed ophthalmic sample includes a selectable link to additional diagnostic information.
11. The method of claim 1 , wherein the test ophthalmic image includes one or more of an optical coherence tomography (OCT) image, an OCT Angiography image, and an en face ophthalmic image.
12. An ophthalmic image diagnostic method, comprising:
displaying a test ophthalmic image on a display, the test ophthalmic image having at least one identified potentially abnormal region, each identified potentially abnormal region being associated with at least one abnormality type, and each abnormality type being further associated with at least one diagnosed ophthalmic sample;
highlighting on the display any potentially abnormal region of the test ophthalmic image; and
following a user-selection of a highlighted region of the displayed test ophthalmic image, displaying at least one diagnosed ophthalmic sample associated with the selected highlighted region, wherein the test ophthalmic image is of a first patient and the displayed diagnosed ophthalmic sample is of different patient.
13. The method of claim 12 , further comprising:
prior to displaying the diagnosed ophthalmic sample, responding to the user-selection of the highlighted region of the displayed test ophthalmic image by displaying a selection of abnormality types associated with the selected highlighted region; and
the displayed diagnosed ophthalmic sample is further associated with a user-selection of one of the displayed abnormality types.
14. An ophthalmic image diagnostic method, comprising:
displaying a test ophthalmic image on a display, the test ophthalmic image having at least one identified potentially abnormal region, each identified potentially abnormal region being associated with at least one abnormality type, and each abnormality type being further associated with at least one diagnosed ophthalmic sample;
highlighting on the display any potentially abnormal region of the test ophthalmic image;
following a user-selection of a highlighted region of the displayed test ophthalmic image, displaying at least one diagnosed ophthalmic sample associated with the selected highlighted region; and
prior to highlighting any potentially abnormal region on the display:
submitting the test ophthalmic image to a neural network trained to identify potentially abnormal regions of an ophthalmic image, the neural network distinguishing between a plurality of abnormality types and having access to a library of diagnosed ophthalmic samples associated each abnormality type, wherein the neural network identifies the at least one potentially abnormal region of the test ophthalmic image and associates at least one abnormality type with each identified potentially abnormal region.
15. An ophthalmic image diagnostic system comprising:
one or more processors; and
one or more computer-readable non-transitory storage media coupled to one or more of the processors and comprising instructions operable when executed by one or more of the processors to cause the system to:
submit a test ophthalmic image to a neural network trained to identify potentially abnormal regions of an ophthalmic image, the neural network distinguishing between one or more abnormality types, wherein the neural network associates at least one abnormality type with each identified potentially abnormal region;
display the test ophthalmic image on a display and highlight on the display any region of the test ophthalmic image identified as a potentially abnormal region by the neural network; and
respond to a user-selection of a highlighted region of the displayed test ophthalmic image by displaying at least one diagnosed ophthalmic sample associated with the abnormality type corresponding to the selected highlighted region, wherein the test ophthalmic image is of a first patient and the displayed diagnosed ophthalmic sample is of different patient the displayed ophthalmic sample being different from the test ophthalmic image.
16. An ophthalmic image diagnostic system comprising:
one or more processors; and
one or more computer-readable non-transitory storage media coupled to one or more of the processors and comprising instructions operable when executed by one or more of the processors to cause the system to:
submit a test ophthalmic image to a neural network trained to identify potentially abnormal regions of an ophthalmic image, the neural network distinguishing between a plurality of abnormality types, wherein the neural network associates at least one abnormality type with each identified potentially abnormal region;
display the test ophthalmic image on a display and highlight on the display any region of the test ophthalmic image identified as a potentially abnormal region by the neural network
provide access to a diagnostic library, the diagnostic library having at least one diagnosed ophthalmic sample associated with each of the plurality of abnormality types; and
respond to a user-selection of a highlighted region of the displayed test ophthalmic image by displaying at least one diagnosed ophthalmic sample, from the diagnostic library, associated with the abnormality type corresponding to the selected highlighted region.
17. The system of claim 16 , wherein:
the display is part of a computing device remote from the diagnostic library;
the computing device retrieves the at least one diagnostic ophthalmic sample over a computer network in response to the user-selection of the highlighted region of the displayed test ophthalmic image.
18. The system of claim 16 , wherein:
the neural network is hosted on a server remote from the computing device and accessible via the computer network; and
the computing device instructs the submission of the test ophthalmic image to the neural network on the host server over the computer network.
19. The method of claim 2 , wherein diagnosed ophthalmic samples in the diagnostic library are from patients different from the patient associated with the test ophthalmic image.
20. The method of claim 14 , wherein the test ophthalmic image is of a first patient and the displayed diagnosed ophthalmic sample is of different patient.
21. The system of claim 16 , wherein diagnosed ophthalmic samples in the diagnostic library are from patients different from the patient associated with the test ophthalmic image.