Article
Author(s):
New software (MatchedFlicker, EyeIC) uses time series retinal photographs to convert change into motion and could improve the accuracy and speed of detecting glaucomatous progression.
New York-A new software program that automatically aligns digital retinal or optic nerve photographs taken several years apart and alternates them to allow clinicians to look for changes over time may improve the speed and accuracy of glaucoma diagnosis.
The technology (MatchedFlicker, EyeIC), which received FDA 510(k) clearance in July, lets physicians go back in time to see whether the conditions of patients who have been followed for many years are getting worse.
The gold standard for detecting glaucomatous change over time is to c ompare a series of stereo photographs taken several years apart. The new software builds on that standard by facilitating the identification of change over time from photographs, said Nathan Radcliffe, MD, director of glaucoma at Weill Cornell Medical College and New York-Presbyterian Hospital, New York.
The new software evolved from a series of patents that its maker acquired and augmented to develop its product.
"You take two photographs and turn them into a movie that flickers back and forth. What that does is turn change into motion," said Ira Wallace, MD, chief executive officer of EyeIC. "The human visual system is very sensitive to seeing motion, especially if nothing else in the photograph is moving, so the doctor sees the changes between the photographs and needs to make the decision about whether those changes represent disease-associated progression or some other change."
Although there are other means of performing automation flicker, such as putting two slide projectors next to each other or overlapping two digital photographs on a computer screen, the new software takes a more sophisticated approach because it automatically aligns photographs at a subpixel level, Dr. Radcliffe said.
"There's a very high level of alignment, and it allows you to detect changes that would otherwise be undetectable," he added.
For example, it has been reported that progression of peripapillary atrophy is associated with glaucomatous progression, but this association is extremely difficult to determine in side-by-side comparisons because the changes can be very small. Additionally, small changes in blood vessels along the neuroretinal rim of the optic disc can be very specific for glaucomatous progression, but these shifts also are quite small and difficult to see.
Aligning the images
"However, with [the software], since everything else about those two photographs has been aligned, changes in the optic disc blood vessels or peripapillary atrophy can be seen fairly easily," Dr. Radcliffe explained.
Describing the way in which the software works, he said that the algorithm requires that the two photographs have at least 20% of the image in common. Even if the eye has undergone considerable change over the years, due to causes such as progressive cataract or retinal disease, the images typically still can be matched.
In addition, digitized photos from any source can be used with the software.
The software looks for any identifiable features in the photographs, such as the crossing of retinal or choroidal blood vessels or chorioretinal scars, and aligns the images along these features to achieve the best statistical match between the photographs. The physician then can determine the rate at which the two photographs are alternated to perceive the change over time. The software can magnify or rotate the images but applies the same changes to the whole photograph. The color and illumination are not affected by the algorithm.
"The physician needs to be aware that changes in the cardiac cycle or differences in the angle from which the two photographs were taken can cause apparent change, which can be differentiated from true change with experience," Dr. Radcliffe said.
He added that the software allows clinicians to look at a multitude of different features of the optic nerve, such as changes in the neuroretinal rim, the retinal nerve fiber layer, the blood vessels, the appearance of disc hemorrhages, or peripapillary atrophy. Most other imaging technologies enable users to look at one or two of those features but not all of them.
Documentation tools
The software includes documentation tools.
According to Dr. Wallace, a mark made on a suspicious feature in any one of the aligned images automatically shows up in the corresponding area of every photograph so that the change immediately can be seen. The mark then can be annotated and printed or saved to an electronic medical record. The images can also be shared with other clinicians.
Dr. Radcliffe said he has used the software in several research studies to look for glaucomatous change over time and found that the images are quite dramatic.
"When you see glaucomatous progression happening in front of your face, it's quite impressive. I've found it to be a great teaching tool. There's also some thought that this might be useful to show to patients to get them more involved in their disease process and help with compliance," he said.
"Since glaucoma is a disease of change, [the software] has a specific role in detecting glaucoma, although it could be used for other retinal diseases as well," Dr. Radcliffe concluded.