American Academy of Ophthalmology Web Site: www.aao.org
Original URL:

November/December 2005

 
Feature
OCT in Practice: Dos and Don'ts
By Miriam Karmel, Contributing Writer
 
 

There's lots of buzz around OCT, but what are its uses in practice, and when should you stick with older, tried-and-true methods of imaging?

Optical coherence tomography may not be as pervasive as the ophthalmoscope, but it has definitely gone mainstream. Manufactured by Carl Zeiss Meditec, and originally used as a research tool, OCT has drawn ophthalmologists to buy thousands of subsequent iterations. Stratus, the most recent, was introduced in 2002. “OCT just exploded on the scene,” said Greg Hoffmeyer, BFA, CPT, ophthalmic photographer and manager of research imaging at Duke University. “OCT is truly a new way to look at tissue in vivo, with widespread applications.”

The technology has radically changed the way the retina is evaluated. Prior to OCT, the only way to obtain high-resolution cross-sectional images of the retina was on pathology slides obtained after enucleation, Mr. Hoffmeyer said. “What we have now with OCT is essentially a virtual biopsy.”

A machine that multitasks. OCT is a technology similar to ultrasound, except that it measures the time delay and intensity of reflected light rather than acoustical waves. In so doing, it can provide three things: cross-sectional images of the macula, topographic images of the optic nerve and measurements of the thickness of the retinal nerve fiber layer. In addition, OCT is noninvasive and can help diagnose and manage a variety of retinal diseases, as well as glaucoma. And it is fast, performing high- and low-resolution scans in a little over a second.

OCT may soon provide standardized interpretation of ocular structure, which co-inventor Joel S. Schuman, MD, said, “brings everyone in the community up to the level of an expert observer.” Dr. Schuman, professor and chairman of ophthalmology, University of Pittsburgh, and director of the eye center there, also noted that other optic imaging devices, GDx (laser polarimetry) and HRT (the Heidelberg Retina Tomograph, or laser tomography), provide a similar function. But OCT is the newest of the technologies and may be on track to become the most widespread.

New to the human eye. OCT reveals details that have never been seen with the ophthalmoscope. It has shown, for example, that macular holes are a result of obliquely oriented vitreoretinal traction, said Michael S. Ip, MD, associate professor, Fundus Photograph Reading Center, University of Wisconsin. “We didn’t know that in the past. We didn’t have any images to show how macular holes were forming. With OCT we can follow the patient from stage one to full-thickness macular hole.”

Not picture perfect. But OCT is not a magic bullet, and it cannot stand alone. “It’s a computerized tool. It’s not an eye doctor,” Mr. Hoffmeyer said. “Not every OCT is a revelation. In fact, you really have to be wary of what you’re getting.” Ophthalmic photographer Marshall E. Tyler, CRA, FOPS, agrees. Photographers, he explained, are still learning how to image the layers of the retina properly, and physicians are learning how to interpret the information. “It truly is new stuff,” said Mr. Tyler, instructor at Wake Forest University Eye Center.

“We’re definitely on the learning curve,” said Frank J. Moya, MD, assistant professor of ophthalmology, Duke University. “We have not plateaued in our ability to determine what’s the best way to use OCT. There’s still so much more to learn.” Still, Dr. Moya uses OCT on all his glaucoma suspect and glaucoma patients.

There’s no reason not to, said Dr. Schuman. Though OCT is still evolving, he said, “We’re at a point where it’s a usable tool.”

Rohit Varma, MD, professor of ophthalmology and preventive medicine, University of Southern California, worries that physicians may be relying too much on overly simplified, and sometimes erroneous, computer-generated interpretations of the scans. OCT is useful for helping to buttress a diagnosis, he said. “It’s one more important tool in your armamentarium. You just don’t want it to be your only tool,” he said.

Following are some ways these ophthalmologists are fitting OCT into their practices.

Uses in the Glaucoma Practice

“The thing that people have been looking for in glaucoma is the one machine that will say, ‘Yes glaucoma,’ or ‘No glaucoma,’” said Dr. Moya. OCT cannot make that call. Yet used in combination with family history and other tests, such as frequency doubling technology, short wavelength automated perimetry or pachymetry, OCT helps Dr. Moya determine which patients might need treatment sooner, rather than later. “OCT is just another piece of the puzzle.”

Nerve fiber analysis. Dr. Moya finds OCT most useful with glaucoma suspects who may not have obvious field defects on standard perimetry. In particular, if he observes thinning or defects of the nerve fiber layer and then sees corresponding abnormalities on other tests, he is more confident of a diagnosis. But, he added, “I personally would not act solely on OCT.”

Macular and optic nerve scans. While Dr. Moya finds the nerve fiber analyzer thickness scan most useful, he also uses the macular and optic nerve scans. The macular thickness test is not as sensitive as the nerve fiber thinning test for detection of glaucoma, he explained. “But if I see thinning of the nerve fiber layer superiorly, and see a corresponding decrease in the macular volume superiorly, that makes me feel more comfortable that I’m observing a true defect in that area.”

Optic nerve head analysis. OCT is useful for red-flagging the optic nerve head. A patient with a very large nerve head might be expected to have a thicker than average nerve fiber layer on the standard scan. “I have found that larger optic nerves have correspondingly thicker nerve fiber layers, as the testing occurs closer to the optic nerve, where we know the nerve fiber layer is thicker,” he said. But if the OCT reveals a large optic nerve head in a patient with a very thin nerve fiber layer, that incongruence might suggest an abnormality.

Setting target pressures. Though there is no literature to support this—only clinical observation—Dr. Moya finds nerve fiber layer images useful in setting target pressures. Seeing how much tissue is lost in the retinal nerve fiber layer helps him determine how aggressive he needs to be. He tends to be less aggressive in patients with a significant amount of cupping but who still have a good amount of nerve fiber tissue left. He’s more aggressive in patients with thinner nerve fiber layers.

Don’t forget to look at the patient. Experts agree that the technology is no substitute for a clinical exam. Dr. Schuman said that an abnormality on OCT always points him back to the patient. “I should be able to see a structural abnormality that corresponds to that OCT finding,” he said, adding that he might not have seen it without OCT sending up a red flag.

Can confuse progression? What OCT doesn’t do as well, said Dr. Moya, is detect progression. Though there are nerve fiber layer progression analyses that compare to baseline readings, he said he can’t be sure that the same tissue is being measured on follow-up exams. Because of that uncertainty, he’s not likely to change the course of therapy, or to initiate therapy, based on a possible progression noted on OCT.

Dr. Varma shares that concern. “It is important to caution people that while we have made important and significant advances in the technology and interpretation, we are not yet there in terms of interpreting and assessing change accurately.” And change, he noted, is the cardinal sign of glaucoma.

OCT cocreator David Huang, MD, PhD, agreed that we don’t yet know whether all the OCT measurements are reliable. “We still need more experience to know when you see a defect on these imaging tests, whether it’s an artifact or correlates with visual field loss now or in the future.” Dr. Huang is associate professor of ophthalmology, University of Southern California.

As Dr. Schuman put it: “For progression, I wouldn’t trust OCT in a vacuum.” On the other hand, his group has shown that more progression events were detected with OCT than with perimetry over a six-year period.1 “That means OCT is more sensitive to progression, but we don’t know about the specificity.”

Still, OCT can be useful if it shows a progression event that also appears on a visual field, Dr. Schuman said. “It can be used in synergy with visual field so you don’t have to repeat the field multiple times.”

While the utility of OCT may be uncertain in detecting progression, Dr. Schuman noted that there’s also a lot of variability in the gold standard—visual field testing. “In a way, we’re searching for the truth and we don’t know which device is giving that to us.”

He cited another study which showed that for glaucoma detection, OCT proved “the most sensitive and specific tool that we have in ophthalmology. Better than IOP; better than visual field.”2 That’s also true for GDx and HRT, he said.

Glaucoma imaging project. A new, long-term study sponsored by the NEI—the Advanced Imaging for Glaucoma (AIG) project—may definitively answer questions about the relevance of new imaging technologies.

The largest arm of the five-year study follows glaucoma suspects and preperimetric glaucoma patients to test the hypothesis that advanced imaging can detect glaucoma progression prior to visual field and optic disc photography. Included are the latest generations of advanced imaging instruments: optical coherence tomography (OCT), scanning laser polarimetry (GDx) and scanning laser tomography (HRT).

“Right now, you might see an abnormality in OCT, for example, and you would interpret it as glaucoma. But you don’t know for sure how well this relates to future changes in visual field or the patient’s vision,” said Dr. Huang, AIG principal investigator. “You just know that there is some anatomic problem.”

But Dr. Huang said the study could lead to better understanding of how to interpret the images generated by these devices. In fact, data from another study recently found that irregular baseline HRT results—used alone or with other measurements—may be associated with the development of primary open-angle glaucoma.3 Results from studies like these may answer important questions: When you see an abnormality or change on an advanced imaging test, does this predict future visual change? Is treatment warranted? (The AIG Study is recruiting normal, glaucoma suspect and glaucoma subjects in Los Angeles, Pittsburgh, Palm Beach and Cleveland. For more information, call the toll-free number 866-450-6782 or e-mail aigs@usc.edu.)

Uses in the Retina Practice

“OCT has really made a big change in how retina specialists think about their patients,” said Dr. Ip. What’s more, he believes it is more widely accepted as a measure for following conditions of the retina than it is for glaucoma. “OCT represents a major advance in ancillary testing as it relates to retinal diagnoses.”

Retinal thickness measurements. In daily practice, he finds it useful for measuring the thickness of the retina in conditions such as diabetic macular edema, retinal venous occlusive disease and even macular degeneration. With AMD patients, for example, OCT is good at detecting subretinal fluid and differentiating it from increased retinal thickness. It also shows the anatomical changes that can occur in a way that fluorescein angiography does not. “Fluorescein is more of a functional assessment, while OCT is more of an anatomic assessment,” Dr. Ip said. Because OCT can detect the vitreoretinal interface with good resolution, it makes it easy to document and follow conditions such as epiretinal membranes and secondary thickening.

OCT is also helpful in distinguishing macular holes from lesions such as epiretinal membrane with a pseudo hole, and it can readily identify stage one macular holes, Dr. Ip said.

Patient education. Finally, he finds OCT to be a useful teaching device. For example, he can show a patient a scan of an eye with a macular hole and a normal eye, and then show a postsurgical scan. He can’t do this with fundus photographs or angiography. But “OCT images are so intuitive that most patients will understand what I’m talking about when I show them the images,” Dr. Ip said.

Watch for pitfalls. The algorithm for measuring inner and outer retinal boundaries is automated, and is prone to making mistakes when taken out of clinical context, Dr. Ip said. If the boundaries are mistakenly placed, OCT can minimize or underestimate retinal thickness measurements. “One has to use retinal thickness measurements carefully and make sure they make sense in the overall clinical context.”

Bright future. Current research with ultrahigh-resolution OCT is expected to offer even better images than standard-resolution OCT.4 Dr. Huang predicts that improved hardware and software will lead to “amazing advances” with this technology. He is now working with a prototype that is 73 times faster than the current model and should provide more detailed anatomic information for both glaucoma and retina applications.

He also noted that his group showed that the accuracy of retinal thickness mapping can be improved with new software. “Future improvement in speed, resolution and image processing will make the results more robust. But we will still need to disseminate the knowledge on how to judge image quality and how to make interpretations in the clinical context.

______________________________
1 Arch Ophthalmol 2005;123:464–470.
2 Medeiros, F. A. et al. Arch Ophthalmol 2004;122(6):827–837.
3 Zangwill, L. M. et al. Arch Ophthalmol 2005;123:1188–1197.
4 Ko, T et al. Ophthalmology 2005;112:1922–1935.

OCT vs. Angiography

Ophthalmic photographer Marshall Tyler, CRA, FOPS, places OCT in the pantheon of retinal imaging devices. OCT is the fifth, and latest, “milestone,” after black-and-white fundus photography, color fundus photography, fluorescein angiography and fluorescein angiography with interference filters, he said.

Lately, however, Mr. Tyler has seen a trend back to fluorescein angiography, which he said is useful for establishing the location of the pathology. OCT, on the other hand, can better track a patient’s response to treatment.

Rather than pit OCT against angiography, ophthalmic photographer Greg Hoffmeyer, BFA, CPT, prefers to regard them as complementary tools. OCT, for example, cannot tell if choroidal neovascularization is classic or occult, cannot gauge circulation for vascular occlusion studies and cannot show tissue perfusion or nonperfusion.

“Comparing OCT to angiography now is like comparing a 2-D sonar image to an aerial satellite image of the ocean. They are both images of the ocean but offer two different types of information,” Mr. Hoffmeyer said. “One isn’t necessarily better than the other, it’s relative to what you are looking for.”

Michael S. Ip, MD, agreed. He regards OCT as “quite an advance from when we just had fluorescein angiography and fundus photography. OCT hasn’t replaced anything. It is an addition to our armament of diagnostic tools.”



Getting the Most From OCT and Other Devices

When one worried patient came to Dr. Varma for a second opinion after being told he needed surgery, Dr. Varma found no evidence of glaucoma damage. The first opinion was based on an abnormal OCT reading, but another OCT scan was normal. And so was the clinical eye exam.

How can two doctors use the same device and come up with two completely different readings? According to Dr. Varma, some eye care providers are making treatment decisions based on overly simplified, and often unreliable, interpretative information generated by OCT (and other optic nerve imaging devices, too). These devices generate numbers based on certain structural characteristics of the optic nerve and then produce color-coded interpretations. With OCT, green is good; yellow is borderline; red is bad. Until recently, there was no way of assessing the quality of the generated data. “So people would look at these printouts and say, ‘Ah, it’s all green; it’s fine.’ Or, ‘It’s red; we need to operate.’” But if the data are unreliable, the physician can make a wrong decision, he said.

In the case of the worried patient, the first opinion was based on an image with Signal Strength of 2. Signal parameters pertain to the quality of the laser signal that was collected from the eye at the time of imaging. Any Signal Strength 7 and higher is considered acceptable. Dr. Varma contends that doctors must check that the signal parameters are in an acceptable range. Unless that is the case, they should disregard the interpretive data.

Dr. Schuman agrees. “You’ll get a measurement,” he said, “but sometimes it’s going to be garbage.” Physicians, he said, must be astute observers to make sure the quality of the image and analysis are adequate. Here is what the experts advise on getting the most from OCT.

Rohit Varma, MD:
“If it’s not good quality, and you think you have a good chance of getting a good image, then repeat the OCT. We do repeat testing for visual fields. We need to hold imaging data to the same standards that we hold other testing data—reproducibility and reliability.”

Marshall E. Tyler, CRA, FOPS:
Provide the photographer with sufficient information. The photographer needs the working diagnosis, as well as the type of scan that’s desired.

Frank J. Moya, MD:
“Early on, the physician needs to take the time to see how the test is being done and do some quality control and be sure the tests are being completed in a proper fashion.”

Michael S. Ip, MD:
Sometimes you can keep retaking a scan and get an artifact. But an experienced photographer can minimize this problem. Keep the surface of the cornea moist. Dilate the pupil. Clear media helps. “In general, patients with less pathology have a better chance of getting a good OCT scan.”

Greg Hoffmeyer, BFA, CPT:
It’s essential to check that the retinal map printout is worthwhile. The maps can be very accurate when [you are] looking for things such as retinal thickening from edema. But when you have pathology anterior to the retinal pigment epithelium, like drusen, blood or CNV, at times they’re relatively the same reflectivity as the RPE, and will either cast a shadow, or light will bounce off this pathology and skew the accuracy of the map data and topographic pseudocolor contour rendering.



MEET THE EXPERTS

Greg Hoffmeyer, BFA, CPT, manager of research imaging, Duke University Eye Center and senior technical analyst at the OCT Reading Center, Duke University. Financial interests: Occasional consultant and lecturer for Carl Zeiss Meditec Inc.

David Huang, MD, PhD, associate professor of ophthalmology, University of Southern California. Financial interests: Receives a royalty for the OCT systems now being sold, and gets research support from Carl Zeiss Meditec Inc.

Michael S. Ip, MD, associate professor, University of Wisconsin, Fundus Photograph Reading Center. Financial interests: None.

Frank J. Moya, MD, consulting assistant professor of ophthalmology, Duke University. Financial interests: None.

Joel S. Schuman, MD, professor and chairman of ophthalmology, University of Pittsburgh, director of University of Pittsburgh Medical Center Eye Center. Financial interests: Receives royalties from intellectual property regarding OCT as licensed by MIT to Carl Zeiss Meditec Inc., and receives research support from Heidelberg and Carl Zeiss Meditec Inc.

Marshall E. Tyler, CRA, FOPS, instructor, department of ophthalmology, section of surgical sciences, Wake Forest University Eye Center. Financial interests: None.

Rohit Varma, MD, professor of ophthalmology and preventive medicine, University of Southern California, and director of the glaucoma service and director of the Ocular Epidemiology Center at USC. Financial interests: None.