(PDF 217 KB)
Since optical coherence tomography burst upon the scene in 2002, it has gradually become a workhorse of clinical practice. The continuing development of new applications prompts the question: What doesn’t OCT do? “Refraction,” quipped Jay S. Duker, MD. His joke isn’t far from the truth.
But, in a more serious vein, he continued: “OCT is the most important ancillary test we do. And it will become increasingly important as the capabilities of OCT expand.” Dr. Duker is director of the New England Eye Center and professor and chairman of ophthalmology, Tufts University.
“OCT has changed the way that we follow patients with retinal diseases,” said Abdhish R. Bhavsar, MD, director of clinical research, Retina Center of Minnesota, in Minneapolis. “It also has changed our ability to detect certain pathologies more readily.” On a busy day, his practice may see as many as 80 patients and use OCT 40 to 50 times. “It gets used a lot.”
Dr. Bhavsar warned, however, that OCT is just another adjunctive tool at the ophthalmologist’s disposal. He mainly uses it when the pathology is in the macula. “I don’t need OCT to tell me if a retina is detached or to fix a retinal tear,” he said. “The clinical exam is still of paramount importance in treatment decisions.”
The Power of OCT
Virtual histology. “Most people overthink OCT,” said Greg Hoffmeyer, who likened the noninvasive imaging technique to a virtual wire cheese cutter, slicing through layers of the retina and yielding images of each slice. Before OCT you couldn’t see that image without excising the retina and putting it under the microscope. “It’s essentially virtual histology.”
In addition, OCT can stack the images to render a cube of light-reflected data that offer volumetric and quantitative analysis. “You’re getting a stunning side-cut view of the retina that we now take for granted,” said Mr. Hoffmeyer, former manager of research imaging at Duke University, now business manager of Retinal Products, Eastern USA, Carl Zeiss Meditec.
A diagnostic tool. In today’s retina practice, the number one reason for using OCT is to determine the presence of subretinal fluid associated with AMD or intraretinal fluid in the case of macular edema, Mr. Hoffmeyer said. “Ninety percent of what ophthalmologists are looking at is the posterior zone, which encompasses the macula and optic nerve. A series of 6- to 9-mm scans should easily tell them what they need to know.”
Beyond that, OCT can sometimes elucidate why you’re seeing a certain condition, Dr. Bhavsar said. For example, vitreoretinal interface disease “is so exquisitely delineated on OCT.” Even though vitreomacular traction, macular hole, and epiretinal membrane are easily detected on clinical exam, you might not see the relationship of the vitreous adhesions to the retina. “With OCT, you have a much clearer picture of what’s really happening inside the retina, on the surface of the retina, and between the vitreous and the retina,” he said.
A tracking device. OCT changes the way we follow patients and consider treatments for AMD and DME, said Dr. Bhavsar, who uses it commonly for AMD re-treatment decisions and often for DME. In AMD, if he sees leakage in the angiogram, he doesn’t use OCT. But for follow-up, it can gauge progression or regression.
And Dr. Duker predicted that once treatments for dry AMD become available, OCT will reveal if the treatment is working.
A window on the choroid. “Until recently, we didn’t have a good way to image the choroid,” Dr. Duker said, explaining that OCT couldn’t penetrate deep enough. “We didn’t know what it looked like in vivo, in disease and in health.”
Both enhanced depth imaging (EDI) software and hardware improvements have changed that. The software, which is available on most SD-OCT devices, flips the image so the choroid is on top. It also oversamples each line scan by taking rapid multiple images at the same point. The software detects any image that’s different from all the others and, similar to Dolby audio technology, eliminates it as “noise.” Dr. Duker said, “With those two innovations you can see the choroid clearly in most patients.”
Is it important to measure the choroid? Probably, Dr. Duker replied. “We’re learning about diseases in which the choroid is abnormally thin.” For example, the choroid in patients with wet AMD is thinner than that in people with the dry form, and both groups have thinner choroids than age-matched individuals without AMD. “We don’t know why. We also don’t know if monitoring choroidal thickness might be another parameter to predict who will go from dry to wet and whether our treatments are working.”
Measuring the choroid may help pin down tricky diagnoses, as well. Idiopathic polypoidal choroidal vasculopathy (IPCV), for example, can look like AMD, except that the choroid is thicker than normal. A definitive diagnosis might improve IPCV treatment, which appears to respond better to a combination of anti-VEGF therapy and PDT or laser rather than anti-VEGF agents alone, Dr. Duker said.
Another condition that features thickening of the choroid, central serous retinopathy, doesn’t respond to anti-VEGF treatment. By measuring choroidal thickness, clinicians may avoid treating for AMD when that’s not the problem (Fig. 1).
|AMD or Not?
|Right eye of a 72-year-old man with 20/25 vision and minimal symptoms after three monthly intravitreal injections of ranibizumab for AMD. This treatment produced no change in clinical appearance. The presence of mildly thickened choroid on this OCT, coupled with relatively good vision and lack of a response to anti-VEGF therapy, suggests the possibility of an alternative diagnosis such as central serous retinopathy or idiopathic polypoidal choroidal vasculopathy.
A crystal ball? Some studies suggest that the condition of the photoreceptor inner segment–outer segment junction, as assessed by OCT, may predict outcomes after surgery for macular hole and other vitreoretinal interface disorders,1,2 said Dr. Bhavsar. He added that large-scale clinical trials are needed to assess OCT’s predictive power.
One Tool Among Many
What are the limitations of OCT? “It’s not really very good for subretinal tumors,” Mr. Hoffmeyer said. Although OCT may reveal a tumor or flat choroidal nevus, it might not measure the base of the lesion as well as ultrasound does. The pigment in the tumor often blocks the OCT signal. Likewise, blood from an intraretinal or vitreous hemorrhage can block the signal. “There are limitations to what this wavelength of light can penetrate.”
Also, a typical scan reveals only a small portion of the retina—6 to 12 mm long. “You’re not getting a pan-fundus image,” Mr. Hoffmeyer said. “You’re getting a slice here or a slice there.”
A continuing role for fluorescein angiography. OCT may be the most important ancillary test in the retina lineup, but it’s not the gold standard, Dr. Duker said. “It shows different information than fluorescein angiography and fundus photography.”
In diabetic retinopathy, for example, angiography reveals more about perfusion in the peripheral retina than OCT does. “OCT can see that to some degree, but fluorescein shows nonperfusion very well,” Mr. Hoffmeyer said.
Fluorescein angiography still plays an important role in Dr. Bhavsar’s practice, too. “I especially use it in questions of initial diagnosis of pathologies,” he said. For example, OCT can detect fluid within and under the retina, but it doesn’t reveal the etiology. The angiogram can tell whether the cause is diabetes, choroidal neovascularization, or central serous chorioretinopathy.
Fundus photography is still important. You sacrifice something by not doing fundus photography, said Mr. Hoffmeyer, explaining that OCT doesn’t render the retina and optic nerve in true color. “In glaucoma, there’s nothing like a 3-D photo of the optic nerve. It’s a near-perfect rendering of reality.” Similarly, he said, “There’s something to be said for a good clinical color photograph of the retina.”
|Pearls for OCT Use
DR. BHAVSAR: First, not all patients need an OCT. And, second, it has to be used as a clinical tool. The most important thing is for us all to learn how to properly evaluate the OCT images and understand the anatomy and pathology that’s being demonstrated. Or, if we are uncertain of the OCT findings, we should feel very comfortable sending the patient on to a vitreoretinal specialist for appropriate diagnosis and follow-up exam.
DR. DUKER: OCT has allowed us to quantify part of the treatment decision. But we’re nowhere near ready to base our treatment decisions on a number derived from an OCT. You still have to ask: What disease does the patient have? What treatments have been given already? How did the macula look on the previous exam? What’s the vision—in the affected eye and the fellow eye? What treatment are you contemplating, and what’s the risk of that treatment?
MR. HOFFMEYER: If you have SD-OCT, take advantage of what it can do. Get a lot of scans over a wide area to rule out any presence of fluids or pathology that you’re looking for. Don’t think you’re wasting time looking through several hundred scans generated by SD-OCT. Some doctors just want to look at the fovea, but there could be pockets of fluid beyond the fovea
Clinical Judgment Is the Key
Challenges abound in reading and interpreting OCTs. “We interpret OCT images in two ways: qualitative and quantitative,” Dr. Duker said.
The qualitative interpretation is fairly straightforward: You look at the image and classify it as abnormal or normal. However, doing this accurately requires knowledge of normal anatomy and a deep understanding of the pathological processes that alter retinal morphology, he said. You also need to recognize the clinical importance of specific features, such as fluid under the retina. For example, if fluid accumulation is not under the fovea, it may not be significant, and the patient may not need treatment. “Location is everything.”
Quantitative reading, such as measurement of the retinal nerve fiber layer, may be trickier. That number may be skewed by artifacts, resulting in an inappropriate treatment decision, Dr. Duker said.
OCT is not the doctor. OCT can confirm what the doctor sees on clinical exam and may even reveal what is not clinically observable. And it provides measures to gauge the effectiveness of treatment. But “as good as it is, it ultimately doesn’t diagnose anything—yet,” said Mr. Hoffmeyer. “That’s up to the doctor.”
Dr. Duker agreed. “Life would be simple if the machine just said, ‘normal versus abnormal; treat versus watch.’ The nice thing about OCT is, we still need to be doctors. This is an ancillary test. We have to put it in context.”
On the Horizon
As OCT technology advances, it will increase the sensitivity and specificity of early disease detection and improve the monitoring of disease progression and therapy. It may also provide insight into retinal biology and functioning at the micrometer level, Dr. Duker predicted. Following are some of the techniques being explored.
Swept-source laser (SS-OCT) provides a greater range of imaging depth and better penetration through ocular opacities. In addition, its faster image acquisition—up to 300,000 A-scans per second—may reduce motion artifact.
Doppler OCT measures blood flow velocity through retinal vessels by assessing light reflectivity changes in the vessels over very short periods of time. It may prove useful in retinal vascular diseases such as diabetic retinopathy and vascular occlusive diseases. Dr. Duker predicted that high-speed Doppler imaging could replace angiography. And though the images aren’t of a high enough quality yet, one day the technique could replace fundus photographs with OCT fundus images.
1 Inoue M et al. Retina. 2011;31(7):1366-1372.
2 Oh J et al. Invest Ophthalmol Vis Sci. 2010;51(3):1651-1658.
Dr. Bhavsar reports no related financial interests. Dr. Duker receives research support from Carl Zeiss Meditec, OptoVue, and Topcon Medical Systems; is on the scientific advisory board for Paloma Pharmaceuticals; and is a consultant for Alcon and EMD Serono. Mr. Hoffmeyer is business manager of Retinal Products, Eastern USA, Carl Zeiss Meditec.
|SD-OCT or TD-OCT?
|Your office has one of the original time-domain OCTs. Is it time to buy a newer, spectral-domain device?
A study comparing TD-OCT and a variety of SD-OCT devices in AMD patients after ranibizumab injections found that SD-OCT was more sensitive in detecting choroidal neovascular activity in this group. The images shown here, from the same patient, were taken with TD-OCT (top) and SD-OCT (bottom); intraretinal (arrowheads) and sub-RPE fluid (arrows) are visible in both views.1 But how significant are these findings to clinical practice?
The SD-OCT is 100 times faster than the original, acquiring 25,000 to 52,000 A-scans per second, compared with 400 A-scans per second for the first-generation OCT, Mr. Hoffmeyer said. But several hundred medium-resolution scans encompassing the area of pathology and beyond will usually tell you what you want to know.
Dr. Bhavsar’s office has tested many SD-OCTs, but he’s sticking with TD-OCT for now. “At this time, our big clinical questions are whether there’s fluid or not, and the degree of fluid. We can see that satisfactorily with our existing machines.”
Dr. Duker agreed. SD-OCT is superior for observing small structural changes of the retina, but if you’re mainly interested in measuring thickness of the retina, “the TD does a pretty good job.”
In fact, SD-OCT alters patient care in AMD in only a small percentage of eyes, Dr. Duker said, citing a study out of Doheny Eye Institute in which readers compared images from both SD and TD devices. The study, unpublished but reported at meetings, found that although SD-OCT revealed much more detail than TD-OCT, that extra information changed care in less than 10 percent of cases.
The bottom line, said Dr. Duker, is that conventional OCT is adequate most of the time because “in diseases of the retina, we’re mostly interested in the full thickness of the retina.”
1 Sayanagi K et al. Ophthalmology. 2009;116(5):947-955.