Optical Coherence Tomography: Detecting Disease Early Through Retinal Imaging

September 2012
Topics: Health Innovation
Your eyes are said to be the window to your soul.

The Telltale Eye

Early diagnosis of individuals with chronic disease is a global health goal because the sooner a physician can detect a disease in a patient, the better the prognosis. In the U.S. alone, more than 6 million people have diabetes without knowing it, according to a 2009 study. That study estimated the cost of undiagnosed type-2 diabetes in medical services and lost productivity to be $18 billion a year.

Eye researchers have observed that many chronic diseases, such as diabetes, hypertension, and cardiovascular and cerebrovascular conditions, leave telltale clues of their presence in the retina. By detecting these clues using novel retinal imaging technologies, physicians can treat the underlying diseases earlier, more successfully, and at a lower cost.

One such imaging technology is optical coherence tomography (OCT), which was introduced in the early 1990s. OCT scans a near-infrared laser beam multiple times across the retina (with no discomfort to the patient) and measures the depth and intensity of the reflected light, from which it forms a three-dimensional image of the retina at a micrometer resolution. These 3-D models allow clinicians to see features, such as individual retinal layers and the presence of macular fluid, that are invisible to the naked eye or to other imaging technologies.

The non-invasive and highly repeatable OCT procedure lends itself well to providing a fully automated screening process. With a simple, painless sweep of a laser, clinicians will be able to diagnose diseases that once would have escaped early detection. Using automated OCT scans and diagnoses, large populations could be screened for early signs of disease. Such an automated screening process would free up clinicians to focus on patients with critical needs.

Optical Coherence Tomography:

OCT Challenges

Taking best advantage of the diagnostic promise of OCT technology presents several challenges. First, for devices such as OCT to serve as standard disease screening tools, they must become more cost-effective, scalable, standardized, and accessible to patients.

Second, the cost and time needed for OCT imaging interpretation must be reduced. Because OCT images incorporate a rich amount of data, it would be impractical for clinicians to manually analyze the retinal images of large populations in search of signs of disease.

Third, clinicians will need to know which diseases leave behind which clues in the retina. This will be difficult because currently there is insufficient data to help researchers identify novel disease predictors.

If we can overcome these challenges, there is an opportunity to cost-effectively perform routine disease screening.


To help develop OCT-based disease screening capabilities, MITRE is pursuing a research project called EyesFirst. The key goal of the project is to work with clinicians and biomedical researchers to advance the state of the art in OCT image analysis techniques to where it is possible to automatically detect early signs of multiple chronic diseases.

MITRE hopes to demonstrate the value of employing an open, standards-based architecture to develop analysis tools that will aid the clinician in interpreting OCT scans and double-checking diagnoses. With EyesFirst, clinicians could compare a patient's history of OCT images to assess disease severity and track disease progression. They could also create statistical models to help identify subtle deviations not visible to the human eye and to classify scans as being normal or diseased.

As a byproduct of the EyesFirst project, MITRE also hopes to establish a database of OCT retinal images tagged with medical history information (minus any personally identifying data). This database could enable correlations between retinal changes and the onset and progression of chronic disease, leading to the identification of novel disease predictors. Such a database will also allow for the validation of automated OCT analysis tools.

EyesFirst is an open-source project; MITRE will freely share our results (including data and OCT image analysis algorithms) with the scientific community. In this way, our ideas and findings can be verified, extended, and put into action.

Self-Service Disease Screening

As we develop faster, better, and cheaper OCT devices, as we discover new retinal predictors of disease, and as we refine automated diagnostic tools and processing algorithms, OCT technology may become a routine part of healthcare exams. Ultimately, there may be self-service kiosks at your local Wal-Mart, similar to today's blood pressure instruments, where shoppers can enter the kiosk, get their eyes scanned, and receive a printed report with the results of the analysis and any necessary prompt to contact a physician for a follow-up exam. Patients will gain more control over diagnostic procedures and over their healthcare data. The role of physicians will shift from diagnostic to therapeutic. OCT technology will be one more step in making our nation's healthcare system more effective and less costly.

—by Salim Semy


Publication Search