Medical	Ability to visually configure findings

Or is it? Like a modern-day scientific detective, Margaret Ballantyne, coordinator of the Canadian Ophthalmology Study Group (COSG), Kingston, Ontario, has been using a visual data analysis (VDA) software package, PV-WAVE^®, a VDA software program developed by Visual Numerics, to help Queens University Ophthalmology Department Chairman Dr. Alan F. Cruess and his team answer that question.

The stakes are high -- not only blindness or sight loss for millions of patients, but potentially millions of dollars in health costs as well.

In the early 1980s, it was demonstrated that laser treatment for some diseases of the retina could prevent people from going blind quickly, which could occur if the situation is left untreated.

Green laser light was found to have the most appropriate wavelength for the treatment of lesions outside the area of central vision, and other research suggested that red laser was more effective for lesions in the area of central vision, also called the foveal avascular zone or FAZ. The COSG set out to answer one question: Is red laser superior to green laser for lesions outside the FAZ?

Krypton red vs. argon green researchers at The John Hopkins University in Baltimore asked Dr. Cruess and his staff at COSG to undertake a comparative study to determine if krypton red treatment really was more effective than the conventional argon green approach. Dr. Cruess treated a random group of patients, some with the new krypton laser and some with the old argon laser. The patients were closely monitored after treatment, and voluminous data were gathered concerning the relative progress of the two groups.

It was at this point that Ballantyne and visual data analysis entered the project. The data collected were in the form of thousands of retinal photographs. Using PV-WAVE, Ballantyne was able to compare successive photographs of an individual patient to determine the rate of growth of laser scars. While the lesions are the preexisting medical condition or symptom of the disease, the scars are introduced during the laser treatment. Both the lesions and the laser scars were monitored in the project.

"In both diseases [AMD and OHS], the retina becomes hypoxic, meaning it desires oxygen," Ballantyne explained. " It's not getting enough, so it creates its own blood vessels to look for the oxygen. These new vessels can cause hemorrhaging, resulting in lesions that raise the retina slightly, lifting it off its lower layer. This is what causes the vision to become blurry or wavy."

But tracking the exact placement and growth of these minute lesions through photographs requires absolute precision. For the purposes for the study, only patients with lesions in a minute band some 2,000 microns wide surrounding the fovea, or center of the retina, were used. Such determinations were made in the lab by Ballantyne, using an Imapro QCS-35 scanner from Japan attached to a UNIX workstation running PV-WAVE to analyze photographic data.

Measurements were compared with data generated from an enhanced image on PV-WAVE, and a cursor was used to draw around the edges of retinal scars. According to Ballantyne, such detailed analysis would be impossible without this technology.

"It couldn't be done. If I want to look at the progression of scar growth, I can take a couple of patients and scan all their photographs in, bring one up on screen and look at the percentage area the lesion is covering," Ballantine said. "Then I can bring up a photo that was taken of the same patient three months later and get the updated percentage area of the lesion. PV-WAVE provides us with all of those numbers and allows us to visually configure our findings in a vast number of ways to calculate the progression of the disease and treatment."

"PV-WAVE gives us the ability to utilize our archived material and view it with high-resolution precision," said Dr. Cruess. "In the future, we hope to develop a package in which images from the eye are directly enhanced and then analyzed by the same sophisticated software."

At one point, Ballantyne received an assist from Mike Mayer, a technical support engineer at Visual Numerics' field office in Cleveland, in adapting the software to the exact needs of the study.

"He helped me write a program," said Ballantyne. "When I was presenting our findings using PV-WAVE to the Association for Research and Vision at its Florida conference, the doctors didn't think that the photographs taken on one day were comparable with the same patient photographed on another day -- the patient probably didn't sit in exactly the same place when he was photographed. They suggested I find some program to account for the change in position, so that I truly was comparing identical areas.

"Mike helped me design a program on PV-WAVE that allowed us to bring up the first photo of the patient, look at the healthy blood vessels flowing through the retina and match them up with the vessels from the same patient's photo months later. Then I could say the lesion covers a larger area six months later and show that indeed it was bigger, that it didn't just appear so because it was from a different camera angle."

Unconventional wisdom

Although preliminary data will soon be published, early results of the age-related macular degeneration study are already in.

"We found that we cannot agree with a popular position that the new krypton red treatment is better," Ballantyne announced. "In fact, our research bears out that the argon green laser is actually more effective than the krypton red laser in preserving visual acuity for AMD patients."

And what does this mean for doctors treating these patients?

"It means that if you have argon green as your mode of therapy in your clinic, keep it. There's no compelling reason to go and buy an expensive krypton laser. They're more expensive and harder to get."

This is good news for anyone working in the field of ophthalmology (and for their patients as well), and Ballantyne sees it as a particularly welcome result for her homeland of Canada.

"In Canada, it's much more difficult to pay for these lasers, because we have socialized medicine and the government has to say Yes to everything. This is good news for our government, because they don't need to run out and supply our health-care infrastructure with new lasers. And taxpayers can be happy about the cost savings."

Through the power of visual data analysis, Ballantyne's group has been able to show that an existing therapy is not outdated and does not need to be replaced.