When Bryan Shaw’s wife first told him she was concerned about their son’s eyes, he brushed it off. She was worrying unnecessarily, he said. It was only later, when the parents went to a doctor for a routine check-up, that he realized his wife was right to worry. Noah, only four months old at the time, had a rare cancer called retinoblastoma. The doctor could see as much from his milky white pupils.
Noah was lucky, in a way. Though his parents had to take out his right eye to stave off the tumors, it could have been worse. Some kids die when the condition isn’t diagnosed early enough. The cancer goes directly down the optic nerve to the brain, and there’s nothing to be done. Of the 8,000 cases in the world each year, about half the children don’t make it to their fifth birthday.
Still, Shaw, who is a professor of chemistry at Baylor, felt guilty, and wondered if there was something he could have done differently. And he realized that perhaps there might have been. When he looked at early pictures of Noah, he could see evidence of leukocoria, or “white eye”–the main symptom of retinoblastoma–as early as 12 days after his birth. He wondered if digital photos might hold the key to quicker diagnoses.
Working with a team of developers, Shaw has since developed an app to do what he was unable to figure out on his own. The iPad version of the app uses the machine’s built-in eye detector to first zero-in on a kid’s eyes and then compare the image against a growing database of leukocoria cases. The iPhone version automatically checks photos as they’re taken and scans existing image libraries as well. The app, available at the end of the summer, will be free to download.
Shaw says parents are probably in a better position to find early signs of leukocoria than professionals are. Pediatricians will do a quick check with a flashlight when a baby is born. But a camera flash is faster than clicking on a flashlight, and gives a more natural look inside the pupil. Plus, parents take thousands of pictures, which raises the chances of spotting a problem. One study showed that 80% of retinoblastoma diagnoses began with parents, and not with physicians.
“Here in America, or Europe or Japan, it will speed up diagnosis. Noah might not have lost his right eye [if we’d caught it earlier] and he certainly would have needed less radiation,” Shaw says. “Overseas, it could save lives. You think about a place like Namibia, where the survival rate is about 45%. Their access to pediatric care is probably going to increase at a much slower rate than access to digital photography.”
In fact, leukocoria is a symptom for several diseases, including pediatric cataracts. The reach of the app could extend far beyond retinoblastoma detection. Eventually, Shaw hopes to embed the software online–for example, on Picasa or Facebook–so anyone can use it with an online image library.
“The challenge is we have to balance the ability of the software to detect the leukocoria with the ability to not crank out too many false positives,” Shaw says.
You don’t want the app beeping red every time it spots white eye. Take a photo from the wrong angle and anyone’s eyes look strange. The app needs plenty of examples before alerting parents they might have a problem. Even then, kids still need to go a doctor to fully checked out. The software is just an early warning system–though a potentially important one.