Computer program lets people with Trigeminal neuralgia diagnose themselves

Trigeminal neuralgia, or TN, is a disorder affecting the areas of the face where the trigeminal nerve's branches are distributed, including the lips, eyes, nose, scalp, forehead, and upper and lower jaws. Often caused by an artery that compresses the nerve, the condition can bring about stabbing, mind-numbing, electric shock-like pain from just a light wind or a finger's glance of the cheek.

Believed to be among the most severe types of pain known to humanity, the most common forms of TN affect 1 in 15,000 to 20,000, but 1 in 5,000 are thought to suffer from some type of facial pain.

People with the condition "are begging to be killed," said Kim Burchiel, M.D., professor and chairman of neurological surgery at the Oregon Health & Science University School of Medicine who sees several new TN cases a week. "I'm telling you, it's total agony."

Just ask James Kirkpatrick of Vancouver. "If I lightly touch my eyebrow, it feels like I stuck my finger in an electrical outlet," said the 47-year-old construction manager, who has suffered from TN for 16 months. "I'll go outside and the breeze hits my face and it feels like the most intense toothache I've ever had in my life. It drops me to my knees almost. I can totally understand why they call it the suicide disease. It affects so many aspects of your life."

Then there's Melissa Hill, 55, a retired California attorney living in Charbonneau who suffered from TN for three years: "It's just very intense. Excruciating is a very good word for it. I would get it down to below my front teeth, on the right side, and right around my eye mostly. I'd get it under control and then it would flare up again, and every time it did, it would take longer and longer to get it under control."

But there is hope. Burchiel has pioneered a new method for classifying and diagnosing TN, and much of the diagnosis can be done by patients themselves. Burchiel's team has developed the world's first online, artificial neural network to be used as a diagnostic tool for facial pain syndromes. It's an artificial intelligence-based computer program built around a classification scheme that categorizes forms of the disorder, and it can be trained to recognize patterns in facial pain data and continually improve its accuracy in predicting the correct diagnosis.

"It's software designed to work the way the brain does," Burchiel said.

An accurate diagnosis means patients can more quickly seek appropriate treatment, he said. "A patient comes in and says, 'I'm having pain,' so a dentist might give him a root canal. The pain comes back, so the tooth is pulled out. People have all kinds of unbelievable things done before somebody finally says, 'You know what? Maybe this isn't your tooth.' To prevent unnecessary procedures, people need to be told early on what they have."

Since 2002, Burchiel has been using an 18-point questionnaire he developed to clinically diagnose facial pain using his novel classification scheme. Since facial pain types are based almost solely on a patient's medical history, the questionnaire is comprised of yes-or-no questions that tend to make a difference in establishing the target diagnosis, such as whether the patient has multiple sclerosis, has suffered a facial injury or has had surgery for facial pain, and whether the pain is on just one side of the face or manifests during certain daily activities, like shaving or eating.

Burchiel uses the responses to place a patient into one of seven TN categories: Type 1, or spontaneous, but severe and brief pain; Type 2, or spontaneous, more constant pain; trigeminal neuropathic pain from facial trauma, oral surgery, stroke or other causes; trigeminal deafferentation pain from procedures that intentionally injure the trigeminal nerve system to relieve other pain; symptomatic trigeminal neuralgia from multiple sclerosis; postherpetic neuralgia from trigeminal shingles; and atypical facial pain, which is more psychological rather than physiological in origin.

Burchiel doesn't think patients should have to wait until a clinic visit to get a diagnosis, so he developed the artificial neural network, which patients can easily, and confidentially, access on the OHSU Department of Neurological Surgery's Web site. Burchiel and neurosurgery department colleagues Farhad Limonadi, M.D., and Shirley McCartney, Ph.D., recently tested the network by asking 100 patients with facial pain to respond to the questionnaire during their first visits. They then interviewed each patient and made an independent diagnosis. Patients' responses and their diagnoses were input to the network, and a genetic algorithm was used to train the program.

The network correctly diagnosed most of the patients. It also correctly diagnosed trigeminal neuralgia Type 1 with 84 percent sensitivity and 83 percent specificity among another 43 new patients who used the network before their clinic diagnosis. Sensitivity is a measure of a diagnostic tool's ability to make a correct diagnosis, and specificity describes its ability to avoid a misdiagnosis.

"We hope that as people use this system, they become more and more informed, because they should be able to make informed choices before they actually get into treatment," Burchiel said.

One treatment Burchiel offers, and which Hill has received, is a surgical therapy called microvascular decompression. The procedure involves entering the brain through a small incision behind the ear, finding and exposing the trigeminal nerve with a powerful surgical microscope, and positioning a piece of Teflon between the nerve and the artery that's touching it and causing the pain.

In most cases, the procedure offers longer-term relief from facial pain than many, less-effective or inappropriate treatments people seek when they haven't been accurately diagnosed.

"People who have this condition are desperate for answers for what they have, so now, anywhere in world, somebody can log onto our Web site and basically diagnose themselves and go to the right resources," Burchiel said. "And we put the resources right there on the Web site."