As reported in Dravet Syndrome News, in a new study, researchers created an artificial intelligence system which could diagnose epileptic conditions more rapidly and easily than existing systems. Epilepsy is an umbrella term for a large number of conditions that cause seizures. Because it is made up of many different conditions, there are many different causes too. Understanding the specific underlying condition of the seizures is vital for finding the appropriate treatment for a given patient. The researchers ran their AI systems on a patient living with the rare disorder Dravet syndrome to test and demonstrate their technology’s capabilities to correctly diagnose epileptic diseases.
Efforts to Distinguish Between Epileptic Conditions
In 2017, The International League Against Epilepsy (ILAE), set out to distinctly categorize the diseases and created a comprehensive set of guidelines for types of epilepsy. To make such guidelines, they gathered data on the causes of the seizures and the presence of other symptoms, as well as common correlations with other medical conditions including psychiatric disorders. To find this information on the many common and uncommon conditions that cause seizures, there needs to be a compilation of information on the many patients with all of this information and the specifics of their disease. They had to find patients and their medical histories, such as type and duration of seizures, triggers, age of onset, genetic testing, and any other unique experiences. This work to create clear boundaries between the specific disorder is necessary before diagnosis is possible. The researchers of this study have created an artificial intelligence to use this information to provide a diagnosis.
How AI Diagnosis Patients
Their program plugs an undiagnosed patient’s data into their system to find which conditions are possibilities, and which are more likely.
“If the only type of seizure is epileptic spasm and the first seizure occurs when the patient is younger than 1.5 years old, we will consider the possibility of the patient having West syndrome. For another example, an early childhood onset seizure with myoclonic-atonic seizure should be considered Doose syndrome.”
As they include more facets and symptoms, they fill in the full picture on conditions that may not already have consistent data collection. This means their technology is not only helping patients receive diagnoses, but it is bettering the system as it diagnoses new patients.
Dravet syndrome only received a distinct disease category in 1978, and much information still needs to gathered. It is known for severe myoclonic epilepsy in infancy which can be dangerous if they last more than 10 minutes, are only on one side of the body, or are triggered by warm-water before a year of age.
The team used an adolescent male who was suspected to have Dravet syndrome when he was 10 due to the seizures he had endured almost of all of his life. They gathered thorough information on his condition and the time line of the onset to eliminate many possible conditions.
They used this rare illness to show how an earlier diagnosis can allow doctors to treat it before it causes more damage. For Dravet syndrome, they note that patients risk speech and psychomotor delay when unaddressed.
The makers of the study acknowledge that the technology must be maintained and continuously updated even once it’s finished. It also is not necessarily conclusive like a full gene sequencing; rather, it narrows in one the possibilities that should be followed up with further testing.
The system is exciting for members of the rare epilepsy community. Receiving a diagnosis is vital to getting access to treatments and to furthering research. By finding a more effective way to define and diagnose conditions of epilepsy, researchers are getting closer to getting those patients the treatments they need when they need them.