Rare Daily Staff
A collaboration, led by Rady Children’s Institute for Genomic Medicine, utilizing machine learning and clinical natural language processing from Clinithink, delivered a provisional diagnosis of a rare genetic disease in a median time of less than a day.
According to investigators, the work represents the first time that rare diseases have been diagnosed using a supervised machine-learning system to analyze and interpret genetic disease testing results.
The diagnosis of a rare disease is a significant challenge. Many rare diseases can present with symptoms associated with more common diseases and some physicians have never seen a rare disease. It can take, on average, 4.8 years for an accurate diagnosis of a rare disease to be made, with patients seeing an average of 7.3 physicians on the diagnostic odyssey. The delays to diagnosis can allow a condition to progress unchecked or worsen because of inappropriate treatments.
In a study published in Science Translational Medicine, scientists documented their effort to diagnose neonatal and pediatric intensive care patients using electronic health records and genome sequencing. The Clinithink platform allowed investigators to use artificial intelligence to extract unstructured data from electronic health records of patients and then link phenotypic information with genetic results to find a diagnosis.
Michelle Clark, a statistical scientist at the Rady Children’s Institute of Genomic Medicine and lead author of the study, created an automated pipeline to analyze the data and deliver a potential diagnosis for hospitalized, often critically ill, children with suspected genetic diseases.
The Clinithink platform automatically extracts all of the clinical information that has been documented about that patient and compares the information to thousands of phenotypes and symptoms that are critical to the diagnosis of rare diseases. The investigators said the process required minimal user intervention, increasing usability and shortening time to diagnosis.
Although the approach the investigators used would need to be adapted for use at different hospital systems, the researchers said such an automated tool could aid clinicians to expedite an accurate genetic disease diagnosis, potentially hastening lifesaving changes to patient care.
“Some people call this artificial intelligence, we call it augmented intelligence,” Stephen Kingsmore, president and CEO of Rady Children’s Institute for Genomic Medicine. “Patient care will always begin and end with the doctor and by harnessing the power of technology, we can quickly and accurately determine the root cause of genetic diseases. We rapidly provide this critical information to intensive care physicians so that they can focus on personalizing care for babies who are struggling to survive.”
Photo: Stephen Kingsmore, president and CEO of Rady Children’s Institute for Genomic Medicine