Treatment search

Machine learning to speed brain injury treatment

A team of data scientists from the University of Pittsburgh School of Medicine in the United States and neurotrauma surgeons from the University of Pittsburgh Medical Center have developed the first automated brain scanners and machine learning techniques to inform outcomes for patients with severe traumatic brain injury. wounds.

The advanced machine learning algorithm can analyze vast volumes of data from brain scans and relevant clinical data from patients. The researchers found that the algorithm was able to quickly and accurately produce a prognosis up to six months after the injury. The amount of data reviewed and the speed with which it is analyzed is simply impossible for a human clinician, say the researchers.

Read more about machine learning in medicine: are machine learning tools the future of healthcare?

Publishing their results this week in Radiologythe scientists’ new predictive algorithm has been validated on two independent patient cohorts.

Co-lead author of the article Shandong Wu, associate professor of radiology, bioengineering and biomedical informatics at the University of Pittsburgh in the United States, is an expert in the use of machine learning in medicine . The researchers used “a hybrid model machine learning framework using deep learning and ‘traditional’ machine learning, processing CT imaging data and non-imaging clinical data for the prediction of outcomes of the patients with severe traumatic brain injury,” he said. Cosmos.

Wu says the team used data from the University of Pittsburgh Medical Center (UPMC) and 18 other institutions across the United States. “By using the machine learning model when the patient is admitted to the emergency room early, we are able to build a model that can automatically predict a favorable or unfavorable outcome or mortality or other potential for recovery” , he said.

“We find that our model maintains prediction performance, which shows that our model captures some critical information to be able to provide this type of prediction.”

Co-lead author Dr. David Okonkwo, professor of neurological surgery at the University of Pittsburgh and practicing neurosurgeon, also spoke with Cosmos. After presenting the same data to a small group of neurosurgeons, Okonkwo says “the machine learning model has significantly outperformed human judgment and experience.”

The success of the first model, based on specific datasets from the first hours of injury, is “extremely encouraging and tells us that we are on the right track here to build tools that can complement human clinical judgment to make the better decisions for patients,” says Okonkwo. But the researchers believe it can be made more powerful and accurate.

“The first three-day window is very critical for better or worse for patients with severe traumatic brain injury. The most common reason someone dies in hospital after a traumatic brain injury is when life-sustaining therapy is stopped, and it most often happens within the first 72 hours,” says Okonkwo.

“If we can build a model based on the value of information from those first three days, we believe we can put clinicians in a better place to identify patients who have a chance of a meaningful recovery.”

The study is one of many using machine learning in different areas of medicine, says Wu. ‘imaging or clinical and machine learning or deep learning to solve many other medical problems, diseases or conditions,’ he says.

“Our study on top of that, another strong study showing, you know, the critical care and severe trauma and brain injury population, how our techniques or how deep learning can provide more information, or additional tools to help physicians like David here provide better patient care. Okonkwo asserts that machine learning tools are not intended to “replace human clinical or human judgment, but to supplement human clinical decision-making.”