Artificial Intelligence Can Predict The Genetics Of Cancerous Brain Tumours In Less Than 90 Seconds

A study claims that using artificial intelligence might speed up the detection and treatment of gliomas by making it possible to check for genetic alterations in dangerous brain tumours in less than 90 seconds.

An AI-based diagnostic screening system called DeepGlioma was created by a group of neurosurgeons and engineers at Michigan Medicine in collaboration with researchers from New York University, the University of California, San Francisco, and other institutions. This system uses rapid imaging to analyze tumor specimens obtained during an operation and find genetic mutations more quickly.

The recently created system successfully identified mutations used by the World Health Organization to define molecular subgroups of the condition in a study of more than 150 patients with diffuse glioma, the most prevalent and lethal primary brain tumor. The average accuracy was over 90%.

Lead author and developer of DeepGlioma Todd Hollon, M.D, stated that the AI-based tool "has the potential to improve the access and speed of diagnosis and care for patients with deadly brain tumors.". , a neurosurgeon at University of Michigan Health and an assistant professor of neurosurgery at U-M Medical School.

Given that the advantages and risks of surgery vary among patients with brain tumors depending on their genetic make-up, molecular classification is becoming more and more important in the diagnosis and treatment of gliomas.

In contrast to other diffuse glioma subtypes, patients with the specific type of astrocytoma can expect to live an additional five years after having their entire tumor removed.

At centers that treat patients with brain tumors, however, access to molecular testing for diffuse glioma is restricted and not consistently offered. The turnaround time for results, according to Hollon, can take days or even weeks when it is available.

For patients with brain tumors, barriers to molecular diagnosis can lead to less-than-optimal treatment, complicating surgical choices and chemoradiation regimen selection, according to Hollon.

Before the development of DeepGlioma, there was no way for surgeons to tell diffuse gliomas apart during surgery. The system was developed in 2019 and uses deep neural networks to image brain tumor tissue in real-time using stimulated Raman histology, an optical imaging technique also created at U-M.

"DeepGlioma creates an avenue for accurate and more timely identification that would give providers a better chance to define treatments and predict patient prognosis," said Hollon.

Patients with diffuse glioma have few treatment options available to them, even when they receive the best standard of care. Patients with malignant diffuse gliomas have an 18-month median survival time.

Less than 10% of glioma patients are enrolled in clinical trials, which frequently exclude molecular subgroups from participation, despite the fact that developing drugs to treat the tumors is crucial. DeepGlioma might encourage early trial enrollment, according to researchers.

According to senior author Daniel Orringer, M.D, "progress in the treatment of the most lethal brain tumors has been limited in recent decades—in part because it has been challenging to identify the patients who would benefit most from targeted therapies". who invented stimulated Raman histology and is an associate professor of neurosurgery and pathology at NYU Grossman School of Medicine.

Rapid molecular classification techniques have great potential for redesigning clinical trial plans and bringing new treatments to patients.