AI predicts brain tumour genealogy in less than 90 seconds; here's how

Researchers have used artificial intelligence (AI) to improve the diagnosis and treatment of deadly brain tumours.

Michigan Medicine's neurosurgeons and engineers, in partnership with researchers from New York University, University of California, San Francisco, and others, created DeepGlioma: an AI-driven screening tool.

It uses quick imaging to swiftly examine tumour samples obtained during surgery and identify genetic mutations.

The tool DeepGlioma, which quickly analyses tumour samples taken during surgery to detect genetic mutations in less than 90 seconds.

It has been stated that this breakthrough has the ability to accelerate the diagnosis process and enhance the quality of care for patients facing life-threatening brain tumors.

Genetic Mutations

In a study involving more than 150 patients with diffuse glioma, a common and deadly primary brain tumor, researchers achieved remarkable results.

DeepGlioma, the AI system, exhibited an average accuracy exceeding 90% in identifying genetic mutations associated with different molecular subgroups of glioma, as defined by the World Health Organisation.

These impressive results were published in Nature Medicine, underscoring the effectiveness of DeepGlioma in molecularly classifying gliomas.

This molecular classification plays a critical role in diagnosis and treatment planning.

By accurately identifying genetic characteristics, DeepGlioma enables the customisation of surgical interventions and facilitates the selection of appropriate chemoradiation treatments for each patient.

Enhancing Timeliness

In the past, access to molecular testing for diffuse glioma has been inconsistent and limited across treatment centres.

Additionally, receiving test results could take days or weeks, causing delays in decision-making and optimal patient care.

DeepGlioma tackles the issue of time-consuming test result delays, which can take days or even weeks, by providing a rapid and reliable method for distinguishing diffuse gliomas during surgery.

It is claimed, this system combines deep neural networks with stimulated Raman histology, an optical imaging technique developed by the University of Michigan, enabling real-time imaging of brain tumor tissue.

If reports are to be believed, this groundbreaking approach shows great potential in enhancing both the efficiency and accuracy of tumor differentiation.

Promoting Clinical Trials

DeepGlioma brings forth opportunities for precise identification and prompt intervention, empowering healthcare providers to tailor personalised treatments and make more accurate predictions about patient prognosis.

DeepGlioma has the potential to address this challenge by facilitating early enrollment in trials, thereby fostering the development and assessment of targeted therapies.