AI-Driven Blood Test Could Aid Earlier Detection of Brain Cancer

WashU Medicine physician-researchers at The Brain Tumor Center at Siteman Cancer Center have co-developed an innovative approach that uses artificial intelligence (AI) to detect brain cancer, potentially leading to earlier diagnoses. Siteman is based at Barnes-Jewish Hospital and WashU Medicine.

The process incorporates a noninvasive blood test and machine learning that analyzes the blood sample for evidence of brain cancer – specifically, circulating DNA patterns associated with brain tumors. It then identifies repeating genomic patterns that indicate the presence of brain cancer. Such tests have already shown success in the earlier detection of lung cancer.

“We now have a method that detects brain cancer based on its unique characteristics, including DNA fragmentation and immune responses,” said WashU Medicine neurosurgeon Dimitrios Mathios, MD, who developed the test with Victor E. Velculescu, MD, PhD, co-director of the Cancer Genetics and Epigenetics Program at Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins.

“The signals we detect come from both the tumor itself and the immune system’s reaction to it,” said Mathios, an assistant professor of neurosurgery and director of the Lab for Molecular Neuro-oncology at WashU Medicine and a Siteman research member.

Mathios and Velculescu published their work April 29 in Cancer Discovery.

Detecting brain cancer before symptoms appear is challenging, often leading to diagnoses at more advanced stages when tumors are larger and located in high-risk areas. This complicates treatment, making it both riskier and less effective. The blood-brain barrier, which protects the brain, also prevents biomarkers — signs of cancer — from entering the bloodstream, further complicating early detection.

Using their new approach, the researchers successfully detected brain cancer in approximately 75% of cases from a cohort of 505 patients in the U.S. and South Korea, according to their paper, and validated their results in a separate group of 95 patients in Poland. In contrast, traditional blood-based liquid biopsy methods have detected brain cancer in fewer than 10% of cases.

A key factor in this success is the detection of immune system changes associated with brain cancer. Brain cancer often leads to immune suppression and alters the immune cell profile in the blood. These immune changes occur throughout the body, bypassing the blood-brain barrier and making them detectable, Mathios said.

In a simulation, the researchers modeled the potential benefits of using their method to screen the 10 million patients who visit emergency rooms or primary care clinics annually due to headaches. Normally, these patients are only referred for brain imaging if a physician suspects a problem. However, the simulation showed that incorporating blood-based liquid biopsy results could help identify nearly 1,700 additional cancer cases in the U.S.

The next step for the team is to conduct a larger prospective trial to confirm these findings in a broader population at higher risk for brain cancer.