AI-Based Liquid Biopsy Approach to Revolutionize Brain Cancer Detection

Detecting brain cancers remains extremely challenging, with many patients only receiving a diagnosis at later stages after symptoms like headaches, seizures, or cognitive issues appear. Late-stage diagnoses often limit available treatment options and result in poorer patient outcomes. Efforts to identify biomarkers in a patient's blood have previously struggled due to the blood-brain barrier, which, while protecting the brain from infection, also prevents potential biomarkers from reaching the bloodstream. To address this issue, researchers have developed a new liquid biopsy method that could transform brain cancer detection by identifying DNA fragments from tumors and immune cells circulating in the blood, potentially allowing for earlier diagnosis. A description of the work has been published in Cancer Discovery.

The innovative liquid biopsy technique, developed by researchers at Johns Hopkins Kimmel Cancer Center (Baltimore, MD, USA), has already proven effective in detecting lung cancer and is currently used in clinical settings for lung cancer screening. This method utilizes machine learning, a form of artificial intelligence (AI), to detect patterns in circulating DNA fragments linked to brain tumors. Additionally, it identifies recurring patterns in the genome associated with brain cancer. In a study involving 505 patients from the United States and South Korea, this approach successfully detected brain cancer in approximately 75% of cases. The team later confirmed these results in a second group of around 95 patients from Poland. In contrast, previous liquid biopsy methods have detected brain cancer in fewer than 10% of cases.

Image: Release of cell-free DNA and altered blood cells in patients with cancer (Photo courtesy of Carolyn Hruban/Johns Hopkins)

A key factor contributing to the success of this approach is its ability to detect immune changes related to brain cancer. Patients with brain tumors exhibit immune suppression throughout their body and a distinct immune cell profile in their blood. These immune changes do not need to cross the blood-brain barrier to be identified. Following this, the team conducted a computer simulation to assess the potential of their liquid biopsy technique for screening the estimated 10 million individuals who visit emergency rooms or primary care clinics annually due to headaches. Typically, these patients are only referred for brain imaging if their physician suspects an underlying issue. However, in the simulation, patients whose liquid biopsy results suggested brain cancer were sent for imaging, which led to the detection of nearly 1,700 additional cancer cases compared to the conventional method. The next step for the team is to design a prospective trial to validate these findings in larger populations at higher risk for brain cancer.

“Our next-generation AI liquid biopsy approach combining DNA fragments and repeating genome patterns may accelerate brain cancer diagnosis,” said Victor E. Velculescu, M.D., Ph.D., co-director of cancer genetics and epigenetics at the cancer center. “It may ultimately allow patients to get care earlier, improving their treatment outcomes.”