machine-learning-can-improve-detection-of-brain-cancer-from-blood

“`html

Harnessing Machine Learning for Enhanced Blood-Based Brain Cancer Detection
Mathios

Dimitrios Mathios, MD, an associate professor of neurosurgery at WashU Medicine, has developed a technique that employs machine learning to identify brain tumors by examining DNA signatures linked to cancer found in the blood. In a recent investigation, Mathios discovered that this machine-learning approach recognized cancer in 73% of instances from a group of patients in the U.S. and Korea. The findings were confirmed in a secondary group of samples from brain cancer individuals in Poland. Conversely, a previous study indicated that traditional liquid biopsies, which tested blood for common cancer biomarkers, identified under 10% of brain tumors.

Mathios’ research was published in Cancer Discovery on April 29.

Identifying brain tumors at an early stage, prior to their growth and spread, has the ability to enhance treatment results. However, the early identification of brain cancer has been difficult, partly because the blood-brain barrier—designed to block harmful substances in the bloodstream from reaching the brain—also restricts certain biomarkers that could indicate a brain tumor from entering the circulatory system, where they could be detected by a blood test.

Detecting brain cancer is challenging without costly scans, which may not be utilized until after the cancer has advanced to a stage where it causes observable symptoms. Nevertheless, even in the initial phases of tumor development, the illness may prompt alterations in the body’s immune system response. Tumor cells can also release DNA fragments into the bloodstream that possess distinct characteristics compared to the DNA released from normal tissues. By investigating patterns consistent with these changes in blood samples, the machine-learning tool crafted by Mathios and his collaborators successfully identified brain cancer without invasive or costly methods.

The researchers approximated that the broad use of this device for screening individuals exhibiting symptoms such as headaches could potentially facilitate earlier diagnosis of as many as 1,700 additional brain cancer cases annually in the U.S., effectively doubling the number of identified patients without significantly increasing imaging study utilization.

The post Machine learning can improve detection of brain cancer from blood appeared first on The Source.

“`


Leave a Reply

Your email address will not be published. Required fields are marked *

Share This