AI Technology-Based Blood Test Identifies Lung Cancer Earlier

Lung cancer stands as the most lethal cancer in the United States, as reported by the National Cancer Institute, and holds a similar status globally, according to the World Health Organization. Annual screening using computed tomography (CT) scans for individuals at high risk can detect lung cancers at an early, more manageable stage, potentially reducing mortality rates. The U.S. Preventive Services Task Force advises that 15 million Americans aged between 50 and 80 who have smoked should undergo screening, yet only about 6% to 10% of those eligible actually receive yearly screenings. The low screening uptake is often due to the time commitment required for arranging and attending screenings, and concerns over the minimal radiation exposure from the scans. Now, researchers have leveraged artificial intelligence (AI) to spot patterns of DNA fragments linked to lung cancer, which has led to the development and validation of a liquid biopsy that could identify the disease earlier. This innovation may enhance the identification of those at highest risk and who could benefit most from further CT screening, potentially increasing screening rates and reducing death rates.

In the last five years, researchers at Johns Hopkins Medicine (Baltimore, MD, USA) have developed a test employing AI to analyze DNA fragment patterns indicative of lung cancer. This method capitalizes on the different ways in which DNA is organized in healthy versus cancerous cells. In healthy cells, DNA is compactly and uniformly structured, similar to a rolled-up ball of yarn. In contrast, the DNA in cancer cells tends to be more disorganized. As these cells die, their DNA fragments, which end up in the bloodstream, appear more chaotic and irregular compared to those from non-cancerous individuals. Through a prospective study, which was published in the journal Cancer Discovery on June 3, the team demonstrated their AI-driven technology's ability to identify individuals who are more likely to have lung cancer based on these DNA fragment patterns in the blood.

The study included roughly 1,000 participants, both with and without cancer, who qualified for conventional lung cancer screening with low-dose CT. Participants were recruited across 47 centers in 23 U.S. states. The research team trained their AI software using specific DNA fragment patterns from the blood samples of 576 individuals, both cancer-afflicted and healthy. They then confirmed their methodology's efficacy on a second cohort of 382 individuals, with and without cancer. Their analysis indicated that the test possesses a negative predictive value of 99.8%, suggesting that only 2 out of every 1,000 tested may be missed and have lung cancer. Simulation studies by the group suggest that if the screening rate could be increased to 50% within five years through this test, it could quadruple the detection of lung cancers and increase the detection of early-stage cancers by approximately 10%. This could potentially prevent around 14,000 cancer deaths over the same period. The researchers plan to seek approval for the test from the U.S. Food and Drug Administration for lung cancer screening and explore its application for other cancer types.

“We have a simple blood test that could be done in a doctor’s office that would tell patients whether they have potential signs of lung cancer and should get a follow-up CT scan,” said Victor E. Velculescu, M.D., Ph.D., professor of oncology at the Johns Hopkins Kimmel Cancer Center. “The test is inexpensive and could be done at a very large scale. We believe it will make lung cancer screening more accessible and help many more people get screened. This will lead to more cancers being detected and treated early.” 

Related Links:
Johns Hopkins Medicine

Visit expo >

Gold Member

Blood Gas Analyzer

GEM Premier 7000 with iQM3

Gold Member

Fully Automated Cell Density/Viability Analyzer

BioProfile FAST CDV

New

Adenovirus Test

S3334E ADV Adenovirus Kit

New

Gold Member

ZIKA Virus Test

ZIKA ELISA IgG