2022 World Conference on Lung Cancer (ePosters)-EP01.01-010. Graphene Based Activity Sensors Detect All Stages of Lung Cancer Using an Evolutionary Machine Learning Algorithm Approach

EP01.01-010. Graphene Based Activity Sensors Detect All Stages of Lung Cancer Using an Evolutionary Machine Learning Algorithm Approach

Back to course

Pdf Summary

Researchers from Hawkeye Bio, Inc and the University of Kansas Medical Center have developed a graphene-based activity sensor that can detect all stages of lung cancer. The study, presented at WCLC 2022, used an evolutionary machine learning algorithm called "Emerge" to improve biomarker analysis and accurately detect early-stage lung cancer. <br /><br />The traditional regression analysis used linear relationships between biomarkers, but the Emerge approach allowed for both linear and non-linear interactions between multiple biomarkers. This resulted in higher accuracy and reproducibility in the analysis.<br /><br />The study examined 256 subjects, with 35% of them having lung cancer. The comparison between logistic regression and Emerge calculations showed that the Emerge approach had better performance, particularly in the validation dataset.<br /><br />The graphene-based activity sensor developed by Hawkeye Bio is sensitive to sub-picomolar concentrations of protease enzymes. It produces a detectable signal when active protease enzymes cleave a fluorescent dye attached to a graphene backbone. In the absence of active protease, the fluorescence is effectively quenched by the graphene backbone.<br /><br />The researchers evaluated 256 serum samples from lung cancer patients, patients with benign lung conditions, and healthy controls. They used a panel of 19 different peptide sensors to measure protease activity. The relationship between protease activity and disease status was analyzed using logistic regression and the Emerge algorithm.<br /><br />Overall, this study demonstrates that the graphene-based activity sensor combined with the Emerge algorithm has the potential to significantly improve the detection of early-stage lung cancer. Further validation in larger cohorts is ongoing to refine the diagnostic performance of the analysis algorithm.

Asset Subtitle

Paul Dempsey, United States

Meta Tag

Speaker Paul Dempsey, United States

Topic Early Detection and Screening - Biomarkers

Keywords

Hawkeye Bio

graphene-based activity sensor

lung cancer detection

Emerge algorithm

biomarker analysis

early-stage lung cancer

protease enzymes

fluorescent dye

serum samples

diagnostic performance