Non-small cell lung cancers (NSCLCs), the leading cause of cancer death worldwide, can be cured if they’re caught before they spread to other organs and promptly treated. Trouble is the current gold standard of screening, the CT scan, generates many false positives. Further, most high-risk people currently do not receive the scans for a variety of reasons, including reluctance to get them done and access difficulties. To address these challenges, a team led by Ludwig Stanford researchers Maximilian Diehn and Ash Alizadeh developed a blood test that detects NSCLC by analyzing extremely rare bits of DNA shed into the blood by tumors. Max, Ash and their colleagues devised a method to differentiate genuine NSCLC-related mutations in such circulating tumor DNA (ctDNA) from mutations found in the far larger pool of circulating DNA that does not come from cancer cells. Integrating that capability with other molecular features of ctDNA, the researchers developed a machine learning-based approach named Lung-CLiP (for lung cancer likelihood in plasma) that they showed can detect between 40% and 70% of early stage NSCLCs. Although CT scans are more sensitive, Lung-CliP generates fewer false positives. Given its ease of use, Lung-CliP could be combined with CT scans to detect thousands of additional early-stage NSCLCs each year—providing, of course, it passes muster in large scale clinical trials. The method was detailed in a March paper in Nature.
This article appeared in the August 2020 issue of Ludwig Link. Click here to download a PDF (2 MB).