Key points from article :
A new blood test using artificial intelligence may help doctors quickly determine whether pancreatic cancer treatments are working, according to a study published in Science Advances led by Carolyn Hruban, formerly of Johns Hopkins and now at the Dana-Farber Cancer Institute. The approach, called ARTEMIS-DELFI, analyses fragments of tumour-derived DNA found in the blood, allowing clinicians to assess treatment effectiveness within just four weeks—much faster than traditional imaging methods.
The technique was tested on patients from two major clinical trials and compared against an alternative method called WGMAF, which also examines tumour DNA but relies on biopsy samples. While both methods outperformed standard clinical tools, ARTEMIS-DELFI proved more practical and broadly applicable, as it doesn’t require a tissue sample and performed well even when biopsy quality was poor.
Senior author Dr. Victor Velculescu emphasized the urgency of rapid treatment decisions in pancreatic cancer, which is often diagnosed at an advanced stage. ARTEMIS-DELFI could enable a “fast-fail” approach, helping doctors pivot to alternative therapies sooner if the initial one proves ineffective.
The team, which includes researchers from Johns Hopkins and Denmark’s Copenhagen University Hospital, plans to validate ARTEMIS-DELFI in future prospective studies and explore its use in other cancers. Early results are promising: a related version of this technology has already shown success in monitoring colon cancer.
