New Approach Filters Biological Noise From Liquid Biopsies to Guide Cancer Treatment

A machine learning model developed at the Johns Hopkins Kimmel Cancer Center can identify whether mutations detected in liquid biopsy samples originate from a patient's tumor or from white blood cells—a distinction that matters when selecting targeted cancer therapies. The research was published in Clinical Cancer Research.

Liquid biopsies analyze cell-free DNA (cfDNA) fragments from tumors circulating in blood samples and are commonly used to identify mutations that can guide therapy selection. However, liquid biopsies can also pick up mutations that accumulate in white blood cells through an aging-related process called clonal hematopoiesis—a phenomenon that is common in older patients and in those who have previously undergone chemotherapy or radiation.

"When you do a liquid biopsy, and you get the report back, and you see mutations, you do not know if the mutations are coming from the tumor or the white blood cells," said first author Jenna Canzoniero. "If you want to select a mutation-targeted drug to treat the cancer, you want to make sure you are targeting mutations in the cancer and not mutations in the white blood cells."

To address this, Canzoniero and colleagues developed a model called plasmaCHORD. It analyzes characteristics of cfDNA fragments, which differ depending on whether they originate from tumor cells or white blood cells, a difference described as distinct "cfDNA fragmentation profiles." It also incorporates factors such as patient age and the type of gene and mutation involved.

The team trained the model on liquid biopsy samples from 225 patients with breast, colorectal, esophageal, ovarian, or non-small cell lung cancer, verifying its accuracy against matched genetic sequencing of tumor and white blood cells. Testing on a separate cohort of 114 patients from another institution using a different sequencing platform showed similar performance. Within that group, plasmaCHORD improved the ability to correctly distinguish tumor from white blood cell mutations from approximately 50% to 83% for a set of clinically relevant mutations.

The team also demonstrated that the model's predictions helped clinicians at the Johns Hopkins Molecular Tumor Board avoid selecting likely ineffective therapies.

"About one-third of mutations detected in tumor-naive liquid biopsies can originate from white blood cells, and our ability to match targeted therapies to each patient's genomic profile depends on our ability to distinguish tumor mutation from biological noise," said senior author Valsamo Anagnostou. "An artificial intelligence model applied to standard liquid biopsy tests could be both clinically valuable and quickly scalable."