Johns Hopkins researchers develop liquid biopsy method that measures epigenetic instability rather than absolute methylation levels, showing high accuracy in lung and breast cancer detection.
Researchers at Johns Hopkins Kimmel Cancer Center have developed a liquid biopsy approach that identifies early-stage cancers by measuring random variation in DNA methylation patterns rather than absolute methylation levels used in current tests.
The method utilizes a new metric called the Epigenetic Instability Index (EII) and successfully distinguished patients with early-stage lung and breast cancers from healthy individuals with high accuracy, according to findings published January 27 in Clinical Cancer Research.
“This is the first study where we are trying to really implement measuring that variation, or stochasticity, into a diagnostic tool,” says Hariharan Easwaran, PhD, associate professor of oncology at Johns Hopkins University School of Medicine and lead study author, in a release. “We immediately found that measuring DNA methylation variation performs better than just measuring DNA methylation by itself.”
Addressing Limitations of Current Liquid Biopsies
Current liquid biopsy blood tests that measure DNA methylation typically detect specific, absolute changes in methylation—a chemical reaction in which a methyl group is added to DNA at individual sites in the genome. However, these tests are usually developed through studying specific patient cohorts and tend to work for those populations but fail to perform as well in broader, more diverse groups.
To develop a better diagnostic tool for cancer screening, the research team analyzed publicly available cancer DNA methylation datasets from 2,084 samples to identify a panel of 269 specific genomic regions, known as CpG islands, which captured most DNA methylation variability across multiple cancer types.
“We identified specific genomic regions that tend to be the most variable in DNA methylation marks during cancer,” says Sara-Jayne Thursby, postdoctoral researcher in Easwaran’s lab and first author on the paper, in a release. “In cell-free DNA in the blood, that variability shouldn’t be high, but if it is, it is indicative of a developing cancerous phenotype.”
High Performance Across Cancer Types
The team trained a machine learning model to distinguish cancer signals from healthy signals, then tested the model using cross-validation approaches. In lung adenocarcinoma, the EII differentiated stage 1A cancer with 81% sensitivity at 95% specificity. For early-stage breast cancer, the tool detected cancer with approximately 68% sensitivity at 95% specificity.
The method also showed promise in detecting signals from colon, brain, pancreatic, and prostate cancers.
“Our hypothesis is that during the earliest stages of cancer development, methylation starts shifting,” says Easwaran in a release. “We can try to pick those signals using these stochasticity metrics, even of early cancer stages, as long as the DNA is shed in the blood.”
Thomas Pisanic, PhD, associate research professor of oncology at Johns Hopkins Institute for NanoBioTechnology and co-lead on the study, adds that early-stage tumors and precancerous lesions exhibiting high degrees of methylation variation may be more resistant to cancer-protective mechanisms and progress more rapidly.
Potential for Clinical Integration
While further validation in larger, long-term clinical cohorts is required, the EII could complement existing screening tools developed at Johns Hopkins, such as DELFI and other DNA mutation-based assays, and be used as a potential secondary triaging measure for clinical use.
For example, if a patient has a high PSA (prostate-specific antigen) test, which often yields false positives, an EII blood test could help determine if a follow-up biopsy is truly necessary, according to Easwaran.
The research was supported by the National Cancer Institute, National Institute on Aging, and National Institute of Environmental Health Sciences, along with several private foundations.
The team is now expanding and improving upon the method to continue developing the EII into a diagnostic tool. Additional authors include researchers from Johns Hopkins specializing in oncology, biostatistics, and pathology.
Photo caption: Hariharan Easwaran, PhD, MSc, Thomas Pisanic, Ph.D., and Sara-Jayne Thursby
Photo credit: Johns Hopkins Medicine
We Recommend for You:
