Using AI to track fragments of tumor-associated DNA in the blood, scientists say they may be close to an accurate test for a silent killer: Ovarian cancer.

It's the fifth leading cause of cancer death in the United States. Ovarian tumors are often lethal because they typically don't cause symptoms in their early, more treatable stages.

“Ovarian cancer is an incredibly deadly disease with no great biomarkers for screening and early intervention,” said senior study author Dr. Victor Velculescu, co-director of the Cancer Genetics and Epigenetics Program at the Johns Hopkins Kimmel Cancer Center in Baltimore.

His team presented its findings Tuesday at the American Association for Cancer Research (AACR) annual meeting, in San Diego.

A highly accurate "liquid biopsy" blood test for ovarian tumors has long been a Holy Grail of cancer research.

In the study, the Hopkins team focused on tiny fragments of the tumor's genetic material present in patients' blood.

“Because cancer cells are rapidly growing and dying and have chaotic genomes as compared to healthy cells, patients with cancer have different patterns of DNA fragments in their blood than patients without cancer,” explained study co-first author Jamie Medina, a postdoctoral fellow at Kimmel.

“By carefully analyzing these fragments across the entire human genome, we can detect subtle patterns indicating the presence of cancer," he said in an AACR news release.

The researchers left it to an artificial intelligence program to analyze the DNA blood "fragmentomes" of women with and without ovarian cancer. The AI program combined that data with measurements of blood levels of two ovarian cancer biomarkers, proteins called CA125 and HE4.

They hoped this mix could produce "a new high-performance approach for early detection of ovarian cancer," Velculescu explained.

The study involved 134 women with ovarian cancer, 204 women without cancer and 203 women with benign adnexal (ovarian) masses.

The results were impressive: The test had a specificity of 99%, meaning there were almost no women who received a false-positive result from their screening.

The test's sensitivity in picking up ovarian cancers varied by cancer stage. For example, it spotted stage 1 cancers 69% of the time; stage 2 cancers 76% of the time; stage 3 cancers 85% of the time and stage 4 cancers 100% of the time.

That level of accuracy significantly outperforms that of blood tests based on levels of the CA125 protein alone, the Hopkins group noted.

Velculescu stressed that larger studies are need to replicate and validate these early findings, but he remains optimistic.

“This study contributes to a large body of work from our group demonstrating the power of genome-wide cell-free DNA fragmentation and machine learning to detect cancers with high performance,” he said. “Our findings indicate that this combined approach resulted in improved performance for screening compared to existing biomarkers.”

Because the findings were presented at a medical meeting, they should be considered preliminary until published in a peer-reviewed journal.