Delfi Diagnostics raises $100M for a new approach to screening blood for cancer

A new method of screening the blood for the telltale signs of tumors promises to offer more chances for a successful result, and Delfi Diagnostics has raised $100 million to help make the cancer test a reality. 

Delfi's technology employs machine learning to profile the entirety of genomes collected from the bloodstream—including DNA from healthy white blood cells, as well as strands shed by tumors—and compares the two to spot any changes.

Cancer DNA can carry a distinct fragmentation profile—or the natural breakages in the molecule and its packaging, which occur within cells during replication—that not only belie the presence of a tumor in the body, but can also trace it back to its original location, as different types of cells parcel their DNA in different ways.

“It's a mutation-agnostic approach for evaluating the millions of potential changes that are a result of fragmentation differences across the genome,” founder and CEO Victor Velculescu, M.D., Ph.D., said in an interview. 

“It’s very different from the targeted approaches that other folks are using, whether it's targeted sequencing or targeted methylation,” said Velculescu, who also serves as co-director of cancer biology and associate cancer center director for precision medicine at Johns Hopkins University.

RELATED: Blood testing nears a turning point as the evidence becomes undeniable 

While some cancer tests may seek out a positive hit among dozens of individual mutations, or gather information from thousands of changes in methylation groups, differences in fragmentation can number in the millions.

“Theirs have many fewer shots on goal, and in contrast, we believe we have many more, and that leads us to be able to have many more opportunities for detecting cancer,” he said.

The method also lends itself well to a machine learning approach, as the test is searching for different patterns, distributions and ratios among larger data sets, instead of small genetic alterations.

A previous paper published in Nature showed the method delivered sensitivities ranging from 57% to over 99% among seven different types of cancer, with a false-positive rate of 2% in samples gathered from 400 individuals.

“We felt this was the best place to look, because it's easier, and then consequently less expensive, to find these differences in packaging in very small amounts of DNA derived from the tumor that are circulating in the blood,” said Delfi’s chief scientific officer, Nic Dracopoli, Ph.D.

Delfi plans to develop its in vitro diagnostic platform into a range of products designed to spot specific cancers, as well as to identify multiple types of cancers from a single sample. Additionally, the simple, bioinformatics-led approach could enable the test to work in potentially any clinical laboratory equipped with a gene sequencer.

RELATED: Foundation Medicine acquires liquid biopsy developer Lexent Bio

The liquid biopsy space reached a turning point over the past year, with companies like Grail and Thrive Earlier Detection staking out significant clinical progress as well as acquisition offers of $8 billion and $2.15 billion from Illumina and Exact Sciences, respectively. Still, Delfi’s approach has been able to capture investors’ attention.

The Baltimore-based company’s series A was led by OrbiMed, with additional funding from new backers Foresite Capital, Northpond Ventures, Cowen Healthcare Investments, Rock Springs Capital and funds and accounts advised by T. Rowe Price Associates. 

Delfi’s previous investors Menlo Ventures, Samsara Biocapital, Illumina Ventures, AV8 Ventures and Windham Venture Partners also joined the round, after delivering $5.5 million in seed money in 2019. 

"If liquid biopsy is to reach its full potential in early cancer detection and save as many lives as we believe it can, tests must be both high performing and broadly accessible in order for large populations to benefit,” said OrbiMed partner Rishi Gupta. “We believe Delfi's test may be ideally suited for broad adoption based on performance, simplicity and cost."