Software detects early-stage cancer DNA using epigenetic markers

UCLA scientists are developing a computer program that can spot early-stage cancer in blood samples.

It’s long been known that tumors cast tiny bits of DNA into the bloodstream, and many efforts are underway to translate that knowledge into new technologies for early detection. But one major challenge will be figuring out where in the body the cancer is hiding.

Now researchers at the University of California in Los Angeles believe they’ve hit upon a way to do just that with a simple blood test. Their technique uses computer software to analyze specific molecular patterns in cancer DNA that’s detected in blood. They developed a computer program that can not only detect tumor DNA, but also identify what type of cancer it is, according to a press release.

The scientists built their program on epigenetic patterns known as “methylation,” which are common in cancer. Methylation patterns differ from one tumor to the next, however, making it possible to build “molecular footprints” of cancers emanating from specific tissues, such as those in the lung or liver, says co-lead author Jasmine Zhou, a professor at UCLA.


How ICON, Lotus, and Bioforum are Improving Study Efficiency with a Modern EDC

CROs are often at the forefront of adopting new technologies to make clinical trials more efficient. Hear how ICON, Lotus Clinical Research, and Bioforum are speeding database builds and automating reporting tasks for data management.

“We also compiled the same 'molecular footprint' for non-cancerous samples so we had a baseline footprint to compare the cancer samples against,” Zhou said in the release. “These markers can be used to deconvolute the DNA found freely in the blood into tumor DNA and non-tumor DNA."

The UCLA team tested their software on blood samples from 29 patients with liver cancer, 12 with lung cancer and 5 with breast tumors. Their program was able to detect early-stage cancers with 80% accuracy, they reported, and the error rate was lower than it was with two other technologies. They published their results in the journal Genome Biology.

Several research groups are working on early-detection technologies based on free-flowing tumor DNA. Perhaps the best funded of those efforts is Grail, a Menlo Park, California, startup that raised $100 million last year from Bill Gates, Bezos Expeditions and other high-profile investors. Grail, a member of last year's FierceMedicalDevices' Fierce 15, was created by gene-sequencing giant Illumina. In January of this year, Grail said it expects to raise $1 billion in its Series B financing—an unheard-of sum in biotech.

Some high-profile scientists are putting out the call for even more funding to be devoted to the development of early-detection technologies. Earlier this week, Johns Hopkins published research suggesting that 66% of cancers are caused by random unavoidable DNA copying errors that occur during the normal process of cell division. The study expanded upon research they released in 2015 that came to a similar conclusion and caused a stir in the oncology community. The researchers believe their study proves that too little attention is given to the importance of early detection.

UCLA’s Zhou acknowledges that one shortcoming of her research is that it focused only on liver, breast and lung tumors. But she was heartened by the ability of the computer program to catch early-stage cancer. “The technology is in its infancy and requires further validation,” she said in the release, “but the potential benefits to patients are huge."

Suggested Articles

At ASH, Dana-Farber researchers presented promising preclinical data on AMG 701, Amgen's long-lasting BiTE for multiple myeloma, but rivals loom.

The takeover, the first of Paul Hudson’s time as CEO, will give Sanofi an IL-2 drug it thinks can become a foundation of future cancer combinations.

A phase 1/2b test showed Roche's antibody mosunetuzumab can trigger durable responses in a significant minority of hard-to-treat NHL patients.