NEW YORK (GenomeWeb) – Adding its name to a growing list, Chinese firm SeekIn is eyeing a new funding round to support clinical validation and commercialization of a DNA-based blood test for early cancer detection.
Similar to US firms Grail and Freenome, SeekIn is using machine learning to train classifiers that can pick up signals in circulating cell-free DNA (and potentially from other molecular information in the blood) that represent the faint signs of an unknown tumor in the body.
Mao Mao, most recently the chief scientific officer at sequencing giant BGI, and now founder of SeekIn, said this week that the risk-reward proposition of early cancer detection seemed best tackled in a standalone startup. Although he did run a group looking at this area at BGI, the approach SeekIn is developing doesn't borrow any specific technology, he added.
"I [came to] see this is one of our greatest unmet medical needs," Mao said, "and the most impactful thing we can do to reduce cancer mortality."
The company has been running on seed funding so far but is anticipating announcing its Series A in the next few weeks, Mao said.
The firm started out using artificial intelligence methods to comb public genomic datasets and train algorithms that can identify cancer cases, based on patterns of DNA alteration in the blood. After demonstrating sensitivity in this public data, Mao said SeekIn is now working on replicating the results in its own cohorts. Researchers are also collecting some evidence from testing the assay in small groups of individuals with elevated cancer risk, which should speak to the method's performance in an intended-use population, as well as its clinical utility.
Mao and colleagues have not yet published data on the mechanics or accuracy of their approach. That said, relatively little information has been made available so far from any of the firms that are pursuing methods for screening healthy people for cancer, though Grail has ramped up its presentation of data at scientific meetings, as has Freenome, other firms like Singlera, and various academic groups.
According to Mao, SeekIn has compared what it is seeing using its own algorithms with data published by Grail, for example, and is confident that it is seeing performance that is at least comparable.
He said that the company will mostly likely publish some of its early results, as well as a description of its methodology and algorithms, in about a year. The firm has applied for a patent in the US for one aspect of its approach, a noise cancellation strategy to help avoid missing minute signals of cancer DNA amidst the large background of normal DNA in a blood sample.
Debate among (and about) the US firms that have stated intentions in the early detection space has focused on questions of sensitivity in early-stage cancers. Just because a technique can differentiate those with advanced cancers from healthy controls does not mean it will be accurate in identifying people with occult tumors in a general population.
But Mao questioned whether showing high sensitivity in case control cohorts of early-stage cancer patients is really the best way to evaluate these methods, considering that the medical field continues to grapple with the question of whether very early-stage tumors even require treatment.
Catching stage-I, or in situ tumors, or even pre-cancers, in other words, may not be what actually leads to big improvement in outcomes, Mao said. Companies working with ctDNA methods need to show that they can pick up aggressive cancers in people who would otherwise not know they carried them, and do it early enough that they can be treated.
With the money it hopes to raise in its planned financing round, Mao and colleagues intend to begin a clinical study next year in which they will apply their approach to a larger group of asymptomatic individuals and track them over time — to see if the assay picks up the presence of cancer before it would otherwise have been diagnosed.
This will support a submission to China's Food and Drug Administration, for approval of the method as an in vitro diagnostic, he added.
According to Mao, the approach has been sensitive, so far, across the cancers that are common in the Chinese population, and the goal is to create a test that is tumor-type agnostic. With a mind toward commercialization, Mao said that he and his team have set the bar for the cost of the test at $150 or lower, with the intention that it could be used for broad population screening.
That said, he recognized, as have other companies in the space, that both regulatory approval and clinical adoption may be easier in the context of specific cancer types — like colorectal cancer, or breast cancer — where there are existing screening paradigms in place and the clinical utility of early detection is pre-established.
SeekIn also has a patent submission for a specific liver cancer indication, though Mao did not detail any commercial plans that the firm has in that vein.