BALTIMORE – Researchers from the New York Genome Center and Weill Cornell Medicine and their collaborators have demonstrated the utility of Ultima Genomics' sequencing platform for sensitive blood-based cancer monitoring, using whole-genome error-corrected sequencing of circulating cell-free DNA (ccfDNA) at high coverage.
In a preprint posted on BioRxiv last year, the researchers showed that the Ultima sequencing platform, a newcomer in the next-generation sequencing field, generates data that is sufficiently accurate to perform the sensitive mutation calling that is necessary for cancer liquid biopsy testing.
Additionally, with the Ultima platform's low sequencing cost, it is now possible to implement genome-wide duplex error-corrected sequencing for circulating tumor DNA (ctDNA) detection in a cost-effective way, they suggested, enabling the clinical adoption of de novo cancer detection and minimal residual disease (MRD) monitoring.
Currently, ctDNA detection usually involves deep targeted sequencing, often informed by a patient's tumor mutational profile. While this strategy can boost the sequencing coverage for specific targets, its outcome is also dictated by the abundance of ccfDNA in the sample.
"Although, in principle, you could sequence patient-specific [targets] or cancer hotspots to, let's say, a million X," said Dan Landau, a researcher at Weill Cornell Medicine and the New York Genome Center and the senior author of the study. "In practice, after about a few thousand fragments, you're exhausting the available material, which puts a physical boundary to the approach of deep-targeted sequencing."
To circumvent this problem, several years ago, Landau's team sought to perform whole-genome sequencing (WGS) on plasma samples.
"Back in 2016, we had this eureka moment: We said, 'What if we do whole-genome sequencing and aggregate signal across the tens of thousands of mutations that are found [in the plasma] genome-wide?'" he said. "That allows us to use the cumulative signal from all of these mutations together in a way that is no longer dependent on the number of genomic equivalents."
In 2020, Landau's team published a proof-of-concept study, demonstrating the value of WGS in liquid biopsy for ultra-sensitive cancer monitoring. Landau is also the scientific cofounder of C2i Genomics, a joint venture and spinout from NYGC and Weill Cornell Medicine that has been commercializing a WGS-based liquid biopsy test for solid tumor MRD detection.
Despite the promise of the whole-genome approach, until recently, sequencing costs still imposed a hurdle for the wide adoption of the method, Landau said. "That put a limitation on the clinical applicability of this work."
To overcome this barrier, more than a year ago, Landau's team started to collaborate with Ultima, which emerged last year with a promise to drastically reduce sequencing costs and deliver a $100 genome.
While the authors did not disclose the exact cost of Ultima sequencing in their paper, Landau said in this study, the cost per gigabase for the Ultima platform is at least seven times less than that of Illumina sequencing, which, according to the paper, is conducted on a HiSeq X or NovaSeq platform using 2×150 paired-end chemistry.
For their recent study, the researchers used the Ultima platform to perform deep sequencing at about 120X coverage on plasma samples from healthy controls, cancer patients, and patient-derived xenograft mouse models. The deep coverage boosted the sensitivity of the assay, enabling ctDNA detection in the parts-per-million range, they wrote.
The researchers also performed a head-to-head comparison between matching libraries of the Ultima and Illumina platforms, which showed comparable copy number and single nucleotide variant profiles between the two technologies. "It’s a new machine, so we needed to establish that the error profiles are compatible with the very sensitive mutation calling we need in the field of liquid biopsy," Landau said.
Encouraged by the results, the authors further utilized the cost advantage of Ultima sequencing by implementing duplex error-corrected sequencing at a genome-wide scale. This resulted in approximately 1,500-fold error reduction in the plasma of patient-derived xenograft mouse models and an error rate of one in 10 million in patient plasma samples.
The researchers then deployed the assay to help detect resectable melanoma in blood without matching tumor sequencing data, demonstrating the assay's ability for de novo cancer monitoring.
"In the liquid biopsy space, this is really meaningful, because for the first time, we can actually get good mutational profiles from just plasma directly at the level of the entire genome," Landau noted, adding that this will allow for the de novo detection of cancer, both as part of monitoring and, in the future, for early detection.
The authors have filed patent applications regarding certain aspects of the technology. "We will be discussing this with commercial partners to try and team up so we can leverage this for commercial development," Landau said. However, he did not say whether C2i Genomics will be exclusively licensing the IP.
"As a clinician, this is very exciting," said Lillian Siu, a medical oncologist at the Princess Margaret Cancer Centre in Canada. The method holds a lot of promise, she added, "especially because it is likely going to be a tumor-naive assay."
Because most existing liquid biopsy tests for MRD detection are tumor-informed, Siu said, clinicians need to have prior knowledge of a patient's tumor sequence in order to achieve targeted sequencing. These include, for example, Natera's Signatera assay, Personalis' NeXT Personal assay, and NeoGenomics subsidiary Inivata's Radar assay.
"If you don't know what you are looking for, then it becomes very difficult," she noted. As such, a tumor-independent approach would be "very attractive" for tumors that do not have archival samples available or are not easily resectable.
Additionally, Siu said, the assay's ability to not only detect ctDNA signals more broadly but also sequence them at greater depth can increase the chance of successful MRD detection, especially when the disease burden is low.
With regard to the test's wider clinical adoption, Siu said a key factor to consider is its actionability. "I think the big question really is: 'Can you do anything about it?'" Siu noted. "Even if we detect [ctDNA], we cannot really do anything because it's too late, or if there are no appropriate interception strategies, there is no point."
"Hopefully, that's not the case," she added, and patients can be cured at an earlier point in their disease development.
Trevor Pugh, a cancer genomics researcher at the Princess Margaret Cancer Centre and the director of the genomics program at the Ontario Institute for Cancer Research (OICR), said he is excited about the study for two reasons.
For one, he said, with the very low detection limit it achieved, the assay presents more opportunity for early cancer detection, enabling scientists to identify signals that were previously hard to detect with targeted sequencing.
"Even if you had the best possible targeted sequencing approach in the world, if that piece of mutant DNA isn't in your sample, you're never going to find it," he explained. "To borrow the Canadian hockey analogy, this is more like putting as many pucks on the net as you possibly can, and hoping that some of them go in."
Moreover, Pugh said, the study offered one of the first looks at the use of the Ultima sequencer in the context of ctDNA detection, offering insights into the new technology's performance when it comes to liquid biopsy.
"Certainly, Illumina sequencing has been the mainstay, and they have also made some strides in lowering the genomic sequencing cost," he said. "This is also where new entrants, like Ultima, are so interesting."
Pugh noted that his academic core lab at OICR, which is currently collaborating with both C2i Genomics and Ultima on cell-free DNA sequencing projects, is working on setting up a protocol similar to the one described in the NYGC/Weill-Cornell study.
"It's certainly a method we're very interested in," he said. "We have spoken with many players in this space who are looking to set up these protocols, and we're making this available through our genomics program as a service."
Still, Pugh said "the big rock to tackle" moving forward is to apply the WGS-based method toward cases beyond cancers with high mutation rates. "This proof of principle is in high mutation rate cancers because you have lots of shots on goal," he said. "There's a lot of innovation and work to be done so this can truly work pan-cancer, rather than just the cancers with high mutation rates."
"I think this is a clever approach to capitalizing on advancements in the development of new sequencing technologies," said Joshua Cohen, an M.D./Ph.D. student at Johns Hopkins School of Medicine.
Working in the lab of Bert Vogelstein, Cohen has also been developing high-sensitivity liquid biopsy assays for MRD detection. He also cofounded Haystack Oncology, a Johns Hopkins spinout that is commercializing a tumor-informed, patient-specific liquid biopsy test that is based on technology devised by the Vogelstein group.
Cohen said one drawback of tumor-informed assays is that they require access to tumor tissue, which can be challenging for certain patients. Meanwhile, with the WGS-based approach, the requirement for sequencing coverage and accuracy can be demanding, he pointed out.
"The challenge is that you are sequencing a much larger swath of real estate. There are many, many more positions that you have to interrogate," Cohen said. "You have to tease out more accurately all the positions that you're looking at — which ones are the ones that are likely to be driving a positive or negative call."
In addition to benchmarking Ultima sequencing's performance for liquid biopsy, Cohen said, another major achievement of the study is that the authors leveraged machine learning-based algorithms to correct the Ultima sequencing data, making cancer signals easily distinguishable from the noise.
Still, given that Ultima is a new player in the field, Cohen said it remains to be seen how it will fare against competitors, especially Illumina. "I think it's a little premature to necessarily say which sequencing technology is going to come out ahead," he said. With Illumina responding to the competition by releasing its NovaSeq X instrument, which promises to significantly reduce sequencing costs, Cohen said he still thinks the company is "very much in the game."
Moreover, Cohen said, there is still "a lot of work that needs to be done" in order to translate the technology described in this study into the commercial space, especially given that it was developed on a new sequencing platform.
Even so, he said he believes the increasing competition in the NGS market is hugely advantageous to the liquid biopsy field in general.
"It's beneficial across the board for the entire field for sequencing to be more affordable," he said. "That just opens the doors for all of these new technologies that are based on next-generation sequencing to be more affordable, which is really key to bringing all of these things to patients."
Echoing Cohen's opinion, Landau said he agreed that competition is a good thing for the community. "As part of the genomics community, we certainly welcome Illumina coming up with new products that are more affordable," he said. "We think that the fact that there is, for the first time, competition in this field is obviously enormously beneficial to the community at large."
"This study makes the point that sequencing cost should no longer be a barrier to the clinical application of plasma deep sequencing for minimal residual disease applications," he added. "I think that these types of data may show that, long term, there is at least a very strong complementary role to whole-genome sequencing of the plasma, considering the current dominance of targeted panels in this field."