Detecting and treating cancer before it invades or metastasizes is the best way to ensure the patient survives the disease. Yet, cancers can be asymptomatic for so long that early detection can be difficult. Many innovators are betting a blood test, or liquid biopsy, could be the tool we’ve been waiting for.
Although the term "liquid biopsy" often refers specifically to analysis of circulating tumor DNA (ctDNA) or whole circulating tumor cells in blood, it can more broadly denote blood and other fluid sample-based tests measuring various analytes, such as proteins or antibodies, for various purposes. Today, liquid biopsy is commonly used in patients with known cancers to facilitate treatment choice, such as a companion diagnostic, and to monitor cancer patients. The use of liquid biopsy to detect early, pre-symptomatic cancers remains a much less explored frontier.
Some companies already offer supplemental screening tests aimed at specific cancers and higher-risk populations, but others are working to develop a "pan-cancer" blood test, one that could theoretically be taken by people with no particular reason to think they have cancer (See Figure 1).
A group at Johns Hopkins University in Baltimore, Maryland made headlines in February this year when it published results of an eight-cancer screening test. A couple of months later, on April 17, liquid biopsy marquee player Grail Inc.reported results of several proof-of-concept approaches to cancer screening that achieved high specificity. And a handful of smaller-scale cancer-specific tests are already available.
“If with a simple annual blood test someone who feels completely healthy could identify that they might have an early-stage malignant tumor, that would obviously be revolutionary,” Thomas Rodgers, senior vice president and managing director of McKesson Ventures, a Grail investor, told Medtech Insight.
“If with a simple annual blood test someone who feels completely healthy could identify that they might have an early-stage malignant tumor, that would obviously be revolutionary,” said Thomas Rodgers, senior vice president and managing director of McKesson Ventures, a Grail investor.
Not only would such a test allow for more targeted treatment and considerably improve cancer prognoses, it would also “dramatically change the landscape of where and how cancer's treated," he said, adding "I think a lot more care would move into the community and lower the cost of care significantly.”
Grail will focus first on a multi-cancer test rather than a pan-cancer test, Rodgers said.
"We're optimistic and confident -- I think you could see as early as next year, [Grail] would have their own test in the market in the US," he said.
Others are more measured in their enthusiasm for early detection.
Howard Urnovitz, chief science and strategy officer of Chronix Biomedical Inc., told Medtech Insight,“I still think we're looking at a minimum of two years, probably four to five years, before there's really a robust, general screening test for cancer.”
Given the stakes, they’ll likely need to be FDA-approved, not laboratory-developed tests (LDTs), he added.
Some observers worry about repeating mistakes made with other early-detection tests whose accuracy was initially oversold.
One example is the overdiagnosis of prostate cancer with prostate-specific antigen (PSA), noted J. Leonard Lichtenfeld, deputy chief medical officer for the American Cancer Society.
“The current thinking is, we develop cancers not infrequently and our bodies actually deal with those cancerous cells. So if you start detecting a disease in people who don't have any disease, you really still have to answer the question, ‘Does this one make a difference?’,” Lichtenfeld toldMedtech Insight. “The further we increase our ability to find cancer earlier, the more likely we are to find these cancers that may not cause a problem.”
“The current thinking is, we develop cancers not infrequently and our bodies actually deal with those cancerous cells. So if you start detecting a disease in people who don't have any disease, you really still have to answer the question, ‘Does this one make a difference?’,” said Leonard Lichtenfeld, deputy chief medical officer for the American Cancer Society.
Clive Morris, chief medical officer of Cambridge, UK-based Inivata Ltd., also believes that a cancer screen will take a while to develop. Inivata, a liquid biopsy company, analyzes circulating tumor DNA (ctDNA) to stratify and monitor cancer patients (Also see "Inivata Names New Chief Operations Officer" - Medtech Insight, 12 Sep, 2017.)
Once a cancer has progressed, the amount of ctDNA spilling into the blood is higher than during the earliest stages of cancer, making it easier to detect and use for stratification and monitoring. By contrast, a small early cancer may shed very little ctDNA, presenting a more daunting detection challenge. (See Figure 2).
FIGURE 2: AT CANCER'S EARLIEST STAGE, DETECTING ITS SIGNAL IN THE BLOOD IS EXCEPTIONALLY DIFFICULT. LIQUID-BIOPSY TECHNOLOGIES ARE MORE WIDELY AVAILABLE FOR TREATMENT STRATIFICATION, MONITORING AND PERSONALIZED TREATMENT PLANS.
“We are not going to try and compete in the screening-for-cancers environment," Morris said. "Technically it’s very difficult, and I think will take a long time. I think [someone] absolutely will get there ... I think it’s a five-to-10-year-project, rather than a two-to-three-year project.”
Estimates of the market opportunity for early-cancer detection liquid biopsy vary widely.
Jay Flatley, the former CEO of Grail’s parent company Illumina Inc., valued the market at $100bn to $200bn, Bloomberg reported on April 6, 2016. The article didn't note whether he specified a particular market. A more cautious 2015 Piper Jaffray report put the future domestic market potential at just $15bn (Also see "VC Deals Analysis: Liquid Biopsy Summons Yet Another Top A-Lister" - Medtech Insight, 8 Sep, 2017.)
Justin Li, GM, controller, and VP of finance at Irvine, California-based Laboratory for Advanced Medicine Inc.. (LAM), which developed a blood test to confirm cancer and monitor recurrence, cited company estimates that put the number of yearly cancer-screening procedures in the US at 80 million. He sees an even bigger opportunity in China. Grail’s 2017 acquisition of Hong-Kong-based startup Cirina Inc.underscores the tremendous size of the Asian market.
According to J.P. Morgan, Li said, “the global NGS (next-generation sequencing) cancer screening, monitoring, prognostic, and theragnostic market opportunity will exceed $76bn in 2020.”
On the reimbursement side, Rodgers expects pathways to resemble those of prenatal screening, and said people will pay out of pocket at first.
“I don't think you will be as focused on CMS (Centers for Medicare and Medicaid Services), given age demographics,” he said. “Assuming the data is strong, I am confident that most commercial payers will eventually pay a few hundred dollars for it."
He added, “In theory, you could have everyone between the ages of 40 and 60 benefitting from having this test performed every couple of years, so that adds up to a pretty big market -- more akin to a major drug.”
“In theory, you could have everyone between the ages of 40 and 60 benefitting from having this test performed every couple of years, so that adds up to a pretty big market -- more akin to a major drug," said Rodgers.
Several companies already have a jump on more modest diagnostic blood tests.
In March, German-based Epigenomics AGgot a vote of confidence from US federal lawmakers for its FDA-approved Epi proColon DNA-methylation blood test to screen average-risk people for colon cancer when lawmakers introduced bipartisan legislation to secure Medicare reimbursement for the test.
Epi proColon is also available in the EU and China. Epigenomics also received the CE mark for Epi proLung, a test that screens for elevated lung cancer risk. A liver cancer test for cirrhosis patients is undergoing a 440-patient trial.
Privately funded LAM, which works with telltale methylation patterns in ctDNA, developed IvyGeneLDT. The test aims to confirm the presence of breast, colon, liver and lung cancers, complement abnormal results on another test, or monitor for recurrence.
For patients with a confirmed cancer of any of those four types, the test’s sensitivity is 86% and its specificity is 88%. IvyGene hit the market in January; some 2,000 units have been sold at $400 each. Though it is not currently FDA-approved as a screening test, Li said LAM has invested heavily in gaining approval.
Additional single-tissue blood tests are in the pipeline, and LAM intends to submit PMA applications for them as class 3 medical devices.
“LAM's ultimate goal is to develop a series of additional tissue-specific screening tests that will help physicians to non-invasively screen for cancer,” said Nick Miner, LAM’s general manager of US operations.
Redwood Shores, California-based Freenome Inc.is somewhat unusual in its field for focusing on multiple analytes. It uses artificial intelligence to analyze cell-free DNA, RNA and proteins for cancer detection (Also see "Sophia Genetics' AI Brings In More Standardization To Liquid Biopsies" - Medtech Insight, 26 Jun, 2017.) The company has argued that ctDNA is not in itself adequate as a cancer-screening tool.
Freenome has raised $72m since its 2014 founding from investors including Section 32 (founded by former Google Ventures CEO Bill Maris) and Anne Wojcicki, co-founder and CEO of 23andMe Inc.. On April 24, 2018, Freenome announced a partnership with Swiss-based proteomics company Biognosys to develop a commercial screening test.
Meanwhile, Seattle-based Integrated Diagnostics Inc’s (INDI)Xpresys Lung 2(XL2) is a two-protein panel that when used together with several clinical factors allows physicians to label as benign some of the 1.6 million lung nodules found annually on CT (Also see "Start-Up Spotlight: Indi, Liquid Biopsy To Rule Out Lung Cancer" - Medtech Insight, 13 Oct, 2017.)
This February, INDI announced XL2’s clinical validation in the 685-patient PANOPTIC trial of stage 1A nodules, which are notoriously difficult to assess. XL2’s negative predictive value (NPV) was 98% for distinguishing benign from malignant nodules -- outperforming PET scans — and the authors calculated its widespread use would allow for 40% fewer biopsies on benign nodules, while misclassifying 3% of malignant nodules. The results appeared online inCheston February 28, 2018.
San Jose, California-based Chronix Biomedical Inc., meanwhile, has deferred plans to take its lung cancer screening test,Delta Dots—which was available for a short time in the UK—through a 510(k) FDA pathway (Also see "New Chronix Liquid Biopsy Test Could Save Doctors On Immunotherapy Costs" - Medtech Insight, 22 Mar, 2017.)
On March 6, the company was granted a US patent protecting a way to identify and monitor personalized tumor biomarkers that comprise repetitive element insertion regions in the tumor genome. Chronix Biomedical combined that approach with measurements of cancer cell genomes’ chromosome-number instability to create what Urnovitz said will be a more cost-efficient approach to cancer screening. He said the test is undergoing a clinical trial, adding he expects results late this year.
Nottingham, UK-based Oncimmune Ltd.is building on the longtime US success of its CLIA-certifiedEarlyCDT-Lungblood test, which detects seven circulating autoantibodies against tumor cells (Also see "Oncimmune Reels In New Strategy Head And More Capital" - Medtech Insight, 27 Mar, 2018.) The test is intended, firstly, to direct people at risk for lung cancer toward or away from lung CT.
“Where the test will be very helpful is in identifying the group that aren't of high enough risk to qualify for CT screening, but by dint of being positive on the test actually become much higher risk for the entry criteria,” Oncimmune's CEO Geoffrey Hamilton-Fairley told Medtech Insight.
Like XL2, EarlyCDT-Lung is also intended to assist with decision-making about indeterminate nodules on CT. Since 2012, the company has sold 150,000 units. In May 2017 the test became available in a validated, CE-marked kit, making it easier to market around the world.
On January 2 this year, the company announced an exclusive license with Hong Kong-based Genostics Co. Ltd., paving the way for EarlyCDT-Lung’s distribution, manufacturing and future development in China, where lung cancer kills 610,000 people annually—the country’s leading cause of cancer death (Also see "Oncimmune Kicks Off 2018 With China Pact For Lung Cancer Test" - Medtech Insight, 2 Jan, 2018.)
Genostics is also making a £10m equity investment in Oncimmune. Subject to approval in China, the two companies expect sales to commence within 36 months.
EarlyCDT-Lung is undergoing a lung-cancer detection trial, the 12,000-patient National Health Service (NHS) Scotland ECLS study. Two-year follow-up results are to be submitted for publication in Q4 2018, and Hamilton-Fairley hopes for a swift subsequent rollout through the UK’s National Health Service.
“There's been quite a resistance by patients just taking on the CT screening program, partly because they have to undertake to screen every year. [But] what the NHS trial has shown us -- and they've run psychological assessments alongside the trial — is that people seem to be very willing to have a blood test, and if that blood test then [multiplies] their risk score … they seem much more willing to enter a CT monitoring program,” Hamilton-Fairley said. “We think once we've got that data, we will be able to make a good case for being really complimentary and supporting CT screening.”
The company has its eye on other cancers.
Oncimmune announced last December it had validated EarlyCDT-Liver, a ten-autoantibody panel to screen for liver cancer. Commercialization is set to begin shortly; CE mark is expected later this year. An ovarian cancer test is also in Oncimmune’s pipeline.
Meanwhile, a multianalyte blood test, dubbed CancerSEEK, developed at Johns Hopkins, offers another glimpse of what's possible.
To detect and localize eight cancer types, CancerSEEK measures both ctDNA and protein biomarkers.
Principal investigator Nickolas Papadopoulos explained that ctDNA’s sensitivity is lower for early-stage cancers vs. later-stage cancers. Adding the protein markers allowed a positive result to be localized to its tissue of origin.
The team ran the test on 1,005 patients with known non-metastatic cancers, and also on a group of cancer-free controls.
CancerSEEK achieved sensitivities ranging from 33% with breast cancer to 98% for ovarian cancer; sensitivities for the other six cancer types fell between those extremes. The test correctly localized the cancer’s tissue of origin for 63% of patients with positive tests.
Specificity was 99%, with only seven of 812 patients receiving false-positive results. The paper appeared inScienceon February 23, 2018.
The team limited the number of DNA regions they checked to 16, in favor of practicality and to limit background noise from errors in sequencing.
“We were willing basically to take a hit on the sensitivity in order to preserve specificity,” Papadopoulos told Medtech Insight. “For a screening test [in] people that have no symptoms, this is a very important parameter.”
A privately funded $50m prospective five-year study of CancerSEEK is underway, testing 10,000 ostensibly cancer-free older women in the first phase and 40,000 more in the second.
“It's a pretty good first step to really kick the tires,” Papadopoulos said.
A first step it is. Averaged across all eight cancer types, the test’s sensitivity for stage 1 cancers was 43%, and that rate would likely be lower in a true screening population.
“I will be happy if we can detect even 50% of early-stage cancers. That's going to make a huge difference in cancer deaths,” Papadopoulos said.
Grail: Size Matters
When it comes to a blood test to detect multiple early cancers, size matters, many argue. On April 17, 2018, Grail announced results of three prototype genome-sequencing assays for lung, ovarian, pancreatic, liver, and esophageal cancer.
In a sub-analysis of the 15,000-person Circulating Cell-Free Genome Atlas Study, Grail performed liquid biopsy on blood samples from people with and without known cancer, including searches for lung, ovarian, pancreatic, liver, and esophageal malignancies.
Analyzing DNA using several approaches and also checking methylation patterns, the assays correctly classified all but five of 580 healthy controls as being cancer-free.
Two more people had been believed to be cancer-free, but were then found to have cancer, leaving just three false positives. At a specificity of 95%, sensitivity ranged from 80-95% for advanced cancers and from 50-65% for early-stage tumors.
Tom Rodgers of McKesson Ventures said his firm broke precedent in two ways with its Grail investment: not only did it write one of its bigger checks, it made that check out to a pre-commercial stage company with considerable clinical risks.
“If you look at who's invested in the company, you'd see that they've got some powerful, impressive and relevant allies,” Rodgers said. That includes companies that could help Grail bring down the costs of sequencing and cloud computing, just as McKesson’s investment in community-care giant US Oncology eases clinical trial logistics and accelerated clinical development for Grail, he explained.
Grail’s size gives Rodgers confidence, as the clinical studies needed to refine a pan-cancer blood test must necessarily involve huge numbers of participants.
“In this segment, that is the key success factor—you have to do the biggest study you can,” Rodgers said. “You need to power these studies to train an algorithm to read those signals, so size matters. And the normal approach of being scrappy and doing things on a shoestring … and all the stuff that the venture community normally loves, doesn't really apply here. You’ve got to do big science at big scale.”
“In this segment, that is the key success factor—you have to do the biggest study you can,” Rodgers said. “You need to power these studies to train an algorithm to read those signals, so size matters. And the normal approach of being scrappy and doing things on a shoestring … and all the stuff that the venture community normally loves, doesn't really apply here."
Skeptics warn it may take longer than we think to not only develop cancer-screening blood tests, but also to learn to apply them responsibly. With PSA and mammography, for instance, the enthusiasm that first greeted them decades ago has gradually given way to a more cautious view of their utility.
“We were told that [PSA] was going to help men,” Lichtenfeld recalled. “We started testing men, we sent men for radiation, we sent men for surgery, and then 20 years later, somebody comes back and says, ‘Well, not so quick, folks—we need to be a little bit more conservative about what we believe about this test,’ because nobody did the right studies up front.
“Someone like myself is left asking the question, ‘How do we avoid making that mistake again, and how can we avoid it without having to wait 20 years to find out?’” he said.
Lichtenfeld is worried about localization as well as overdiagnosis. Tracking down the location of a cancer that shows up in the blood might not be straightforward, since certain mutations are shared by a variety of cancers, he noted.
“The technology is advancing, and, I do believe [it] will advance to the point where it will become an early detection tool,” Lichtenfeld said. “However, the science to confirm the benefit of that technology is going to lag behind.”
San Diego-based company Pathway Genomics Corp.learned in September 2015 that the FDA is also bearing this concern in mind.
The agency sent Pathway a letter that questioned the company’s LDTCancerIntercept Detect, a 10-cancer early-detection test for use in high-risk populations.
“It is unclear how the literature that you cited, addressing the presence of circulating tumor DNA (ctDNA) in already-diagnosed patients, is adequate to support the expansive claims of screening for early cancer detection using ctDNA in undiagnosed patients for up to 10 different cancers with the CancerIntercept Detect,” the agency wrote. The test is no longer offered.
But some believe confidence in early cancer-detection blood tests is a matter of doing the studies.
“To me, this will come down to how much other confirmatory work you need to do for the patients who test positive to characterize the risk of prognosis and the likely treatment path,” Rodgers said. “And the more those answers come with that initial test, the more successful it will be from both a clinical and a commercial standpoint.”
“As cards are getting turned over, I think they are showing the science works,” he added, “and, with a sufficiently powered algorithm, you can read the signal.”
Papadopoulos said early detection won’t catch everything, and that it will be important to find a balance between over- and under-diagnosis.
“Of course, it's do no harm,” he said. “But we need—in cancer especially—we need to have those wins.”