Skip to main content
Premium Trial:

Request an Annual Quote

Exact Sciences Marker Panel Detects Early Liver Cancer With Triple Sensitivity of Current Biomarker


NEW YORK – Exact Sciences is most well known for its colorectal cancer diagnostic Cologuard. But during several of the company's recent quarterly earnings conference calls, CEO Kevin Conroy has also spoken about its plans to release a liver cancer diagnostic test, which he has said has "superior performance … compared to the alpha-fetoprotein [AFP] test," which is the current standard testing modality for hepatocellular carcinoma (HCC).

Now, the company has released data showing that its panel of biomarkers is nearly three times as sensitive as AFP testing for the detection of early-stage HCC. The blood-based test, developed in collaboration with the Mayo Clinic, is composed of four methylated DNA markers and two protein markers.  

According to John Kisiel — an associate professor of gastroenterology and hepatology at the Mayo Clinic and co-author of an abstract Exact presented last week at the annual meeting for the American Association for the Study of Liver Diseases — AFP is not considered powerful enough to be used as a standalone test. It's used in combination with ultrasound — most commonly, patients at higher risk of developing HCC because of cirrhosis or longstanding hepatitis B infection are monitored with AFP and ultrasound every six months. However, even with the addition of ultrasound, this type of testing has about 63 percent sensitivity for detecting early-stage cancers, Exact also noted.

"We think that the data are a very compelling indication that this new panel of markers is substantially better than AFP," Kisiel said. "The difficulty with ultrasound is that while it's the current standard of care, it is also often insensitive for early-stage cancers. And although it's relatively widely available, it depends on a human operator to take the pictures, and it depends on patients coming in to have it done."

He further noted that ultrasound doesn't always work, particularly in patients who may have a larger central abdomen. "You may not be able to see their liver very well with an ultrasound," he said.

With Exact's new panel, which the company is planning to launch in the second half of 2020, those problems are ameliorated.

Researchers from Exact and the Mayo Clinic collected blood samples from 137 radiographically diagnosed HCC and 313 age-matched controls with benign liver disease without structurally apparent HCC. They assessed 10 methylated markers that had previously been reported as having some association with HCC, as well as several protein markers including AFP. With a specificity of 90 percent, the researchers found that their panel of six markers detected early-stage HCC with a sensitivity of 71 percent compared to 25 percent for AFP. The panel also detected Stage A HCC with a sensitivity of 74 percent compared to 30 percent for AFP. Overall, the panel's sensitivity was 82.6 percent compared to 47.1 percent for AFP. At 90 percent specificity, AFP's sensitivity was 45 percent.

The researchers chose to include AFP and an AFP isoform called AFP-L3 as the two protein markers in the panel, Kisiel said. AFP-L3 is represented as a ratio of a lectin protein to total AFP — the AFP-L3% assay is used to help identify people at risk of developing HCC, and studies have shown that AFP-L3% test results higher than 10 percent can be indicative of early HCC.

They further found that the addition of methylation markers gave the test more diagnostic power than the addition of DNA mutation markers would have. Although DNA mutation is valuable for understanding what types of treatments patients may benefit from, mutations in tumors are very heterogeneous from patient to patient and even among clusters of different cells within the same patient.

"The tumor may have different mutational profiles so that the mutation spectrum of a cancer can evolve within a single individual. And as they're given treatment, that treatment may select out certain cells and allow other ones with a different mutational profile to grow," Kisiel said.

On the other hand, tumors of one type tend to have similar methylation profiles and vary much less than DNA mutations from patient to patient. Data generated by Kisiel and his colleagues in colon cancer showed that the DNA methylation levels in primary tumors, sites of metastasis, and sites of distant recurrence were very similar to each other.

"That gives me some optimism that DNA methylation could be a very reliable way to track tumor growth or recurrence or shrinkage moving forward," he said.

Before the panel can be used in the clinic, though, there are some factors that have yet to be determined. A full clinical workflow must be vetted to ensure efficient testing standards. Most likely, Kisiel said, the test would be distributed in a kit with the appropriate blood tubes. A patient could have the blood drawn at any local laboratory or doctor's office, but the tubes would most likely be sent back to a central lab for testing, which would then report results to the ordering physician.

The protein-based biomarkers run on several standard commercially available platforms, he added. The methylated DNA markers are run using a proprietary, methylation-specific PCR technology that's very similar to what is used for Cologuard.

"Because it's targeted and just looking at a few small candidates, it is much faster and much cheaper than sequencing," Kisiel noted.

Optimal timing of initial and follow-up testing also have yet to be determined, but it would most likely be administered every six months like AFP. The frequency with which AFP and ultrasound are currently used is based on the speed at which tumors grow, Kisiel said, so that would be unlikely to change regardless of the testing modality being used.

In an email interview, Conroy noted that the company plans to encourage use of the test for early detection of HCC in patients at risk for developing the disease.

"Detection in the early stages can help improve survival rates, but current HCC testing has a low participation rate — less than one third of people adhere to routine testing," he said. "Clinicians and patients need more accurate, convenient testing options to help combat rising incidence."

He also said that Exact is currently evaluating its regulatory strategy for this test and has plans to continue meeting with the US Food and Drug Administration.

"Exact Sciences was recently granted Breakthrough Device designation by the FDA for the HCC test, which expedites development, assessment, and review processes to provide patients and healthcare providers with timely access to new technologies," Conroy added.

Research on the biomarker panel is ongoing, though Conroy declined to speculate on the possible completion of follow-up studies. Although there has been some speculation among analysts that the test might be named Liverguard (in the same vein as Cologuard), Conroy said it hasn't yet been officially named.

According to Kisiel, this type of biomarker panel could also be eventually used to track the progress of liver cancer or to gauge a patient's response to treatment.

"We have not studied that concept with liver cancer yet. We likely will," he said. "We do have some emerging data that the same type of technology can be used to monitor patients for recurrence, for instance, of colon cancer. That's data we haven't presented yet but hope to soon. So, I would say that tests that would be capable of detecting cancer at an early stage could also be applied to assess how patients were responding to therapy and to follow them over time."

He also noted that the implications for the test are broad. Although testing for the panel was not conducted outside of the context of surveillance of patients with liver cirrhosis or hepatitis B, the cancer testing field is moving towards tests that could screen for multiple types of cancers in healthy patients, he noted.

Though each cancer will probably have its own methylation profile, necessitating different methylation markers in a panel for each different cancer type, there are likely some methylation markers that would apply to many different cancers, making such a pan-cancer methylation panel a possibility.

"We might be able to develop a screening test for multiple different cancer types in the same population. But then, once we found a cancer, we would probably need something that was more specific to that tumor to follow it forward in time," Kisiel said. "The idea there would be like your cholesterol test — you would get a blood draw perhaps annually that would screen for multiple different types of cancers. While we don't currently have such a test available, the idea would be that this might ultimately replace or complement other cancer-based screening that people are already getting."