NEW YORK – Johns Hopkins spinout Previse recently published a clinical validation study for its epigenetic Esopredict assay for predicting the risk for patients with Barrett’s esophagus to develop high-grade dysplasia (HGD) or esophageal adenocarcinoma (EAC).
The Halethorpe, Maryland-based gastrointestinal health company, formerly known as Capsulomics, said that its assay provides physicians with the means to more confidently tailor care plans for individual patients, and that the results of its study push epigenetic risk stratification testing one step closer to eventual inclusion in clinical care guidelines.
"Currently, there aren't any risk stratification tests included in clinical guidelines," said Previse Cofounder and CEO Dan Lunz.
Barrett’s esophagus is a premalignant condition in which chronic acid reflux damages parts of the esophagus, which can contribute to abnormal cell growth leading to cancer.
At present, diagnosis and surveillance relies heavily on pathology, using techniques such as H&E staining of biopsy slices to scan for the presence and growth of any cellular abnormalities.
However, Lunz explained, pathology alone has been unable to accurately assess an individual's risk for disease progression, with roughly half of those deemed at the lowest risk for progression by pathology going on to develop esophageal cancer. Even when caught early, such as at stage I, esophageal cancer carries an average five-year survival probability of approximately 47 percent.
On the flip side, Lunz said, ambiguities associated with pathology interpretation also sometimes result in overtreating. Current guidelines call for more frequent surveillance testing when low-grade dysplasia is identified, but some studies have found that a second pathology review may return a patient's status to non-dysplastic in up to 85 percent of studied cases.
"We need a better test," Lunz said.
First launched last year, Esopredict is a lab-developed, PCR-based assay that measures DNA methylation of the p16, HPP1, RUNX3, and FBN1 genes.
These methylation changes occur early on in abnormal cell growth and before conditions such as HGD and EAC can be detected clinically. Combined with a patient's age, these results help physicians estimate the likelihood that patients' premalignant conditions will advance and determine how often to monitor or manage patients with potential progression towards cancer.
Importantly, Lunz said that no additional patient samples are needed to run Esopredict. Rather, the test makes use of a subsample of a standard biopsy.
"There's no additional workflow for the physician or samples taken," Lunz said, "which can be an inconvenience or even a risk for patients."
DNA is extracted from the biopsy sample, followed by bisulfite conversion and methylation-specific PCR at Previse's CLIA laboratory in Baltimore. Results are run through the company's proprietary algorithm, which provides an "Esoscore" — a rating between zero and 100 and used to categorize patients from lowest to highest risk (risk rising with the value of the score) — and a five-year prevalence-adjusted probability of progression, which takes into account the Esoscore.
"It's a complementary piece to the diagnostic work that the pathologist and physician [are] doing, looking at different risk factors and risk variables," Lunz said.
In the retrospective case-control study, published recently in the American Journal of Gastroenterology in collaboration with Johns Hopkins University School of Medicine, where the test was first developed, investigators assessed biopsies from 240 patients with Barrett's esophagus from six medical centers.
While the overall risk for neoplastic progression was approximately 5 percent, risk scores generated via Esopredict indicated that patients predicted to be at a higher risk were, on average, 6.4 times more likely to progress than lower-risk patients and 15.2 times more likely to progress than the lowest-risk category of patients. Individuals in the highest-risk group had an approximately 22 percent risk of developing HGD or EAC, compared with a nearly 7 percent risk among those in the high moderate group, a 3.25 percent risk for those in the low moderate group, and just above a 1 percent risk among those in the lowest-risk group.
In the study, Esopredict showed a 99 percent negative predictive value among 62 patients in the lower-risk category, indicating a potential opportunity to decrease surveillance among some patients. Conversely, from 26 patients with non-dysplastic Barrett's esophagus who later progressed, the assay classified 14 of them as being at higher risk, indicating the potential for targeted increased surveillance and preventive measures such as endoscopic eradication therapy.
Rebecca Fitzgerald, a professor of cancer prevention at the University of Cambridge, called the study "interesting" in an email, while cautioning that "very limited conclusions can be drawn at this stage since they have just 17 cases in a validation cohort and then an additional set of 31 progressors."
The investigators from Previse and Johns Hopkins acknowledged this in their report, writing that this group was too small to allow statistically significant conclusions to be drawn.
Fitzgerald indicated, however, that epigenetics provides a promising avenue for developing assays such as Esopredict.
"We have seen from the multi-cancer early detection tests in blood that methylation markers are highly specific," she said.
Lunz agreed, saying that the company continues to generate further data through more studies aimed at demonstrating the performance of Esopredict and the impact it can have on patient care and on improving outcomes.
Previse is currently comparing biopsies from the same patient taken at different times, as well as from different locations in the esophagus taken at the same time, to evaluate whether and how risk changes with location or over time.
"We expect some exciting new data will be published in the coming months," Lunz said, "and even more by the first quarter of 2025."
Additional studies will also include more patients of other genealogical backgrounds, as the majority of those in the present study were of European ancestry.
Previse was originally spun out from Johns Hopkins as Capsulomics, and its technology is exclusively licensed from the university although Lunz said that it has since submitted patent applications for additional, non-specified intellectual property.
Insurance coverage for Esopredict is an evolving process, and Lunz said that while the assay has obtained coverage from "a range" of payors, including Medicare Part B and Medicare Advantage, he expects further agreements to come as clinical evidence builds. Without insurance, Esopredict costs $2,750 out of pocket, although the company offers financial assistance plans for households earning under certain financial thresholds.
"We're excited about the value that [Esopredict] is going to add to the doctor-patient relationship," he said.