Skip to main content
Premium Trial:

Request an Annual Quote

Berg Exploring Commercialization Paths for Predictive Pancreatic Cancer Liquid Biopsy Test


NEW YORK – Berg Health, a biotech firm using machine learning to study diseases and inform drug development, has set its sights on commercializing a liquid biopsy biomarker test that can guide treatment decisions for pancreatic cancer patients.

Around 10 percent of pancreatic cancer patients survive for five years after diagnosis because their disease is often detected in later stages. Pancreatic tumors are typically detected using imaging, though sometimes high CA19-9 levels in the blood can point to the presence of cancer. Usually, by the time CA19-9 levels increase to a concerning level, the cancer has already progressed to an advanced stage.

Levels of CA 19-9, combined with CEA and CA125 levels, can also be used to identify which patients are likely to benefit from adjuvant chemotherapy or radiotherapy. Standard treatments for advanced pancreatic cancer include FOLFIRINOX — a chemotherapy regimen including leucovorin calcium, fluorouracil, irinotecan hydrochloride, and oxaliplatin. Gemcitabine, 5-FU, nab-paclitaxel are among other approved chemotherapy options.  Berg, headquartered in Framingham, Massachusetts, is hoping the biomarker test it is developing will improve oncologists' ability to personalize treatment for pancreatic cancer patients and improve their outcomes.

To identify the biomarkers in its test, Berg evaluated samples from 34 patients who were treated with FOLFIRINOX and enrolled in Project Survival, a collaborative project between Berg and the Pancreatic Cancer Research team to identify early detection, prognostic, and treatment-predictive biomarkers. Separately, Berg analyzed samples from 42 patients enrolled in a previous clinical trial of its investigational agent BPM31510, a molecule designed to correct cancer cell metabolism and reactivate cell death.

Berg profiled patients' samples using its Interrogative Biology platform, and the biomarkers were identified using its multi-omics machine-learning technology called bAIcis. That platform combines proteomics, metabolomics, and lipidomics data from patients' samples along with information from clinical records to develop cause-and-effect networks and perform statistical analysis. Using this method, the company has identified four protein biomarkers associated with keratin-regulated processes, WNT signaling, and cell cycle regulation that can predict patients' survival outcomes on FOLFIRINOX.

The researchers found that using the four biomarkers they could predict whether patients would survive less than or greater than 266 days on FOLFIRINOX, the average overall survival for patients who receive the treatment, with a positive predictive value of 88 percent and a negative predictive value of 93 percent. The test had a sensitivity of 94 percent and a specificity of 87 percent.

"When patients [with stage IV pancreatic cancer] only have on average six months to live, the trial and error to determine whether gemcitabine [plus] abraxane or FOLFIRINOX is better for them may cost valuable time," said Berg CEO Niven Narain. "If a patient is not going to respond to FOLFIRINOX, they don't have the two or three months to figure that out. By the time they fail the treatment, they may not respond to another treatment, so that's why you see some of these patients unfortunately expire so quickly."

Narain hopes the biomarkers BERG has identified, once validated in further studies, will help guide those decisions and help doctors avoid putting patients on treatments that they won't respond to. He added that the biomarker test could help doctors make a "more data-informed decision" at diagnosis for patients who don't qualify for surgery.

Although the biomarker validation studies are forthcoming, Berg is already starting to explore commercialization opportunities for a future test. The company has put together an advisory board to help take its research through validation and on to the market.

Narain noted that Berg previously developed a prostate cancer diagnostic in collaboration with the US Department of Defense and Uniformed Services University's Center for Prostate Disease Research. That multi-analyte test screened patients' blood taken before surgery and characterized their likelihood of progression.

Narain said his team is taking the lessons from that experience in prostate cancer and applying it to this biomarker test development program in pancreatic cancer. He hopes that by bringing on a partner that the commercialization process for the pancreatic test will be more "time efficient."

"Berg is not a commercial product company, we do clinical validation using our platform," he explained. "We're trying to decide if we should take this one out on our own or if we should partner with another diagnostics company or a lab."

While it figures out the best commercial path for the pancreatic cancer biomarker test, Berg has been sharing its ongoing research with physicians. The company presented its latest biomarker research at the 2021 European Society for Medical Oncology Congress last month. Berg is also working with the nonprofit Alliance of Families Fighting Pancreatic Cancer (AFFPC) to raise awareness of its research and forthcoming test. They hope to present the validated platform and supporting research to pancreatic cancer specialists at a future AFFPC meeting.

Meanwhile, the biomarker research within Project Survival is continuing. As part of that effort, Narain noted that researchers are using Berg's machine-learning platform to look for biomarkers in various types of samples, such as tumor tissue, blood, saliva, and urine.

Berg's investigational drug BPM31510 is also being explored in a Phase II trial as a neoadjuvant treatment for glioblastoma multiforme. In that study, researchers are similarly collecting samples to try to identify predictive biomarkers.

Narain said Berg will continue analyzing the pancreatic cancer data in the hopes of identifying other predictive biomarkers for the disease, like those that can support earlier diagnosis.

"In other cancers, survival is a few years, and you may be able to pull off failing a drug and then having a response [to another treatment]," Narain said. "Pancreatic cancer patients just don't have that time. In all its complexity and in the five years of doing the work [on Project Survival], the story we found here is very simple. [We developed a] tool to get ahead of what's going to work for patients as soon as possible."