Skip to main content

Berg, DoD Using AI Platform to Ready Diagnostic Test for Prostate Cancer

Premium

CHICAGO (GenomeWeb) – A longstanding collaboration between bioinformatics firm Berg and the US Department of Defense is getting ready to bear fruit.

Berg is, according to Senior Vice President and Chief Analytics Officer Slava Akmaev, in the "final stages" of commercializing a diagnostic test for prostate cancer, in conjunction with the DoD Uniformed Services University's Center for Prostate Disease Research.

An announcement about the product, which helps clinicians distinguish between benign prostate hyperplasia and actual prostate cancer, is expected soon. "Current testing such as [prostate-specific antigen] is not very good in differentiating those two conditions," Akmaev said.

Framingham, Massachusetts-based Berg entered into a research partnership in 2013 with the Center for Prostate Disease Research at Walter Reed National Military Medical Center in Bethesda, Maryland to find biomarkers for that cancer. Last year, Berg entered into a similar arrangement with DoD's Clinical Breast Care Project to apply its artificial intelligence platform, named Interrogative Biology.

Under the prostate partnership, Berg has been able to access DoD's prostate cancer biobank to cull and analyze data related to various stages of the disease, Akmaev said. "Using our artificial intelligence technology, [we have been able to] decipher what molecular mechanisms are at work in patients at various stages of cancer development," Akmaev said.

"It's multi-omic information," he explained. It includes not just genome data, but also the proteome, metabolome, and the lipidome of the biological samples.

While Berg mostly focuses its work on oncology, neurology, and endocrinology, it is a multifaceted company. Berg has an ongoing relationship with the University of Miami's Leonard M. Miller School of Medicine in cardiovascular disease.

Products and services cover pharmaceutical research and development, drug discovery, diagnostics, health IT, and digital health. "We're not a one-trick pony. We have a very vibrant drug discovery and drug development program. Our entire pipeline is internally developed," Akmaev said.

Berg also has operations in diagnostics as well as in precision medicine, and the latter is not limited to genome sequencing and interpretation. For example, Berg has its own mass spectrometry facility in Framingham that can process samples for proteome analysis.

The common thread holding most of this work together is patient stratification for population health management and for clinical trials, based on the safety and efficacy associated with various biomarkers, Akmaev explained.

For analytics services, Berg is targeting research institutions, large and mid-size pharmaceutical firms, and health IT vendors, plus the company has had "several conversations" with payer and provider organizations, Akmaev indicated. "This is really focused on analyzing big data in health IT, looking at patient records, pharmacy information, [and] claims data," he said.

Berg supplies its clients with proprietary reports covering risk factors, predictive features, and algorithms. "Either we or our partners implement in their software these predictive features at the point of care," Akmaev said.

The company collaborates with IT vendors who can integrate Berg's intellectual property into their software, generally more hospital information systems than electronic health records.

In January, Berg announced a data partnership with Becton Dickinson's MedMined inpatient safety surveillance monitoring system. BD has software to integrate that information into hospital information systems, though not electronic health records.

Interrogative Biology  is "the engine behind the scenes," Akmaev explained. The Berg Analytics division of the company "provides this artificial intelligence engine component that can crunch the data as the back-end data analytical support."

One pharma-related effort underway is combining clinical records and molecular data.

"That's a very exciting project because we are going to be getting biological samples from our collaborators and then doing this multi-omic profiling, and then using our artificial intelligence," Akmaev said. "We have a technology basis to link disparate data sources — proteomics, lipidomics, metabolomics, genomics, and clinical information — and [create] actionable data that can be commercialized or used in our partners' commercial pipeline."

Berg plans to provide more details on this pharma partnership in about a month.

Already public is a partnership announced in May with the Dystrophic Epidermolysis Bullosa Research Association of America. Berg is looking for clues into epidermolysis bullosa, a rare disorder of the connective tissue that can have a wide range of genetic variations.

In population health management, Berg is involved with Project Survival, a multiyear collaboration led by Beth Israel Deaconess Medical Center in Boston. That team, which also includes the Pancreatic Cancer Research Team (PCRT), Cancer Research And Biostatistics (CRAB), and several other teaching hospitals, seeks biomarkers for pancreatic cancer.

"This project is focused on recruiting pancreatic cancer patients, collecting longitudinal biological samples across multiple years, following those patients, collecting their outcomes, and then doing extensively deep molecular profiling of those biological samples for multi-omic signatures," Akmaev said. Berg's AI software will help develop predictive diagnostic tests for physicians to stratify patient risk in hopes of speeding diagnosis of this notoriously aggressive form of cancer.

"Right now, treatment of pancreatic cancer is a huge, unmet medical need. There really are not many options for those patients, and life expectancy is really short," Akmaev noted.