Berg Hopes Big Data Will Lead to Breakthrough for Pancreatic Cancer

An estimated 40,560 people will die from pancreatic cancer this year.

The disease is the twelfth most common cancer with only 7.2% of those with the disease surviving five years after diagnosis, according to the National Institutes of Health.

It’s known as a silent killer. One that has, in recent years, taken the lives of some of America’s most beloved celebrities including former Apple (NASDAQ:AAPL) CEO Steve Jobs, and actor Patrick Swayze. It is the kind of cancer that goes virtually undetected until it’s in advanced stages, leaving the patient with very few options to live. Worse, its expected, within five years, to surpass breast and colon cancer as the second most deadly cancer in the United States, according to research from the Pancreatic Cancer Action Network, published in the American Association for Cancer Research’s Cancer Research journal.

But a biopharmaceutical company near Boston called Berg is doing something new to try and change that.

In April, the company announced a seven-year initiative to bring its big-data methodologies to the pancreatic cancer playing field. It's invested $7 million to complete what it calls “Project Survival,” Berg formed a partnership with heavy hitters in the medical community including the Cancer Center at Beth Israel Deaconess Medical Center and a Harvard Medical School teaching hospital. In addition, the Pancreatic Cancer Research Team managed by Cancer Research and Biostatics is also teaming up with Berg, offering a 48-site multinational network of cancer centers and expert scientists who are dedicated to finding a cure for pancreatic cancer.

Through the partnership, the team’s goal seems simple: To find biological markers (biomarker for short), or indicators, of pancreatic cancer, and then work to develop new courses of treatment to help those diagnosed survive.

But the process isn’t as cut and dry as it might seem. There are no known biomarkers for pancreatic cancer, so if Berg and its partners are able to identify one, it would be groundbreaking.

Doctor Donghui Li, a professor in the Department of Gastrointestinal Medical Oncology in the research division of cancer medicine at the University of Texas MD Anderson Cancer Center has been working on a control study for pancreatic cancer since 1998 that focuses on environmental and genetic responders.

“It’s everyone’s dream to find a biomarker for early detection [of pancreatic cancer],” she said. “But it still remains a big challenge.”

That’s because pancreatic cancer is relatively rare and by the time patients are diagnosed, the cancer is in a late stage, and the patient is given about six months to live with very few treatment options available.

“Individual hospitals and institutions don’t see many patients each year, and they die very fast, before they can be included in a study,” Li explained. “For an epidemiology study, it’s hard to get to the patient in time. That’s the first big challenge. From the clinical point of view, it’s difficult to give a correct diagnosis because the diseased organ is so deep in the abdominal cavity, it’s hard to reach.”

Li went on to explain other problems that prevent doctors from correctly identifying pancreatic cancer at an earlier stage. For example, there are risk factors for pancreatic cancer including smoking and type II diabetes. But if 40% of the population smokes – who do doctors screen? It’s not realistic to be able to screen all high-risk smokers in one single study, or everyone who has type II diabetes.

She added diagnosis of the cancer primarily depends right now on costly CT scans, ultrasounds, and biopsies that are prohibitively expensive for many patients.

“We definitely need broad-based, non-invasive biomarkers,” Li said. “From past research, we’ve all come up with the same concept: There’s probably no one single biomarker to cancer. We probably need three to five or more to work together as a battery of biomarkers to really tell a part of pancreatic cancer from another disease like pancreatitis, diabetes, or other benign desires.”

That’s where Berg comes in. The goal for the research teams working together in this partnership is to identify a biomarker that doctors could screen for at any patient’s annual check-ups.

“The first part of the process is getting the right tissue samples from these patients,” Niven Narain, Berg president and CTO said.” Jim Moser from the Beth Israel Deaconness Medical Center in Boston is going to be doing very unique types of surgeries that he’s pioneered in the field over the past 15 years where he’s able to get pieces of pancreatic cancer tissue in addition to, in some cases, normal, healthy tissue from the same patients. That’s the really amazing starting material that we’re going to have to start this process.”

The unique part about Berg’s approach is that it can, from one patient, obtain both healthy and diseased tissue to compare and contrast the difference. At the same time, the team will analyze both the healthy and diseased cell make up including genes, proteins, metabolites, lipids, etc. that would, in theory, lead them to identifying a biomarker for the cancer.

“We’re able to run the same biological tests [on both kinds of tissues] so that we can understand what’s gone wrong in a part of the pancreas that’s leading to a cancer state compared to that healthy tissue in the same patient,” Narain said.

From there, all of the data is put into the Berg Interrogative Biology Platform, a proprietary artificial intelligence application that allows the scientists to integrate all of that molecular data with the patient’s clinical and demographic information. What comes out, Berg contends, is a much more effective and efficient way to develop drug-treatment plans that offer a more personalized approach: It would suggest, based on a specific patient’s biological make-up, their demographic information, and clinical information, a specific drug, or combination of drugs to treat the cancer – or a way to ID the cancer exists before it becomes late-stage.

"Don’t get me wrong, this is not easy. The folks who have tried are excellent. But if you can get the best starting materials from those brilliant guys and apply it to our technology, we feel confident we can do this.”

- Niven Narain, Berg President and CTO

That process, which Narain expects to take about two to three years, will allow the scientists to use their tissue and clinical samples to pinpoint potential biomarkers for pancreatic cancer, which they’ll then put through a validation phase to ensure their data is accurate and confirm they’ve found the global biomarker for pancreatic cancer.

“This step is to take the output of that data from the artificial intelligence and start to validate that by going into either blood samples, serum samples, urine samples in these patients so that we can see, based on what the AI is predicting for being the difference between healthy and disease: Do we actually see these things showing up as a differential in blood samples or urine samples. That would be the last step in the validation when we would have discovered and validated the first ever biomarker for pancreatic cancer.”

It All Lead to This

For Narain, the motivation to begin the project was easy to find.

“No one knows how to tackle this,” Narain said. “I hear oncologists tell patients they can only do hospice or palliative care. For moms who want to see their kids graduate or get married – I’ve seen that happen hundreds of times and it bugs the hell out of me.”

Right now, Li said one of the biggest disappointments of the years spent researching this disease is that there are no viable treatment options other than surgery.

“For early-stage pancreatic cancer, if they’re still resectable, surgery is the only cure. Only 20% of pancreatic cancer is diagnosed at that stage, though. And because a tumor already involves a large vessel or organ you can’t take out, among those 20%, only 25% can live up to five years afterward because the tumor will come back or metastasize to other organs.”

She went on to say targeted therapies exist to treat cancer, but none really work for pancreatic cancer. Not only that but clinical trials are offered to diagnosed patients, but she said typically, the patient gains only about a month of life from those, and that’s seen as a “big breakthrough” for the disease.

For Narain, this new project represents an armor or tool kit that will be within reach in a matter of years, not decades. If all of the information is out there, floating in different places, waiting to be matched with the right set of scientists with the right toolset...why have biomarkers not been identified?

Narain said it’s not for a lack of effort. It’s that the approach has been completely different.

“Those [attempts] have been hypothesis-driven as opposed to taking samples from live patients, running them through a platform and through AI. It’s tough. Don’t get me wrong, this is not easy. The folks who have tried are excellent. But if you can get the best starting materials from those brilliant guys and apply it to our technology, we feel confident we can do this.”

To prove it, Narain points to Berg’s work with the Department of Defense, a five-year partnership the company started last year to identify biomarkers for prostate cancer. In that time, the scientists working on the project have completed the clinical validation of the biomarkers identified, and this year marks the start of clinical trials on the treatment kit.

But perhaps the most promising part is the path to discovering the biomarkers for pancreatic cancer is the exact same one Berg used with the DoD to treat patients with prostate cancer. All of the logistical components the scientists need to even begin the project, and the parameters within which they need to operate, are already to scale and set, ready for the next project.

“We know how to do this [with pancreatic cancer] because of our DoD relationship,” Narian said. “We know how to set up certain protocols and what controls for healthy tissue are, certain technology elements, and how to set up AI to analyze the different samples, then put in the clinical data to match the samples. It’s all been fine-tuned and those are all locked processes.”

Li said there’s hope at the end of the tunnel…it’s just a matter of finding the right pathway to the exit.

“Technology wise, there’s a lot of a breakthrough like next generation sequencing, micro-ray approach which looks at many proteins and antibodies all together,” she said. “At MD Anderson, we have a moonshot program to find low-hanging fruit to find something and make an impact on cancer in five years. This is one of the things everyone is interested in. But I’m not that optimistic to think we can really make a major breakthrough in a short time period.”

Narain, however, is.

“The interrogative platform has allowed us to really delve so deep and uniquely into patient populations and develop drugs like BPM31510 that can be matched with specific patient populations and get away from this one-size-fits-all approach to medicine to really a new paradigm that allows us to select the patients that are going to do the best on certain treatments,” he said.

He said if it works, that approach will lead to core changes in cancer research. He said once his team and partners have figured out how to take on prostate cancer effectively, breast cancer will be the next deadly cancer squarely in their crosshairs.

“We put a light at the end of the tunnel through this relationship, and our job is to drive the train out of the tunnel,” he said.