MolecularHealth, Backed by SAP Founder, Debuts Software to Match Patients' Genomes with Best Drugs
When executives at medical software developer MolecularHealth were working on a new product for oncologists, they consulted with a cancer center that was grappling with an emerging trend: About 100 patients a year were paying for their own genomes to be sequenced. Jeffrey Marrazzo, chief business officer of MolecularHealth, a Swiss-founded company that recently set up its commercial office in New York, won’t name the cancer center, but he recalls the conversation vividly. “These patients were walking into their oncologists, handing them disks containing their genomic data, and saying, ‘I don’t want you to treat me how you treat every other breast cancer or lung cancer patient—I want you to treat me based on what’s on this disk,’” Marrazzo, says. “I’m not going to tell you that tomorrow every patient is going to pay for their genome to be sequenced, but with price points falling, it opens up a new world in terms of patients willing to push the system.”
MolecularHealth has spent the last eight years developing software platforms that its founders hope will place the company at the leading edge of that movement towards personalized medicine. In January, the company—which is backed by Dietmar Hopp, former founder and CEO of software giant SAP—began marketing its first two products, both of which are designed to translate massive amounts of genomic data and scientific evidence into better treatment choices for patients. “Our idea is if we really want to improve therapies, we must understand the individual’s molecular setup and map that to what’s known about disease pathways,” says Friedrich von Bohlen, CEO and founder of MolecularHealth.
Towards that end, MolecularHealth is introducing what it calls a decision-support software platform for oncologists. Here’s how it works: A physician treating a patient who’s just been diagnosed with cancer can input everything that’s known about the case, from the exact tumor type, to the presence of disease-promoting proteins in the patient’s lab tests, to the patient’s entire genome, if he or she has it. If a patient has a specific genetic variant that may make him or her respond best to certain therapies, the software will pick up on that.
MolecularHealth’s program then combs the broader universe of medical literature on the particular cancer being treated, and delivers back to the oncologist a ranked list of all the treatment options. The suggestions aren’t limited to approved drugs; the software will also pull up information about clinical trials that the patient might enroll in, or even suggestions for drugs that aren’t FDA approved for that specific cancer, but that other oncologists are commonly using to treat it, Marrazzo says. “What the doctors get is cancer-specific prioritization,” he says. “The technology takes raw information and turns it into something that’s useable in a clinical context.”
MolecularHealth has been fine-tuning its decision-support software with the help of the prestigious MD Anderson Cancer Center in Houston, TX. The hospital is using the platform in its newly formed Institute for Personalized Cancer Therapy, which has a mission of incorporating genetic data from individual patients into its treatment plans. “Physicians who work in oncology are beginning to understand that the ability to pull data from multiple sources and build alternative treatments from that is the wave of the future,” says Dan Fontaine, senior vice president for business affairs at MD Anderson. “We’re all enthusiastic about the idea that the next great breakthrough will be driven by the capabilities of computational science.”
MolecularHealth founder von Bohlen and his startup team spent much of their first decade in stealth mode, working through a series of technological hurdles. One of the toughest challenges, he says, was inventingtechnology that could automatically curate usable information from the scientific literature on every cancer drug that’s on the market and in pharmaceutical pipelines. “The data is dirty. For example it has spelling errors, and many drugs have different names at different times,” von Bohlen says. “That was one of the many hurdles.”
Von Bohlen, who previously founded a German bioinformatics company called Lion Bioscience, founded MolecularHealth in 2004 with a small amount of seed financing. The following year, he met SAP’s Hopp, and the two bonded over their shared passion for applying technology advances to personalized medicine. They started an investment firm called Dievini Hopp Biotech, which went on to become MolecularHealth’s sole investor.
When it came time to commercialize its software in the U.S., von Bohlen chose New York City as a hub and called upon Marrazzo to run it. Marrazzo had some experience in personalized medicine: He helped start Generation Health, an Upper Saddle River, NJ-based company that made technology to help health insurers decide which genetic tests are the most useful and how they should be covered under health plans. CVS Caremark bought a majority stake of Generation Health in 2009. Marrazzo says he had a particular idea of what the next big thing in genomics would be—an idea that matched up perfectly with MolecularHealth’s vision.
The oncology community has historically embraced tools that make personalizing treatments easier. Drugs that include tests to identify patients most likely to respond based on their genetic profiles, such as Novartis’s imatinib (Gleevec) and Genentech’s trastuzumab (Herceptin), have become blockbusters. And recently, Redwood City, CA-based Genomic Health (NASDAQ: GHDX) has found a willing audience for its cancer assay, Oncotype DX, which probes the biology of breast tumors to predict how likely they are to recur. “These are examples of the genome finally helping us,” says Rajiv Datta, medical director of the Gertrude & Louis Feil Cancer Center at South Nassau Hospital in Valley Stream, NY.
Any software tool designed to incorporate genomic data into treatment plans would be welcome, Datta says. “The human genome is so vast it’s hard to identify what we are looking for and how we should use it,” he says, adding that as the price of mapping the genome continues to fall, he expects more physicians will demand tools that help them make sense of all the extra data they can collect from patients. “Just as a lender would look at a homebuyer’s credit score, we will look at a person’s genetic map,” he says. “Anything that helps us do that will be at the forefront.”
MolecularHealth’s second software platform is called MASE, which stands for Molecular Analysis of Side Effects. MASE was designed to help drug developers and others who are concerned with drug safety to predict and ultimately avoid potential problems, such as dangerous drug interactions. The technology scans the biomedical literature and then pulls together information such as what targets a particular drug hits in the body, and which genes are involved in metabolizing it. On February 14, MolecularHealth announced it had struck a five-year agreement with the FDA to evaluate and refine MASE.
Marrazzo says MolecularHealth has started marketing MASE to pharmaceutical companies, insurers, and others who work in the field of drug-safety science.
As for the decision-support product, MolecularHealth has installed it at several cancer centers and will be officially rolling it out in the next month or so, Marrazzo says. Although the company is focusing on oncology, the platform could ultimately be adapted to other diseases, Marrazzo says. “The premise here is to provide a system that consumes patient information—both clinical data and molecular data—and transforms it into a list of potential treatment strategies,” he says. “Our goal is for medical decision-making to result in the safest and most effective treatment choices for patients.”