CNBC | Daniel Bukszpan
Tempus, a Chicago-based technology company that offers genetic testing and aggregates clinical information, initially focused on cancer patients. But with the coronavirus pandemic spreading across the world, it’s joining the fight against Covid-19.
Tempus was founded in 2015 by its CEO, Eric Lefkofsky, who previously co-founded the e-commerce company Groupon in 2008. But while Tempus’ mission may seem far removed from that of an online discount marketplace, Lefkofsky had personal reasons for branching out.
“My wife was diagnosed with breast cancer about five years ago, and I was amazed how little data was actually used as a part of her therapy, largely because our system makes it hard for doctors to access data when making real-time clinical decisions,” he said. “It became clear to me that I needed to try and tackle this problem, and I founded Tempus soon afterwards.”
Tempus, which ranked No. 6 on the 2020 CNBC Disruptor 50 list, has three primary lines of business. It offers genomic tests that physicians can order, its most popular being a test that looks for over 600 genes associated with cancer. (It does not offer any direct-to-consumer genetic testing, like a 23andMe.)
It also partners with academic medical centers and community health-care systems to organize and aggregate clinical data, primarily data held in electronic health records. The clinical data is de-identified to protect patient privacy, then typically used in research projects to help understand patterns. Tempus claims to have one of the largest molecular and clinical data libraries.
There is crossover between these two businesses, as data from genomic tests that run in its Chicago and Atlanta labs can be de-identified and added to the molecular and clinical data library.
Finally, it curates and sells de-identified data to pharmaceutical companies working on drug discovery.
Tempus is a covered entity under the Health Insurance Portability and Accountability Act (HIPAA), so all of the data is scrubbed of patient name, date of birth, address and any other data point that could identify an individual.
Early on, Lefkofsky felt that the Tempus platform should apply to a broad cross-section of patient data. For this reason the Tempus platform was designed to be what the company calls “disease agnostic,” which means it can be applied to patients suffering from many kinds of illnesses, including cancer, depression, diabetes, cardiovascular disease and Covid-19.
Using real-world Covid-19 patient data
In April, Tempus began offering a polymerase chain reaction (PCR) test for Covid-19 from its Chicago and Atlanta laboratories. It currently expects to build its testing capacity to approximately 10,000 tests a day in the coming months. The test, offered through physicians, was rolled out parallel to the start of a research project to gather data on Covid-19 patients and look for patterns that can help physicians treating patients as well as drug-discovery efforts. Procedures may vary from patient to patient, but this is where Lefkofsky thinks the company’s AI platform can do the most good.
“We are focused on collecting real-world evidence for patients that are Covid-19 positive to help doctors stratify patients based on risk and help triage patients to the optimal therapeutic path for them, based on their unique phenotypic characteristics,” he said.
Dr. Scott Gottlieb, resident fellow at the American Enterprise Institute and special partner at New Enterprise Associates, sits on Tempus’ board of directors. He joined after serving as commissioner of the Food and Drug Administration, where he spearheaded efforts to advance the use of artificial intelligence to create individualized patient care.
“I was impressed by the company’s ability to rapidly scale its sequencing operations to support major academic institutions and community oncologists across the country and get data to patients and providers,” he said. “They were focused not only on the back-end technologies to enable these opportunities, but equally important the front-end interfaces to make this information useful at the point of care.”
It also launched an initiative whose goal is to aggregate data for 50,000 patients who are Covid-19-positive. The hope is that this data will help reveal genetic characteristics that cause specific Covid-19 outcomes, as well as treatment practices for the illness.
Gottlieb added that Tempus is partnering with employers and health-care providers to support Covid-19 testing in states where restrictions are being relaxed, and employees are going back to work, so they’ll know what to watch out for.
“We’re going to have to be more aggressive in detecting and identifying respiratory viruses,” he said. “To these ends, Tempus is also working on a next-generation sequencing panel for respiratory pathogens, including Covid-19, to diagnose patients and create a library for infectious disease research that could reduce the impact of future epidemics.”
Current limitations to AI use in medicine
While Tempus’ use of health data as a way to accelerate research shows promise, there are limitations.
Dr. Jose Morey, Chief Medical Innovation Officer for Liberty BioSecurity and advisor for MIT Solve and NASA iTech, said that one problem in relying on AI in the case of a new virus is that there isn’t enough data at the moment.
“You have to have large data sets to be able to train, test, and validate,” he said. “When you have something new and novel like this, they just don’t exist. … It will come eventually, but not yet.”
He added that the technology is also only as good as the data, and currently, there are systemic flaws in the way medical data is gathered and recorded. For example, he said that data in medical centers could be “dirty.”
“This means that the data is not structured to be plugged into the math that is AI,’” he said. He added that health-care data is also frequently “siloed,” or separated into unconnected compartments. For AI to do its job, it requires large amounts of interconnected data, which Morey said doesn’t exist in the U.S. health-care system.
“Health-care entities, such as device manufacturers, electronic medical record vendors, and hospitals, are not incentivized to share data amongst each other,” he said. “This is why there are so many AI companies doing incremental algorithm and application development.”
While this is a challenging moment for the U.S. health-care system, Tempus has been recognized by the medical community and investors as a company whose platform shows great promise.
The company has raised $5 billion from such investors as Baillie Gifford, Franklin Resources and T. Rowe Price. It has also entered into a partnership with CVS Health, which allows oncologists in the CVS Health/Aetna network to use the results of its genomic sequencing tests.
Covid-19 research, like cancer research, has yet to yield a cure. However, Gottlieb said that he sees AI tools as indispensable for cancer patients, infectious disease patients and others in the future.
“As we collect more data and apply AI tools to large datasets, we can direct care to patients at greatest risk,” he said. “Artificial intelligence could help us solve a number of serious public health problems and improve outcomes for patients with infectious diseases.”