Why do some patients respond to immuno-oncology drugs when others don’t?
It’s one of many million-dollar questions in medicine that confound companies, researchers, and clinicians alike. And the really frustrating part? We know where many of the answers lie. They’re trapped in electronic health records (EHRs) and siloed by disparate health systems.
Chicago, Illinois-based Tempus is working to extract that information at scale. The two-year-old company recently began offering an operating system, dubbed Tempus O, designed to structure, cleanse, and annotate clinical data.
Tempus O is one part of an end-to-end service that also includes full DNA and RNA sequencing at its CAP/CLIA-certified labs. However, the company found many of its clients wanted to focus on the data organization component, CEO Eric Lefkofsky said, to harness the phenotypic, therapeutic, and outcome and response data.
“People want to structure this data clinically because they believe that having that data at their fingertips will help them provide better care to their patients,” Lefkofsky said via phone. “And people want to structure that data for research because, obviously, it’s paramount for them to understand: Are there particular characteristics leading some people to have outsized positive or outsized negative responses to any therapeutic?”
It’s harder than it sounds.
While most medical records have been digitized in recent years, they’re not exactly “user-friendly” resources. They were designed for medical billing, after all. A lot of the important data is tied up in free text — those hastily written progress notes.
Tempus O taps into some sophisticated workflow tools, including optical character recognition and natural language processing, which extract meaning from this text. Those notes can then be compared and organized within a larger dataset, along with insights from research databases, images, and scans. Lefkofsky said the company has also built up a team of abstractors that can manually input data when necessary and review the finished work.
Of course, all of this is done at scale, to power real insights.
“To give you some perspective on that, we expect to structure around 400,000-patients worth of data in the next 12 months,” Lefkofsky said.
That’s almost one-quarter of the 1.7 million Americans expected to be diagnosed with cancer in 2017.
Unsurprisingly, there is big demand for this type of software.
In September, Tempus closed a $70 million Series C round co-led by New Enterprise Associates (NEA) and Revolution Growth. That cash injection brought the startup’s total funding to $130 million. (Lefkofsky, a serial entrepreneur, has also invested a significant amount of personal money).
It’s not the only player in the game. Palo Alto, California-based Syapse closed a $30 million Series D round in November, for a total of $71 million raised.
While Syapse also works to bring fragmented clinical, molecular, treatment, and health outcomes data together, it is focused on the software component. Rather than perform the sequencing, it has collaborations with various labs.
In a July interview with MedCity News, Lefkofsky said part of his company’s edge has been the ability to do it all, to simplify the relationship with the client.
However the molecular data is generated, both Syapse and Tempus agrees that the two data sources need to be analyzed together.
“As a company, we’re most interested in the combination of both the clinical data and the molecular data,” Lefkofsky said. “When you have the molecular data you can also answer the holy grail question, which is ‘why.’ Why are these patients responding well? Why are these patients not responding well? For cancers, that’s a molecular question often.”