Definiens CEO Thomas Heydler is leading the charge to put his company at the forefront of personalized medicine, starting with a big data approach in order help treat patients.
The imaging software company, started by Nobel Physics prize winner Gerd Binnig, seeks to identify and analyze biomedical images in order to help doctors treat patients as early and efficiently as possible.
The company’s Definiens Cognition Network Technology emulates the human brain in attempting to understand and anatomizes image, determining meaningful data within its context, cutting out noise to focus on important information.
The company is expanding into various diagnostic fields through a series of partnerships, including a recent deal with Metamark Genetics which made Definiens the sole provider of automated image analysis for Metamark’s ProMark prostate cancer prognostic test. Munich-based Definiens also closed a recent deal with Clarient Diagnostic Services to provide its software for immuno-histochemistry testing in breast cancer.
“We are able to datify tissue, extract all the information, and correlate that with other data sources," Heydler told us. "We use that, on the one side, to automate existing applications and essays in tests, but more and more collaborate with diagnostic companies to actually develop new tests that are desperately needed in oncology to do better patient stratification, to do better predictive tests for better treatment decisions."
In an interview with MassDevice.com, Heydler talked about where he thought personalized medicine was going and what role Definiens wants to play in it.
MassDevice: How did you land with Definiens?
Thomas Heydler: I have been with the company now for 10 years as the CEO of the company. Before Definiens, I was spending a few years in the U.S., in the Bay Area. I ran a California-based software company, and when I returned to Europe Definiens was presented to me and I found it to be one of the most exciting places to put my work in.
It was founded by Gerd Binnig, who won the Nobel prize in 1986 in physics, but he kind of evolved and went into the software business and developed a very, very unique technology. It is an analytical technology which can be used to analyze data much in the way we humans are doing that, using a cognitive- and context-knowledge-centric approach to analyze data. I found it so exciting that I said, "This is really really cool and where I can help you is I can take that raw technology and turn it into a scalable business.” That is what we are working on, which is a very exciting play.
MassDevice: How does big data relate to personalized medicine?
TH: To look to the core technologies, it is a technology used to analyze data. Originally it was independent to our text data, numerical data or images. Over the course of the last couple of years, we have basically taken that technology and really honed in on a very specific space, which is healthcare. Within healthcare, we’ve been focusing on tissue images, as part of this pathology. And what the core of the technology does is it is able to systematically extract tons of information out of the tissue.
It not only fully quantifies the information, but really understands the context and pathology of the tissue. Basically, you think about gene sequencing on the one side and you think about tissue datafication on the other side, where we pull out a huge amount of data out of the tissue. Having that information at your fingertips, it is a tremendous amount more information than a pathologist can even think of. This is already the first step of the big data approach.
But now you have all of this big data coming out of the tissue. Now you can, in the instance of personalized healthcare, think about a patient group – about 1000 patients – where you are looking for a differentiator to differentiate long-living patients to short-living patients. So now you have this patient cohort on this one side, you have the tissue, you also have the patient outcome information of those patients. Now you can use that and correlate the outcome data of the patient with features you are finding in the tissue.
That is really a great opportunity to find new tissue markers. This is a big data play, right? You are potentially talking about terabytes of data. Now you can actually take that information, those markers you find, and make that the foundation for going forward with new diagnostic tests or predictive tests.
I think this is the uniqueness about Definiens We are able to datify tissue, extract all the information and correlate that with other data sources. We use that, on the one side, to automate existing applications and essays in tests, but more and more collaborate with diagnostic companies to actually develop new tests that are desperately needed in oncology to do better patient stratification, to do better predictive tests for better treatment decisions.
MassDevice: How does over-treatment play into healthcare cost?
TH: If you look at the prognostic tests, you are trying to avoid unnecessary treatment. That is where you are using the prognostic tests. You know unnecessary treatment means unnecessary cost, whether you are talking about chemotherapies which can cost in the hundreds of thousands for treatment or whether we are talking about surgery. This also reduces patient pain, so I think it improves the quality of a patient’s life significantly. As far the prognostic side is concerned, I think there is a lot of money at stake.
MassDevice: So you are saying that without all the guess-work, you can bring the cost down substantially?
TH: Absolutely. We are calling this tissue phenomics, because we have really looked at the phenotype, which is represented in tissue, in order to understand how the disease has indentured itself in a given patient at a given time. It is very complimentary to the genomics side, because there is a lot of buzz around genomics right now and genomics and gene-sequencing is doing a great job of understanding the human genome, really understanding the sequence information. It only gives you a first stage prognosis as to how the disease develops. Only in very few cases can you have an exact prediction.
The Angelina Jolie case is an example where you have a relatively clear path from a gene mutation to an outcome, but in most cases you don’t know, so you really have to look and ask, “How does the gene mutation and the genomic code translate into tissue?” to be able to make an exact statement. To really do an exact prognosis and a much better prediction, the tissue is unavoidable.
Definiens is in a unique position to be able to have the technology to extract all the information from the tissue and drive better prognostic tests. Our goal is, over the next few years, to establish the tissue phenomics paradigm as a compliment to the genomic paradigms, that through the combination of molecular diagnostics and tissue diagnostics you get the best possible diagnosis and, from there, the best personalized treatment for the patient.
MassDevice: Definiens recently closed a Metamark deal for prostate cancer prognostic tests. What is the significance of that?
TH: Metamark is one of our typical clients where we had spent 2 years jointly developing that test. It’s all based on Definiens technology. The recent announcement was about the promark test going through clinical validation and really starting to commercialize the tests, taking the clinical samples at the beginning of 2014.
The exciting thing for us is that it is a real high-need, high-benefit area, prognostic prostate. I think about over 80% of the surgery is over-treatment and unneeded. The Metamark guys are trying to really do a much better prognosis to get better treatment options in order to avoid over-treatment of prostate patients, which I think will improve a lot of patients’ lives and will be successful going forward.
MassDevice: Personalized medicine has something of a motto: "Predict, preempt and cure." Does that fit for Definiens?
TH: I think that is pretty general, because what you try to achieve with personalized medicine is to find the most specific treatment to get the highest success rate. The further you go into the screening and predictive side the more you are talking about relatively broad data. You are, on a very high level, segmenting patients.
There might be a certain risk, but it’s not really able to give you certain treatment decisions. That was one of the problems with 23andMe, which was shut down by the FDA. They were too vague with coming up with some kind of prognostics and really having a lot of substance.
In turn, I think that this is not what I think personalized medicine is. Personalized medicine is, once you have an exact diagnostic profile of a patient, you know the status of the disease, what the specific disease is and how do you treat that.
Specifically in oncology, when you look to any expression of cancer it is critical to look into the protein structure of the tissue to make a decision. When people are getting [diagnosed with] cancer, I think normally they are not only having one, but, because of heterogeneity, having to deal with multiple cancer types at the same time. That’s why you really have to understand which cells are critical, how they are regionally and geographically distributed in the tissue and how you can tackle that.
Once you have really understood that, you can start to come up with a personalized treatment. I think this is true personalized medicine – not too much on the screening side. Yes, you can indicate you are at risk or you’re not at risk. When it comes down to medicine and treatment, you really have to understand the diagnosis. This is where molecular diagnosis can help and where tissue diagnosis is absolutely critical to make that judgment.
MassDevice: Definiens has had some changes in its leadership, with the addition of a new COO, CMO and a new clinical team. What was behind this move?
TH: The company started as an IT software-centric company, based on data analytics, and then over the course of the past couple of years we have been honing in on the healthcare space. We focus much more now on tissue diagnostics and digital pathology. With that, we are definitely evolving the position of the company to become more of a healthcare company and a tissue diagnostics company.
As a result of that, a key focus on my side is to really bring in healthcare and medical expertise into the company. Some of the announcement – with the chief operating officer having a strong background from Philips Healthcare and with the chief medical officer being a professor of pathology and a practicing pathologist that has spent significant time in the industry – is a perfect blend for us, having seen both worlds. We are constantly hiring people from that domain, we are adding expertise so we can be a strong partner to our clients.
We have also added a scientific advisory board and added key pathologists as members. Over the course of 2014, we will grow the board.
MassDevice: Definiens was also named Frost & Sullivan’s global company of the year. What does this mean to you?
TH: We are very honored and excited about that because because it shows that what we are doing with tissue datafication, which is really driving digital pathology forward, is also the foundation for new tissue diagnostics. That is why we won that award in those categories. I think we are really proud because there are some key players in that market and having a company like Definiens receive that award, we are very pleased, obviously. We are going to continue to work hard to continue to earn that type of award.
*This interview was edited for clarity.
At DeviceTalks Boston, Tyler Shultz will give attendees an inside look at Theranos and how he was able to sound the alarm after he realized the company was falling apart. Shultz will take attendees behind the story that everyone is talking about: the rise and fall of Elizabeth Holmes and her diagnostic company, Theranos.
Join Shultz and 1,000+ medical device professionals at the 8th annual DeviceTalks Boston.