Why Flatiron is betting on data quality, not hype, to power the next era of cancer care

You’ve taken the helm at Flatiron at a pivotal moment in your history. How do you define what the company’s core mission is when you are entering into this new chapter?

As I’ve taken over as CEO, I’ve been framing my vision in five pillars, which I think will be critical for our success.

Those are becoming a talent development engine; doubling down on technological and scientific innovation; delivering and delighting for our clients across all of our businesses; simplifying Flatiron; and lastly, of course, financial performance.

My vision anchors on supporting a unified global team in delivering more representative, higher-quality data faster and at scale, to hold our spot as the industry’s preferred evidence partner.

What matters most is building a strong, growing business that makes a real impact in cancer care and science. Focus and discipline are what will give us the resources to continue to invest in science, technology, and people, while also being financially sustainable.

Flatiron has built a significant real-world evidence engine in OncoEMR®. How are you planning to evolve that platform technologically and scientifically?

Our approach to global evidence is focused on turning global oncology data into decision-shaping evidence. I think of that as four things.

First, advancing the most important questions in cancer care. At Flatiron, we’re redefining how real-world evidence powers cancer research and care globally, combining direct electronic health record (EHR) access with trusted scientific and regulatory expertise to generate answers to oncology’s hardest questions.

Second, building the world’s most representative oncology evidence network. We’re expanding our global footprint to more than 10 million patients by partnering with national health systems, regional networks, and independent hospitals, to reflect care in every setting.

Third, using artificial intelligence (AI) and large language models (LLMs) to unlock deeper insights. Flatiron is responsibly applying advanced AI to move beyond what is happening, or what did happen, in the real world, to why something happened and what if. Extracting meaning from unstructured data, uncovering drivers of outcomes, and predicting future trends through approaches like digital twins.

And then, lastly, shaping the future of oncology worldwide. By bringing together global scale, cutting-edge AI, and deep expertise, we’re helping biopharma, regulators, and researchers make faster, smarter decisions that transform cancer care.

If I could add a couple of other points. Our Panoramic data solutions, which are LLM-powered, deliver access to Flatiron’s full patient network, unlocking up to 25 times larger patient cohorts across six-plus cancers, with several more launching next year.

And lastly, we’re built on direct EHR data captured via our OncoEMR® medical record system, and that Panoramic data delivers rich longitudinal evidence and supports endpoint analysis across the entire patient journey, from diagnosis through advanced disease.

If you could go into a bit more detail about artificial intelligence, and how machine learning is helping to power the way that Flatiron’s collecting and interpreting data.

Yes, absolutely. At Flatiron, we’ve been harnessing the power of machine learning and LLMs since well before I started, five-plus years ago.

The difference with Flatiron is, one, we’re a science-first organization with a relentless focus on data quality. That comes with a hesitancy to over-index on novel methodologies that aren’t aggressively validated and ready for center stage.

Second, while we haven’t always led with ML and AI in our company story in the market, we’ve been publishing about it in peer-reviewed articles since 2019.

Responsible AI in healthcare isn’t just about innovation. It’s about accountability to the patients whose lives depend on it. One of the big unlocks for us in applying these technologies was developing a framework for evaluating data quality, which we call the VALID framework. That’s what enables us to know, at both a disease level and a variable or endpoint level, how our models are performing on accuracy, which is so critical to the conclusions we draw.

Flatiron’s use of AI and LLMs is to unlock deeper insights. We’re responsibly applying advanced AI to move beyond what is, as I mentioned, to why and what if. Extracting meaning from unstructured data, uncovering drivers of outcomes, and predicting future trends through approaches like digital twins.

Having mentioned peer review, do you work hand in hand with academia in that respect?

Yes, a total of 1,920 publications that are either Flatiron-authored or have utilized Flatiron’s data. Scientific rigor and publication are a huge part of what we do.

Would you be able to describe some of the new therapeutic or biomarker areas that Flatiron’s intending to deepen with the scientific engagement?

Yeah, absolutely. We’re really excited about this.

We’re continuing to innovate with AI to help us ask better questions and get to answers faster. Obviously, our therapeutic focus is oncology.

The first bucket is digital twins. It’s a somewhat generic phrase, but we think of these as advanced predictive models that help us go beyond descriptive questions of what has happened and move into predictive and causal analytics, answering questions like what will happen or why did this happen.

Predictive models allow us to simulate the what if. How would this patient have responded if they received a different treatment? At their core, these are virtual representations of patients that have been built using data from larger cohorts of patients with similar diseases.

At Flatiron, we incorporate a wide range of information into these models, from clinical characteristics to genomics to socioeconomic factors and more.

We see these models supporting clinical trial design by simulating outcomes of different patient populations, contextualizing trial results by predicting control arm outcomes for single-arm trials or even increasing the statistical power of late-phase trials. That’s one bucket.

The other bucket is what we call ‘physician insights’. This is one of the most exciting things we have going on, because it helps uncover the why behind clinical decision-making. What’s driving a physician’s choice around efficacy, safety, convenience, or even patient access and insurance?

For our biopharma partners, that means objective, real-time insights at scale, not based on perception or recall like traditional market research, but grounded in actual clinical behavior, including things like offering a clinical trial to a patient.

Beyond the commercial value, this is about improving patient care. Helping the industry understand unmet needs, how treatment decisions are made in the real world, and how we can design interventions, or trials, that better reflect clinical reality by contextualizing the considerations that went into those decisions.

To give a tangible example, this approach can shine a light on radioligand therapy uptake in prostate cancer. You might find that reimbursement for the prostate-specific membrane antigen (PSMA) biomarker scan is the barrier to uptake. That’s not a traditional real-world data output sitting in the EMR as a structured field. It’s context on why that treatment decision was made.

Our proprietary LLMs allow us to filter through millions of patient pages of information to surface that context on why decisions were made.

Concerning those physician insights, are you working in partnership with clinicians to be able to take those insights on the ground? Is that done using an app, or how are you getting that data?

It’s actually part of the context entered into the electronic medical record (EMR), in unstructured fields. Often, it’s the documentation of the reasoning behind the decision.

I’d add that we’re not naming individual physicians and saying it’s Dr. X doing Y. It's an aggregated sentiment of what’s going on in the field as we see it.

I almost think of it as a Bloomberg Terminal for what’s happening in real time in the treatment and care of cancer patients.

Regarding Flatiron’s scientific advancements, how are you translating them efficiently into commercial value?

As I mentioned earlier, my first priority stepping into the CEO role was continuing to lean into making Flatiron a talent development engine.

That’s because achieving these results is only possible with the incredible team we’ve built here and continue to foster.

This diverse combination of expertise across our teams and functions is critical. From deep business understanding to the scientific levers required to drive clinical value, to having medical oncologists, technologists, and medical informaticists on staff alongside traditional functions, that’s what allows us to tackle these problems in a truly cross-functional way.

With regards to Roche, your parent company since 2018, how are you leveraging Roche’s capabilities while preserving Flatiron’s agility?

Flatiron has always had really strict firewalls in place with our parent company, Roche, and we remain an independent subsidiary.

We do that because the vast majority of our partnerships in biopharma and on the clinical side are with peers of Roche, not specifically with Roche itself.

While we remain focused on our core growth engines at the point of care and in global evidence, we see Roche as a great supporter, but not someone we engage with on the specifics of the business.

How do you see the competitive landscape currently in oncology and real-world data? What would you say is the Flatiron ‘secret sauce’?

One thing I’d say is that cancer is such a terrible disease that we try not to think in terms of competitors, but as peers; people trying to make the world better on behalf of all of us who will be impacted by cancer, one day.

That said, the accelerating pace of industry-wide innovation is exponential, both outside our organization and within it. It’s a clear indicator that real-world data and evidence are increasingly important to partners across the industry.

The space is getting more crowded as demand grows. But, in our view, we’re the only company that can offer this level of connected global oncology insight with one platform, one standard, and one ecosystem.

I’d also point to our strong relationship with the community oncology setting through our OncoEMR suite of solutions and the incredible clinicians providing care in those settings.

With regards to global expansion, what are you seeing as the biggest strategic risks, and how are you mitigating them?

Let me give a bit of context on how we think about protecting our lead.

We power the largest and most globally diverse collection of oncology data available today. Our commitment to data quality continues to differentiate us in the marketplace.

Looking ahead, we’ll continue to differentiate in a few key ways.

One is scale and data advantage. What sets us apart, in the AI era, is our access to high-quality, source-level clinical data that’s curated, validated, and linked across thousands of providers and millions of patient journeys. That depth and fidelity create a foundation that few others can replicate.

The second is speed and scientific rigor. AI is only as good as the data and science behind it. Flatiron combines AI-driven efficiency with scientific validation, allowing us to generate insights faster without compromising the rigor that healthcare demands.

One of our favorite framings internally is “move fast and validate things.” A big part of that validation is our nearly 2,000 total publications.

Most recently, we issued a curtain-raiser for the upcoming San Antonio Breast Cancer Conference, including a spotlight poster on using LLMs to explore real-world use of GLP-1 medications among breast cancer patients and their potential impact on treatment outcomes.

Being able to generate timely, relevant answers to emerging scientific questions is what makes Flatiron unique.

What do you see as being your top priorities for revenue growth over the next three years? How do you expect the business model to evolve?

We’re entering the next three years with a relentless focus on where we can lead, grow, and make the biggest difference.

Our two growth engines are the global evidence business, where we operate not only in the US, but also in Japan, the UK, and Germany, and our point-of-care community oncology business.

Building the world’s most comprehensive global evidence platform and scaling our offerings through AI-driven data extraction is critical.

Cancer treatment options have increased by roughly 90 percent annually since 2019. Cases are projected to grow from 20 million to 32 million annually by 2050. There are more people affected and more treatment options than ever, and it’s increasingly complex.

AI data extraction enables us to generate more meaningful data from more patients, support cohort selection and analysis, and make it easier to find and evaluate patients.

We think of this as making data more liquid and accessible, helping biopharma and researchers move faster, make better decisions, and, ultimately, accelerate delivery of innovative treatments to patients.

Thinking about the future, what would you like Flatiron’s progress to reveal about your personal approach to leading in data-driven healthcare?

Flatiron’s story has always been grounded in science and technology. It’s about learning, validating, and improving. We don’t chase hype. We build trust.

That’s what will continue to set us apart.

We’re entering one of the most exciting moments in cancer care. Science is evolving fast, and with the right data and insights, we can help accelerate that progress responsibly.

I want to continue building the world’s most comprehensive, scientifically rigorous cancer data platform. To go beyond descriptive analytics and provide predictive power, not just showing what’s happening in cancer care, but uncovering why it’s happening and modeling what will happen next.

We’re pushing our teams to lead with an anticipatory mindset, thinking about the questions customers and clinicians will ask before they even ask them.

Most importantly, we want to do this while continuing to set the standard for quality and scientific innovation. We’re the only company that can offer this level of connected global oncology insight with one platform, one standard, and one ecosystem.

I’m excited about redefining what’s possible in oncology evidence generation.

And one final thought we never forget at Flatiron: behind every data point is a clinician caring for a person experiencing cancer. That’s what grounds us and why we’re so committed to advancing the future of oncology.

If there were one piece of insight you’d want to pass on to the next generation of executives in healthcare, particularly diagnostics, what would it be?

I’d come back to something I mentioned earlier. When you have the privilege of operating in healthcare, scientific rigor and quality are mandatory.

At the end of the day, there’s a clinician and a patient on the other side of every data point. Decisions must be grounded in high-quality data and rigorously assessed approaches, because the decisions we’re making can be truly life-altering.

What questions should we ask Nathan next? Let us know in the comments.