Greg Talbert, Head Personalised Healthcare Partnering

Partnering to fulfill the promise of personalised healthcare

By Pierre Valette, Emmy and Peabody-Award winning journalist and TV producer.


Raju Kucherlapati, Ph.D.,  has seen a lot of medical advances in his lifetime. As a contributor to the National Institute of Health’s Human Genome Project, the Harvard Medical School genetics professor helped sequence the first human genome. But even when the ambitious project was completed in 2003, he remained cautious about how quickly genomics could help advance cancer care. 

“Fifteen years ago,” he recounts, “if somebody asked me, ‘Do you think that there is a future when we would be able to take care of cancer patients appropriately?’ I would have said, ‘Probably not in my lifetime.”

When asked the same question today, Kucherlapati answers differently. “We can actually see the light at the end of the tunnel, where cancer is treated as a chronic illness rather than a terminal one. It’s already beginning to happen,” he says. That’s because we’re on the verge of a paradigm shift in cancer treatment—replacing a one-size-fits-all approach with a new standard of care that treats patients based on the unique characteristics of their disease.

This is the promise of personalised health care, agrees Josh Lauer, the Global Head of Personalised Healthcare Market Development at Roche. Cancer care has reached an inflection point, he says, thanks to “an explosion of information and new tools to identify and make use of the most clinically relevant information.”

“Roche is a committed believer in this space,” Lauer says, “but we have the self-awareness to know that the challenge is too complex to do it all on our own. We must partner with others.”

Dr. Gaurav Singal, a member of the Harvard Medical School faculty and Vice President of Data Strategy at Foundation Medicine, one of Roche’s partners in the space, identifies the three key components necessary to fulfill its promise. “First, physicians need an array of precise, effective therapies. Second, you need high quality diagnostics to define which patients will benefit from which therapies. Finally, you need the infrastructure and informatics to allow the frontline provider to know how to make that match.”

Three of Roche’s most recent partnerships map directly back to these steps.

Breaking down data silos

The first is Foundation Medicine—a molecular insights company founded in 2010. Based in Cambridge, Massachusetts, the company uses a proprietary combination of precise DNA sampling and sophisticated algorithms to provide a comprehensive genomic profile for each patient’s cancer. Then the company culls the world’s most up-to-date published cancer literature to match each patient’s genomic profile with the best-known treatment option or clinical trial.

To date, Foundation Medicine has analyzed more than 200,000 patient samples as it continues to refine its algorithms and increase the precision with which it can identify genomic alterations known to drive cancer growth.

We do more sequencing in a week than all U.S. academic medical centers put together.
Dr. Gaurav Singal

Dr. Singal, puts this scale into context. “We do more sequencing every single week, than every academic medical center in the U.S. put together. The beauty is that we’ve broken down the data silos to centralize the data into a harmonized and anonymized set, and enabled access to commercial and academic researchers worldwide.”

This type of profiling was the difference between life and death for Bryce Olsen, a data engineer at Intel, who suffered from a rare form of advanced metastatic prostate cancer that was unresponsive to the standard treatment. “Genomic profiling was the key that opened up a door to a clinical trial,” says Olsen, who has now been in remission for three years. “Without that profiling, patients like me have no idea what’s driving their disease.”

Unfortunately, for many cancers, there are still no effective cures. Melanie Nallicheri, Foundation Medicine’s Chief Business Officer, believes that the best way to accelerate the research to find those cures, is for biopharma competitors to come together. “I think that Roche and all our research partners understand that our data must be openly shared, and that competition needs to happen on the basis of the merits of the individual therapies, and that the best therapies should win.”

This vision requires, in the words of Nallicheri, “a sea change where everyone with advanced cancer is tested.” “For Foundation Medicine to be successful in advancing cancer research,” says Nallicheri, “comprehensive genomic testing needs to be more ubiquitous.” Roche’s partnership and recent acquisition has helped with these efforts—with Foundation Medicine’s genomic testing now available in more than 60 countries.

Machine analyzing cancer specimens for certain genomic alterations and biomarkers
Machine analyzing cancer specimens for certain genomic alterations and biomarkers

Recovering lost nuggets of wisdom

A high-quality genomic data set is only one piece of the puzzle. Researchers need an equally strong clinical data set to fully understand patient populations, treatments, and outcomes. Traditionally, that has been hard to come by, with most clinical data lost when a patient dies, or limited to clinical trials.

Enter Flatiron Health, another Roche partner, and recent acquisition.

For years, Dr. Amy Abernethy, the first Medical Director of Flatiron Health—has been vexed by this problem. In 2013, she took to the TEDMED stage to tell the story of “Janet,” a patient of hers who died from a melanoma. Dr. Abernethy lamented that once Janet was no longer eligible for a clinical trial in which she had enrolled, all her clinical information—including potential “nuggets of wisdom—was lost. There are tens of thousands of patients like Janet—including all their nuggets of lost wisdom.

Reception desk at Flatiron with company purpose wirtten at the back wall
Flatiron’s products connect community oncologists, academics, hospitals, life science researchers and regulators on a shared technology platform.

Abernethy has spent her career trying to remedy the problem. In 2009, she helped advance the HITECH Act which subsidized the near universal adoption of Electronic Medical Records (EMRs) in the United States. Digitizing patient information was an important first step, but according to Abernethy, “We were taking what was currently a metal filing cabinet with manila folders, and now making it digital. But it was still digital paper, locked in a digital cabinet. None of the data that had been digitized was actually usable for research purposes.”

At first, Flatiron thought a solution could lie solely in natural language processing and artificial intelligence. Abernethy wasn’t convinced. “I was very vocal that you needed clinical experts to clean up the prose and turn it into quantifiable data.” Heeding her advice, Flatiron developed software that allows a highly-trained team of data aggregators to pull and verify all relevant data points out of the medical records in a consistent fashion – using tools like natural language processing to help make the process more efficient.

In 2014, the company delivered its first data set, which included a cohort of 200 patients with melanoma and 2,000 with lung cancer.

Abernethy, a self-described data geek, describes it as a “magical moment.”

“I immediately had a view, a window into cancer care that no one had ever known before,” she says. “What treatments patients were receiving, how long they were living, what biomarker testing had been done.”

Since then, Flatiron has aggregated and normalized the data of over 180,000 patients from OncoEMR, its EMR system used by over 280 community oncology practices—and its interfaces with seven major academic centers. Flatiron uses this de-identified data in partnership with the world’s leading biopharma companies, academic medical centers, regulators and patient advocacy groups for research into everything from supporting label expansions, to understanding patients who are not treated in clinical trials because they meet exclusion criteria.

Flatiron’s Amy Abernethy giving a presentation
Flatiron’s Amy Abernethy has recently been appointed FDA Principal Deputy Commissioner.

As the use cases for its data expands, Flatiron asks itself, “How good is the data set going to be in answering the questions?” They grade the data set within the context of those questions, and then continually improve it with the target of making it, “regulatory grade.”

Recently, Flatiron has embarked on an ambitious data-enhancement project that combines its real-world data set with Foundation Medicine’s genomic one. The resulting longitudinal Clinico-Genomic Database (CGDB) is able to anonymously link patients’ unique biomarkers with their clinical data—including specific treatments and outcomes. “In cancer research, this is considered the Holy Grail,” says Harvard Medical School’s Raju Kucherlapati. “And no one has done this commercially, at scale before.”

“It’s really only been a dream until now,” says Dr. Singal of Foundation Medicine. “I think we’ve created the first step and we’re on the cusp of creating something that is going to transform the field—allowing us to accelerate and expand research, with the aim of developing more targeted therapies, more efficiently.”

A learning health network

Fulfilling the promise of personalised healthcare is also what motivated Jonathan Hirsch, founder of Syapse, another Roche partner. “We started the company back in 2008 with the core realization that oncology was moving towards the use of genomic information alongside traditional clinical data to segment patients into smaller populations and start directing treatment to those patients based on these data—which eventually became known as personalised healthcare.”

Individual pointing to a screen graphic in Syapse’s clinical decision support tool
Syapse managed to create a ‘learning health network’.

In order to fulfill that promise, Hirsch and his team created an oncology clinical decision support platform for integrated health systems which, with increased consolidation, deliver the majority of cancer care in the U.S.. Over the past ten years, Syapse built what has become the U.S.’s dominant oncology precision medicine platform. Their customers include the largest integrated health systems which together represent over 10% of cancer care. This patient population complements that of Flatiron which largely serves independent community practices and clinics.

Like a true platform company, Syapse owns the means of connection and knowledge transfer rather than the means of laboratory testing or EMR management. Syapse is data agnostic, meaning their platform can ingest data from third-party EMRs and where their clients do their genetic testing—whether in house or externally. Like Flatiron and Foundation Medicine, they aggregate and harmonise that data at scale, and then provide oncologists with the tools they need to deliver insights from that data at the point for care, including surfacing outcomes associated with therapies in real-world settings and matching patients with relevant clinical trials.

What differentiates Syapse, according to Hirsch, “is that we’ve managed to create a ‘learning health network’, the first of its kind in the industry, allowing health systems and physicians participating in the network to engage in the sharing of insights and best practices that can be derived from the data.” In its first few years, Syapse shared data only within the walls of each client’s health system. Now, through the Learning Health Network, the company breaks down those barriers and makes its entire de-identified data set is available to all members of the network—provided that participating health systems share their findings with the entire network.

Syapse founder Jonathan Hirsch in an interview
Syapse founder Jonathan Hirsch and his team created an oncology clinical decision support platform for integrated health care systems.

The Syapse Learning Health Network creates a virtuous cycle where the entire network gets smarter as it gets bigger, drawing intelligence from all participants. Syapse’s life science partners, like Roche, also add to its intelligence. Hirsch explains, “Roche has an incredibly long history in conducting clinical trials and studies, and a deep understanding of the rigors of regulatory approvals. That expertise helps our clinical team understand what’s needed to enhance and validate the real-world data, and to scale our insights and offerings to advance patient care.”

It’s this continued ability to scale the size and intelligence of the network that motivates Hirsch. “When I see the improvement of care delivered and see more patients gaining access to better treatments, that’s tremendously motivating,” says Hirsch. As proof of Syapse’s success, he references a series of studies that tracked patients at one of his health system clients over two years. “We found that the patients treated through the precision medicine program had a doubling of progression-free survival, and a 10% reduction in total cost of care per month.1

Leveraging the power of the entire ecosystem

Greg Talbert, Head Personalised Healthcare Partnering
Gregg Talbert hopes with PHC to gain a better understanding of the targeted cancer treatments that are needed next.

Gregg Talbert, Global Head of Personalised Healthcare Partnering for Roche, recognizes that although a lot of progress has been made, there’s still a long way to go. For every patient like Bryce Olson who is responsive to targeted treatments, many cancer patients still aren’t. “But what’s beginning to change,” according to Talbert, “the reason we’re on this personalised healthcare journey in the first place, is that we are continually improving our ability to segment specific patient populations and to gain a better understanding of the targeted treatments that are needed next.”

Partners like Syapse, Foundation Medicine, and Flatiron are making that possible. And Roche believes that their true potential can only be realized, if the partners continue to pursue their independent objectives, and leverage the power of the global cancer research ecosystem.

References

  1. Journal of Oncology Practice: DOI: 10.1200/JOP.2016.011486 Journal of Oncology Practice 13, no. 2 (February 2017) e108-e119. [Internet; cited 2019 Jan 4]. Available from: http://ascopubs.org/doi/full/10.1200/JOP.2016.011486

Tags: Science, Innovation, Personalised Healthcare, Digital health