Pfizer Turns to Health Social Network to Find Patients With Rare Lung Cancer Genetic Variants

NEW YORK – In the pursuit of new genetic targets for lung cancer treatments, researchers at Pfizer have teamed up with Boston Children's Hospital, genetic research company Citizen Genetics, and Inspire, a social network for patients and caregivers, to recruit difficult-to-find patients who may have a genetic resilience to the disease.

Within this study, which started in February, researchers want to find patients with a family history of autoimmune diseases who have either recovered quickly from lung cancer or not gotten sick at all despite having risk factors. Researchers are particularly interested in whether these patients have genetic variants that may have both contributed to their lung cancer resiliency and may be driving autoimmune-related symptoms.

The lung cancer project is the first study within the Patient Forward Access to Clinical and Technological Research (PFACTR) initiative, within which Pfizer is working with diverse collaborators to identify patients with rare genetic variants for research. Specifically, for the lung cancer study, the drug giant is hoping to find a shared genetic variant among study participants that could lead to a new genetically targeted treatment for either the autoimmune conditions or lung cancer.

The challenge in finding these patients is that they aren't seriously ill and may not normally be considered for genetic analysis. Still, understanding these "resilient" patients could help researchers identify new genetic alterations to target when developing cancer treatments. 

That's where Inspire comes in. Stefan McDonough, executive director of genetics at Pfizer, described the Inspire platform as "an extraordinary resource" because patients are already appropriately consented and are eager to share their medical information to advance treatments. "I would be wearing out multiple sets of shoes going from investigator to investigator, clinician to clinician" to collect this sort of patient data, McDonough said.

Inspire provides an online platform for patients with cancer, autoimmune, and rare diseases, as well as their caregivers, to share their experiences. With more than 2 million members and 15 million posts, the platform is a gold mine of patient data.

Users can create a profile on the Inspire platform, which may include information about their diagnoses, treatments, and test results. Based on these profiles, patients and caregivers find others with similar experiences.

Researchers in industry and academia can also partner with Inspire to search these profiles to collect data from users who have consented to being contacted about medical research. Within the lung cancer study, an institutional review board has ensured that the data collection methods have met ethical standards.

By mining the data within the Inspire platform, the researchers were able to find more than 100 participants who fit the specific lung cancer and autoimmune disease history profile in just a few weeks.

Experts within Inspire are doing the initial screening and outreach to identify users who meet the lung cancer study criteria. Specifically, researchers are looking for two types of participants: those who have been diagnosed with atopic dermatitis, eczema, or lung cancer, and those who are related to people diagnosed with those conditions.

Individuals who agree to partake in the study are referred to Citizen Genetics for genetic sequencing.  Researchers at Citizen Genetics and Pfizer will then compare the participants' genetic profiles against the sequence data from the Human Genome Project to identify potential gene variants to study further in the lab.

The Manton Center for Orphan Disease Research at Boston Children's Hospital is responsible for the clinical administration of the study, and participants will also receive genetic counseling from the Manton Center.

"We have been able, with our collaborators, to find people who do have very interesting mixes of oncology phenotypes and autoimmune phenotypes," said McDonough. "Because we are looking for families, this gives us some hope that the conditions may be genetic."

In identifying the right study participants based on Inspire user profiles, the experts at the social network have to transform unstructured data in users' health profiles and forum posts into structured data that can be queried for the specific characteristics of interest.

Researchers from Inspire and Stanford University demonstrated their ability to cull structured clinical data from the Inspire platform in a 2018 JAMA Oncology paper. They built a natural language processing engine to search for reports of adverse reactions to an EGFR inhibitor and two checkpoint inhibitors on the platform's forums, said Richard Tsai, senior VP of marketing at Inspire.

Using this method, the researchers said they would have been able to identify treatment-related adverse events discussed on Inspire's social network an average of seven months before these adverse events began appearing in clinical reports. The study also identified an adverse drug reaction that had previously gone unrecognized by clinical researchers and the medical community, but which patients had discussed on the Inspire platform for more than 11 years.

"We had confidence in structuring the data for a medical entity and we have this machinery of providing value to patients," said Tsai, highlighting that the platform is also "able to cluster people in different ways longitudinally across time."

For this project, Pfizer chose lung cancer because it is a commonly occurring tumor for which there is a need for better treatments. This allows for a large population of affected patients within which researchers can hunt for the individuals with the rare genetic and clinical features they are interested in studying.

But ultimately, as the research progresses within the project, the immune-specific biological questions may lead Pfizer to broaden the study criteria beyond lung cancer. "The most interesting patients are those with many autoimmune diseases regardless of tumor type, and we may look at this as well," McDonough said. "We are really looking for immune mechanisms, not tumor-intrinsic mechanisms, since there’s no doubt that the immune system can be manipulated to produce therapeutic benefit."

The "holy grail," McDonough said, is a patient or an entire family who has some combination of lung cancer and an autoimmune disease like rheumatoid arthritis, atopic dermatitis, or eczema. As an example, he presented a hypothetical example of a family with five siblings who all developed lung cancer, but three also had an autoimmune disease that spontaneously resolved. The experience of those three siblings would suggest they may have a genetic resistance to lung cancer, he explained.

Pfizer, the company that developed crizotinib (Xalkori), one of the earliest examples of a lung cancer treatment for a genetically defined population, is hoping is to apply the discoveries from this study to develop new targeted lung cancer treatments.

The learnings from this study will be "hypothesis-generating," according to McDonough. The drugmaker is hoping to find variants in genes with large effect sizes. Pfizer scientists will then try to pinpoint the biological effects of those variants in various disease models and engineer those effects within the drug development process, he explained.  

While the researchers are not certain what they will find within the study, a potential treatment that could emerge from this research could work similarly to the osteoporosis treatment romosozumab (Amgen's Evenity), which mimics inherited variants in the SOST gene associated with stronger bones, McDonough explained. If a druggable genetic variant associated with lung cancer is found in this study, Pfizer's R&D division would try to develop a drug that phenocopied that genetic effect. 

But a lot of research and development work needs to happen before the findings from this study can be translated into a treatment, and ultimately, it may never happen given the difficulties in the drug development process, McDonough acknowledged. Regardless, the researchers will publish their results for others to pursue in new directions.

However, McDonough also recognized that this type of research in a genetically distinct patient population, facilitated by the Inspire network and advances in genetics, may not have been possible a few years ago.

"It takes patients, it takes Inspire, it takes really crack academic investigators, and a pharma company behind it," McDonough said. "We're part of the ecosystem, and we are seeing all these different areas that in my memory used to be completely separate really move together for the good of everyone."