The healthcare ecosystem is experiencing an acceleration in the uptake of the precision medicine paradigm due to many technological advances, including the utilization of big data. Across therapeutic areas, new technologies are allowing researchers to use big data to investigate ways to better diagnose disease, predict which patients will benefit from which therapies, and feed clinical outcomes back into research for the next therapeutic breakthrough. The challenge for the scientific community is to determine how to best synthesize and utilize this data throughout the clinical process. Industry leaders gathered Wednesday afternoon at the 2018 BIO International Convention to discuss these issues at the premier session of the Precision Medicine and Diagnostics Track entitled, “Heralding in a New Era of Precision Medicine”.
The distinguished panel was moderated by Dr. Thomas J. Lynch, MD, Executive Vice President and Chief Scientific Officer at Bristol-Myers Squibb. Panelists included: Isaac Kohane, MD, PhD, Marion V. Nelson Professor & Chair of the Department of Biomedical Informatics at Harvard Medical School, Neal Meropol, MD, Vice President of Research Oncology at Flatiron Health, Melanie Nallicheri, Chief Business Officer & Head of Biopharma at Foundation Medicine, and Scott Solomon, MD, The Edward D. Frohlich Distinguished Chair Professor of Medicine at Harvard Medical School.
The panel discussed how the scientific community is using big data to identify predictive biomarkers as well as how diagnostics and next generation sequencing (NGS) are speeding research. There is real opportunity to analyze large datasets—claims data, health records and genomics—and machine learning to make narrow clinical trials that are more successful. We are now able to leverage genomic information at all stages of discovery and throughout the development process. Nallicheri discussed how we need both scale and depth of data to continue the identification of biomarkers to allow us to make pinpoint diagnoses and determine whether a patient will be able to respond to therapy. Real world data offers a chance to link clinical data to genomic data longitudinally to think about the next therapy; however, experts and regulators have been cautiously optimistic about the efficacy of real world data.
Artificial intelligence (AI) is accelerating the use of big data in drug development and patient care. Meropol explained how we should be able to unleash the power of the data but the problem is that electronic health records (EHR) data is unstructured—free text that is often too difficult to image. Since AI imaging is insufficient in these cases, humans are still required in the process. To understand the richness of this data and to make it scalable, we need to reduce the need for humans to interpret the text in the clinical records.
The ultimate promise of AI lies in the prospects of prevention prior to any symptoms or diagnosis. Ideally, after taking a simple panel of tests at initial screening, the AI would eventually be able to read the patient’s EHR to predict the totality of a patient’s health risks and the likelihood of responding to a particular therapy. Many AI techniques will replace current techniques that read imaging data, and the health applications could be very disruptive—in a good way—for the healthcare system.
The new data privacy laws enacted by the European Medicines Agency (EMA) could have a chilling impact on big data. The restrictions could both deter patient participation and scientific uses of the data. Before jumping in with both feet on the big data bandwagon, the public has expressed that they want to know that their data is safe, so the industry must double down on data safety and privacy. Some patients may simply be unwilling to participate— for instance, not wanting to ‘know’ that they may have an incurable disease down the road. Much of the worries around data privacy are driven by negative stories in the media. To solve this, it may just be a matter of educating the public to understand that many database providers now de-identify the datasets and anonymize profiles, and that disease risks can be disclosed upon request rather than being automatically provided to the patient.
The new era of precision medicine is becoming a convergence of technological advances of many disciplines, including biology, computer science, statistics, and engineering. The many trends in precision medicine—including the challenges and opportunities for big data—will help the industry deliver products with better safety and efficacy across a wide range of disease areas. The evolving technologies, including diagnostics and AI, will help us identify the large majority of patients with undiagnosed health problems that could benefit from these technological opportunities.