By Deborah Borfitz
March 13, 2019 | Medical knowledge will soon double about every three months, while continuing to grow in complexity. Oncologists are challenged to keep up with the latest research on hundreds of different cancer types as well as evidence on actionable genetic mutations that make cancer care increasingly personalized.
Delivering the best possible care also means multidisciplinary alignment to decide on the most suitable diagnostic tools to apply and settle on a treatment path with patients and their care team, says Okan Ekinci, chief medical officer for Roche Diagnostics Information Solutions at F. Hoffmann La-Roche Ltd in Basel, Switzerland.
Exploiting value from data—more than 2,300 exabytes of it come next year—is the path to personalized healthcare where patients receive treatment based on the molecular profile of the tumor or other unique characteristics, Ekinci says. While in the past cancer drugs were used to treat all patients and later companion diagnostics for subgroups, today’s improved data resolution allows for better predictions about which patients may benefit from which drugs.
Expanding personalized healthcare will require more longitudinal and structured patient data collected in the disease continuum—from diagnosis and therapy to the ultimate outcomes—as well as greater collaboration across care settings, says Ekinci. He foresees a connected health system that includes not only physicians and hospitals but also nonprofits, medical associations, academic institutions, and regulatory bodies.
The Roche cloud-based NAVIFY Tumor Board decision support software, launched in 2017 and now used in eight countries in Europe and North America, integrates and displays relevant aggregated data into a holistic patient dashboard to aid oncology care teams in reaching clinical care decisions. As of last fall, the NAVIFY portfolio also functions as an ecosystem of clinical decision support apps, the first two of which match patients to clinical trials and find relevant scientific publications. Partners in the effort include global players such as GE Healthcare and Accenture.
The NAVIFY platform currently offers integration to the electronic medical record and ingests data from radiology, pathology, and molecular testing, says Ekinci. In the future, it will also pull in more information directly from images and other primary data sources.
For the last half century in oncology, survival has doubled and the options for treating and monitoring patients has soared, says Ekinci, making survivor management burdensome for hospitals. “We now know that every patient is different and that their tumor has a molecular fingerprint or multiple ones, which keep changing.” Physicians also need to decide whether to treat patients with a single drug or a drug combination, currently with insufficient guidance as to which is best because “data does not feed into a global system to help others with a similar fingerprint constellation.”
With so little data being shared between institutions currently, it is hard for treating oncologists to know which genetic mutations are actionable and sensitive to which drugs, says Ekinci. A treasure trove of data being generated by wearable devices and patient-reported outcomes surveys also goes untapped as a driver of therapeutic choices, improved patient satisfaction, and lower hospitalization rates.
All of that will be enabled by clinical decision support software that puts advanced analytics to work atop large, aggregated data sets, Ekinci says. But it needs to be accessible at the point of care and physicians, already “feeling overwhelmed” by technology, must see the value for themselves and their patients. One benefit for patients comes with faster evidence generation for regulatory approval of drugs, he adds. Clinical trials increasingly incorporate real-world evidence, which is also used to better detect and avoid toxic reactions.
From Modeling to Digital Twins
As a decision support tool, the NAVIFY Tumor Board cuts preparation time in half for the lead oncologist because most of the data needed for therapy decisions is automatically extracted from the electronic health record, says Ekinci. The tool could perhaps be used in a similar fashion by hospitals to improve the diagnosis and management of other diseases for which therapy can be increasingly personalized—e.g., depression, which is relatively common among cancer patients—or to address any other challenge along patients’ care journey. NAVIFY’s capabilities around interfacing to digital pathology and molecular profiling with structured data is expanding, Ekinci adds.
Roche and GE Healthcare already work on a digital collaboration platform that helps with the early detection of patient deterioration due to sepsis—a difficult-to-diagnose condition associated with high mortality rates—so treatment can begin sooner.
Notably, Roche is one of nine companies piloting the FDA’s Pre-Cert Program intent on adjusting the approval process for fast-cycle digital health solutions like NAVIFY based not on their merits individually but how products get designed, developed, validated, and implemented in clinical practice, says Ekinci. NAVIFY also aligns with data protection and privacy regulations in both the U.S. (HIPAA) and the European Union (GDPR).
Bringing together systems biology and more data on patients’ condition might one day allow a simulation layer to be added to model their health status and predict the effects of various treatment options, including side effects. Much like how knowledge and data analytics on materials properties allow a mechanical engineer to forecast when a jet engine turbine blade may break, Ekinci says, digital twin technology for modeling individuals might be used to “work out the options and, to a certain extent, predict the outcome.”
Bringing simulation to clinical routine maturity will take time and require advanced analytics applied to large cohorts of longitudinal patient data, Ekinci says, but “the end game is more confident decisions, not replacing healthcare professionals.”