In an evolving regulatory era, speakers highlighted the role of real-world endpoints used for evaluating cancer drug therapy outcomes at a forum sponsored by the Friends of Cancer Research (FOCR) and Alexandria Real Estate Equities, Inc., in Washington, D.C. Real-world evidence is defined as the clinical evidence derived from data on the uses and potential benefits and/or risks of a medical product outside of a traditional clinical trial.
The ultimate goal of using real-world evidence to assess outcomes is to help facilitate faster drug approvals, thus bringing safe and effective therapies to cancer patients sooner. The case for real-world evidence is strong: many traditional clinical trials take too long and are too expensive; such evidence captures a broader array of data on a more representative population, since only a small, selected fraction of the population participates in clinical trials; real-world evidence can be used as supplemental data with clinical trial data; it can send early signals of drug efficacy or safety issues; and real-world data can capture patients' quality-of-life experiences that may not be reported in clinical trial data.
But speakers at the 8th Annual Blueprint for Breakthrough Forum emphasized that real-world evidence is only as good as the quality of the data on which it is based. Discussion questions at the forum included the following: What opportunities or incentives exist to help improve the format, quality, and validity of real-world evidence? What opportunities exist for FDA decision-making to be supported by real-world evidence? What opportunities exist to expand to other endpoints such as patient-reported outcomes and patient-generated health data?
Noting that a clinical trial does not necessarily reflect the breadth of data that can be collected in a more diverse real-world setting, FOCR has been working to create a standard dataset curation process and validated framework so data can be made operational for real-world evidence. At the forum, organization released a white paper of preliminary findings from a collaborative pilot project on advanced non-small cell lung cancer (NSCLC) showing the benefits and limitations of real-world evidence, which was derived from electronic health records (EHRs) and medical/pharmacy claims. EHRs capture data from individual patient encounters, while claims data provide a view of a patient's history over a period of time.
“It's already here; FDA is already using it,” stated keynote conversation speaker, Acting FDA Commissioner, and former Director of NCI Norman E. “Ned” Sharpless, MD, said of real-world evidence. “The FDA is eager to use it.”
But, he noted this is an area where there is much room for improvement in terms of data quality. As an example of how the FDA is currently using real-world evidence, Sharpless cited the drug palbociclib, which was approved in April 2019 in combination with hormone therapy for male patients with ER+, HER2- advanced or metastatic breast cancer. Palbociclib had already been approved for female breast cancer patients; male breast cancer patients comprise less than 1 percent of breast cancer cases. In approving palbociclib for male breast cancer, the FDA concluded that the safety profile in male patients was consistent with that in female breast cancer patients.
“Biological understanding of the problem solves everything,” said Sharpless, noting that increased understanding of science in oncology has led to smaller, nimble trials and faster drug approvals. He said that his previous job at NCI helped him in this regard when he was appointed acting FDA commissioner. Sharpless noted the increased scientific understanding of cancer has led to a new era in clinical oncology. “It used to be that you couldn't use the word ‘cure,’” he said. But today, “I think it's a really exciting time in oncology,” and some of the new therapies, such as CAR T-cell therapy, are curative.
The FOCR's ongoing new pilot project on use of real-world evidence for advanced NSCLC, known as RWE Pilot Project 2.0, has yielded some encouraging preliminary findings, said Jeff Allen, PhD, President and CEO of FOCR. Perhaps the most encouraging is the consistency and harmonization of datasets from different sources. In the words of the white paper released at the forum, this new pilot study, the second sponsored by FOCR, “leveraged parallel analyses from common data elements across multiple data sources to assess three frontline treatment approaches in real-world patients” with advanced NSCLC. It is a retrospective, observational analysis derived from EHR and claims data. The datasets include all HIPAA-compliant, de-identified data available for eligible individuals up to a cutoff date of March 31, 2018.
The pilot project includes datasets from multiple databases provided by different, collaborating sources. Data sources included Aetion, ASCO CancerLinQ/Concerto HealthAI, COTA, Flatiron Health, IQVIA, Kaiser Permanente, Mayo Clinic, McKesson, SEER-Medicare, Syapse and Tempus. Allen said it was remarkable that different datasets could come together with such alignment in this collaborative project. Agreeing was Amy Abernethy, MD, PhD, Principal Deputy Commissioner at the FDA. In a keynote address, Abernethy called this alignment “pretty remarkable.” She said the pilot project highlights for her “how important it is to do the hard work.”
The white paper noted: “It is possible to coordinate the efforts across numerous real-world oncology data organizations to reach high-level alignment on important data elements and definitions for real-world endpoints in the context of a focused research question.” The depth of data varied across data providers and each dataset had distinct characteristics; the data demonstrated the ability to provide insight into recent trends in clinical cancer care.
But real-world data can suffer from quality problems, such as missing elements. One salient fact emerging from the new pilot study is that verifying and determining the date of death may prove challenging. As the white paper stated, “Although discharge status and some diagnosis codes may be a source of mortality information, it is often incomplete.”
While there is a tsunami of patient data that is becoming available, obtaining death data in a timely fashion is a problem, confirmed Lawrence (Larry) Kushi, ScD, Director of Scientific Policy in the Division of Research for Kaiser Permanente Northern California. He noted that differing data sources may make it hard to follow a patient through the treatment course. For example, a patient may change insurers and may be treated in different settings in different parts of the country, as happens with “snowbirds” who spend the winter in Florida.
Agreeing with Sharpless on the need for better data quality for real-world evidence was Sean Khozin, MD, MPH, Associate Director of the Oncology Center of Excellence at the FDA and Founding Director of FDA's Information Exchange and Data Transformation. Khozin said data quality issues include missing information. “At what threshold is the data quality good enough?” he asked.
Asked by Oncology Times if real-world evidence could ever replace clinical trial data, Khozin said that in the immediate future such evidence could be useful in expanding the drug label; drug repurposing; and moving a drug from off-label use to labeled use. “If it rises to the threshold of evidence, we can put it on the label,” he said. Like Sharpless, he cited palbociclib as an example of a case where real-world evidence was used by the FDA to expand the indication for an approved drug.
Khozin said there is a need to build pathways for routine data collection, which could boost the potential value of real-world evidence. He sees the distinction between retrospective and prospective data narrowing as real-world evidence is used more frequently, and cited ASCO's TAPUR (Targeted Agent and Profiling Utilization Registry) as an example of that trend. TAPUR is a non-randomized, open-label clinical trial that aims to describe the performance (both safety and efficacy) of FDA-approved targeted anti-cancer drugs prescribed for patients with advanced cancer that has a potentially actionable genomic alteration. The study aims to define potential signals of drug activity.
From the patient's point of view, real-world evidence can be valuable in making treatment decisions, said breast cancer survivor Christie Mangir, MS, a Project Manager for Provider Education at the Association of Community Cancer Centers and co-founder of a digital health startup focused on patient-reported outcomes. Mangir said she was diagnosed at age 30 with an aggressive breast cancer and received two opinions on two different treatments. She noted that the availability of real-world evidence “could have helped me make this decision.” Now 3 years post-treatment, she said that when she was diagnosed she could find no clinical trial data on patient quality of life and patient experiences for her two treatment experiences. During her treatment, she said she had months when she was “basically bedridden” and had multiple visits to the emergency department.
FOCR's research focus on real-world data is an ongoing initiative. The organization published the results from its first collaborative real-world evidence pilot project 1.0 in JCO Clinical Cancer Informatics; this pilot project demonstrated that different datasets could be used to extract several real-world endpoints in a relatively consistent manner (2019; doi: 10.1200/CCI.18.00155).
“RWE is an important tool for continued learning about treatment outcomes over time,” the FOCR's Allen commented on this first pilot project. “This partnership explores the potential use of different endpoints derived from electronic health data, and demonstrates how different data sources can reproducibly provide important information on patient populations often excluded from clinical trials.”
Peggy Eastman is a contributing writer.