Friends of Cancer Research released data Tuesday that support the use of real-world evidence (RWE) outcomes — including real-world progression-free survival (rwPFS) — as surrogate endpoints for overall survival to evaluate a treatment’s long-term effects.
In the pilot program, a working group led by the non-profit Friends evaluated electronic health records (EHR) and claims data from non-small cell lung cancer patients receiving anti-PD-L1 mAb Tecentriq atezolizumab and anti-PD-1 mAbs Opdivo nivolumab and Keytruda pembrolizumab in the postmarket setting. Friends measured the correlation between several real-world endpoints and OS using Spearman’s rank correlation co-efficient, which ranges from 1 when the variables are identical to -1 when they are fully opposed.
Data from six RWE data sets show positive correlation between real-world OS (rwOS) and real-world time to next treatment (rwTTNT), with correlation co-efficients ranging from 0.36 to 0.7 (n=1,779), as well as between rwOS and real-world time to treatment discontinuation (rwTTD), with correlation co-efficients ranging from 0.62 to 0.88 (n=6,709).
Two data sets were evaluable for rwPFS and both showed a positive correlation between that endpoint and rwOS, with co-efficients of 0.75 (n=4,337) and 0.84 (n=142). Furthermore, real-world time to progression (rwTTP) was also positively correlated to rwOS in two evaluable data sets, with co-efficients of 0.6 (n=2,286) and 0.63 (n=55).
Friends presented the data at its The Future Use of Real-World Evidence meeting Tuesday in Washington and also published the findings in a white paper.
The paper’s authors wrote, “Further validation is required to determine whether these endpoints are reliable surrogates for OS outside of a traditional clinical trial and whether they can support regulatory and payer decision-making.”
Friends noted that the range of OS observed in clinical trials that supported approval of checkpoint inhibitors is “highly similar” to the range observed in real-world populations, suggesting that, at least in this case, the clinical trial data are “generalizable to the broader population.”
Results also suggested treatment effects as evaluated by real-world endpoints remained consistent for several demographics, such as age and sex. Differences emerged for disease stage or line of treatment.
Separately, FDA proposed a $100 million medical data system under President Donald Trump’s FY19 budget to incorporate new data from EHRs and other sources to improve real-world evaluation of approved products, Commissioner Scott Gottlieb wrote in a blog post Tuesday.
“Relying on health claims information was the state of the art at the time that we built these systems. Now we have the capacity to use clinical data derived from electronic health records to develop faster reporting on the performance of medical products in real world medical settings,” he said.
The new system would rely on medical data within de-identified EHRs across about 10 million lives.
The 21st Century Cures Act requires the agency to develop guidance for RWE as a requirement for regulatory approval in new indications of approved drugs or post-marketing trials (see “Realities of Real-World Evidence”).
The pilot program working group included Cota Inc. (New York, N.Y.); Flatiron Health (New York, N.Y.), IQVia Holdings Inc. (NYSE:IQV) and OptumLabs (Cambridge, Mass.) and its co-founder Mayo Clinic. Data also came from the Cancer Research Network and the Patient-Centered Clinical Research Network (PCORnet) of the Patient-Center Outcomes Research Institute (PCORI).