Real-world electronic health record data on tumor progression appears to be a good surrogate for overall survival, a finding that could help to improve estimates of benefit for complex cancer therapy regimens for use in both clinical trials and treatment planning, investigators say.
A study of electronic health record (EHR) data on more than 30,000 patients with non–small cell lung cancer (NSCLC) showed that real-world data on progression-free survival (PFS) showed a high degree of correlation with overall survival (OS), reported Sandra D. Griffith, PhD, of Flatiron Health, New York, and colleagues.
“As survival rates have improved and more than one therapy line has become typical, OS has become an inadequate metric for ascertaining the benefit of some cancer treatment interventions. Intermediate endpoints are essential for evaluating treatment benefits in real-world contexts. This study presents a scalable, feasible, and replicable approach to yield an EHR-generated progression variable ready for incorporation into large, contemporary real-world analyses,” they wrote in JCO Clinical Cancer Informatics.
The investigators studied retrospectively whether it would be feasible on a large scale to extract clinically relevant endpoints from an EHR-derived database that relies on technology-assisted abstraction of data from patient charts.
Their cohort included 30,276 patients diagnosed with advanced NSCLC from January 2011 through February 2018 who had two or more visits documented on EHR and who had been started on one or more lines of systemic therapy. Of this group, 16,606 had one or more documented tumor progression events, and 11,366 either died, discontinued therapy, or started on a new therapy.
The investigators found that among 20,000 evaluable patients, correlation of real-word PFS with OS was moderately high (Spearman’s P = .076), although for 11,902 patients with a progression event who died the correlation of EHR data with actual OS was somewhat lower (Spearman’s P = .069).
The strength of the correlations was verified in an duplicate random sample reviewed by independent abstractors.
The median time to abstract data on disease progression from individual electronic health records was 18 minutes.
“More work is needed to document how real-world progression under different treatments relates to treatment effects observed in clinical trials; understand the impact of different imaging and assessment cadences—for example, informative censoring; and clarify the relationship between real-world progression and other endpoints, such as real-world tumor response. Similar work is also needed in other tumor types,” Dr. Griffiths and colleagues wrote.
The study was supported by Flatiron Health, a subsidiary of Roche. Dr. Griffith and multiple co-authors are employees and stockholders of Flatiron Health and/or Roche.
SOURCE: Griffith SD et al. JCO Clinical Informatics. 2019 Aug 12. doi: 10.1200/CCI.19.00013.