(V) Regulatory Benefit-Risk Assessment of Oncology Drugs: A Systematic Review of FDA and EMA Approvals

Publication Number: 312

Background: EMA and FDA have policy goals of strengthening benefit-risk (B-R) capabilities but how this has been translating into regulatory practice is unclear. Details regarding how regulators have been structuring and reporting their B-R assessments in recent years can inform product development planning by key stakeholders.

Objectives: To summarize and compare publicly available reports of regulatory B-R assessments for oncology drugs approvals in January 2015 through December 2020.

Methods: A list of oncology drug approvals between 2015 and 2020 were identified through review of FDA and EMA annual reports. Publicly available regulatory review documents were screened to extract information on clinical context, evidence, and additional B-R considerations. Results were stratified by agency, year, orphan designation and review type.

Results: Data were extracted from 233 regulatory reviews including 64 new EMA submissions (39 orphan and 25 non-orphan indications) and 70 new FDA submissions (35 orphan and 35 non-orphan indications). In 2015, there were few reviews (32%) with EMA effects tables or FDA B-R tables, compared with ≥93% each year starting in 2017. Among new submissions, there were notable differences between agencies, with FDA approvals more frequently based on single-arm trial evidence (orphan [54% vs 40%] and non-orphan [40% vs 21%] indications) and conducted as accelerated/priority reviews (orphan [83% vs 60%] and non-orphan [71% vs 13%] indications) than EMA, respectively. EMA more commonly based decisions on overall survival data (orphan [84% vs 40%] and non-orphan [54% vs 37%] indications) than FDA. For EMA and FDA, respectively, few assessments included quality of life data (9.4% vs 0%), patient experience data (7.8% vs 8.6%), or real-world data (1% vs 0%). No assessments included results of quantitative B-R assessments. EMA more often granted marketing authorization without conditions (orphan [60% vs 9%] and non-orphan [69% vs 51%] indications) than FDA, with a requirement to conduct post-authorization studies more often required by FDA (64%, N=45) than EMA (41%, N=26). Data were also extracted from 99 EMA label extension submissions; with no reviews of label extensions reported by FDA (which is common practice) for comparison.

Conclusions: Differences in B-R assessments by regulators can inform early planning by sponsors to address multiple health authority needs. Notably, while both agencies have public policy targets of extending their use of patient experience data and real-world data, these are still being rarely applied in regulatory B-R assessment of oncology drugs.