Mismatch between mortality burden and number of FDA registration trials in highly lethal cancers.

Abstract

2072 Background: Treatment successes in cancer are achieved through new drugs tested in clinical trials. However, drug discovery has been disparate across cancer types for various reasons. We sought to investigate if the number of trials used to support United States Food and Drug Administration (FDA) drug approvals is proportional to the incidence and mortality burden of highly lethal cancers, i.e. those with an expected relative mortality of >5% per Cancer Statistics, 2020 (Siegel et al.). Methods: All FDA labels for 258 antineoplastic cancer drugs approved as of January 2020 were reviewed for citations of registration trials supporting initial approval and additional indications. Trials were identified by matching described characteristics (e.g., patients enrolled, clinical trial NCT codes) to publications indexed on HemOnc.org. Trials were labeled by cancer type studied and type of trial (randomized vs non-randomized). Results: We identified 559 registration trials in total. Results for the six highly lethal cancers are shown in the table. The percent of registration trials was roughly proportional to incidence, but not mortality burden. For example, despite the 22% expected mortality burden of lung cancer, it had a share of only 11% of registration trials whereas breast cancer has an expected 7% mortality burden, with a share of 14% of registration trials. Chronic myeloid leukemia is expected to cause 1,130 deaths in 2020 (0.2%) and has had 20 registration trials (3.6%). The highly lethal cancers had a higher rate of randomized trials supporting approval than other cancers (84% vs 56%, p<0.001 [Chi-square]). Conclusions: While the findings may in part be due to disease biology (e.g., pancreatic ductal adenocarcinoma has proven resistant to many novel therapies), our evaluation highlights a potential mismatch between resources and needs. Randomized trials were more often used to support new drug approvals in highly lethal cancers. These findings will be important in regulatory policy. [Table: see text]