In my previous blog, I described my methodology and reported on my analysis of the outcomes and endpoints of the top five Asian cancers by sponsor type in a subset of phase II-III completed and terminated trials with one of four efficacy primary endpoints. Here, in the second part of this blog series, I will report on my analysis of outcomes and endpoints by disease.
In this subset of 440 trials, slightly fewer trials overall had positive primary outcomes than negative at 49% vs. 51%, respectively. However, the ratio of positive to negative primary outcomes differed in each of these diseases. Outcomes were equal in Breast, more negative in Gastric (52%), Lung (54%) and Liver (58%) and more positive in Colorectal (57%).
The distribution of primary endpoints differed among trials in these diseases as follows. For Lung, PFS and OS were almost equally chosen as the primary endpoint, while for Colorectal, the choice of DFS, PFS and OS was roughly equivalent. OS was the clear favorite in Gastric and Liver, and PFS was the clear favorite in Breast. TTP was rarely chosen as the primary endpoint for any trials in these diseases.
In addition, the association of primary endpoints with either positive or negative primary outcomes differed in these diseases. DFS was associated with slightly more positive primary outcomes in Colorectal, clearly more positive in Lung but with more negative in Gastric and Liver. Likewise, PFS was associated with clearly more positive primary outcomes in Lung but with slightly more negative in Liver. OS was associated with almost equally with positive and negative primary outcomes in all the diseases except Lung, with clearly more negative. TTP was only associated with positive outcomes in Lung. All four primary endpoints were associated with termination due to negative reasons, but only PFS and OS were associated with termination because of an early positive outcome.
Finally, the use of PGX biomarkers to select or stratify patients differed by disease. Breast cancer trials had the highest incorporation with 66%. No incorporation of PGX biomarkers was found in Liver cancer trials; minimal incorporation in Gastric cancer trials and 13-15% incorporation in Colorectal and Lung cancer trials. Yet while the use of PGX biomarkers for selection or stratification was most widespread in Breast trials with all of the four primary endpoints, this use of PGX biomarkers was associated most clearly with positive outcomes in Lung cancer trials with PFS as the primary endpoint.
For my final blog in this series, I will continue my analysis of the outcomes, endpoints and the use of PGX biomarkers for selection or stratification in this set of later-stage oncology trials by start year and then wrap-up the series with a summary of the trends and correlations.