Systematic review of time to subsequent therapy as a candidate surrogate endpoint in advanced solid tumors

Plain language summaryIn a recent study, researchers investigated how we can measure the effectiveness of cancer drugs. They focused on a specific measure called 'time to next therapy', which is the duration between two treatments patients receive. By analyzing the relationship between time to next therapy and disease progression, they discovered a strong correlation. This suggests that in the future, time to next therapy could potentially help to measure how well a cancer treatment works. However, when it comes to predicting patient survival, the relationship was not as strong. This implies that time to next therapy is not a reliable indicator of patient survival. To fully understand whether time to next therapy can effectively measure the effectiveness of anticancer drugs, further research is necessary. Tweetable abstractA systematic review showed that time to subsequent therapy correlates with progression-free survival, and therefore may be useful to measure the treatment effect of anticancer drugs in advanced solid tumors. Aim: Time to subsequent therapy (TST) is an end point that may complement progression-free survival (PFS) and overall survival (OS) in determining the treatment effect of anticancer drugs and may be a potential surrogate for PFS and OS. We systematically reviewed the correlation between TST and both PFS and OS in published phase 2/3 studies in advanced solid tumors. Materials & methods: Trial-level correlational analyses were performed for TST versus PFS (by investigator and/or central review) and TST versus OS. Results: Of 21 included studies, nine (43%) used 'time to first subsequent therapy or death' (TFST) as the TST end point; 11 (57%) used different definitions ('other TST end points'). There was a strong correlation between TFST and PFS by investigator (medians: R-2 = 0.88; hazard ratio [HR]: R-2 = 0.91) and TFST versus PFS by central review (medians: R-2 = 0.86; HRs: R-2 = 0.84). For TFST versus OS there was medium/poor correlation for medians (R-2 = 0.64) and HRs (R-2 = 0.02). Conclusion: TFST strongly correlates with PFS, but not with OS.