Header: This week, JAMA offers a large meta-analysis of different endpoints in cancer screening trials, and Bach offers a conservative op ed with regard to endpoints other than mortality itself.
##
Find Feng et al here.
To oversimplify, imagine this is a classic surrogate biomarker study. You have the best biomarker, [say] mortality, at 5 years, and you have an early biomarker, incidence-of-late-stage, at 1 year, and you ask, if you made decisions on "late-stage," at 1 year, how well does that predict if you had waited for "mortality" at 5 years? Or again, it's the same logic as asking if you have only PFS today, how well does that predict OS later.
Find Bach here.
The FENG systematic review and meta-analysis explores the use of late-stage cancer incidence as an alternative endpoint to cancer-specific mortality in randomized clinical trials for cancer screening. Analyzing 41 trials, it finds mixed results: while incidence of late-stage cancer correlates highly with mortality for ovarian and lung cancers, this is less so for breast, colorectal, and prostate cancers. This suggests that while late-stage cancer could be a suitable endpoint for some cancer types, it is not universally applicable across all cancers, highlighting the need for cancer-type specific trial endpoints.
The BACH opinion piece critically evaluates the use of late-stage cancer incidence as an alternative endpoint to cancer-specific mortality in screening trials, as discussed in the FENG study. BACH emphasizes the inconsistencies in correlation between late-stage cancer and mortality across different cancers, highlighting risks of misinterpreting screening benefits. He argues that despite the shorter trial durations offered by using late-stage cancer as an endpoint, cancer-specific mortality remains the more reliable measure for assessing the true efficacy of cancer screening tests.
[AI, Chat GPT 4]
