Clinical endpoints in trials of palliative radiotherapy: A systematic meta-research analysis

Up to 50% of radiotherapy courses are delivered in palliative intent for various indications. Despite the large number of treated patients, we know little about the choice of endpoints in trials of palliative radiotherapy. Our primary aim was, therefore, to analyze primary endpoints in trials of palliative radiotherapy.

We conducted a pre-registered (https://doi.org/10.17605/OSF.IO/GMCAF) meta-research analysis searching Pubmed/MEDLINE, EMBASE, CENTRAL, and "ClinicalTrials.gov" for clinical trials of palliative radiotherapy published 1990-2020. Endpoints were categorized in "patient-centered endpoints", including overall survival and patient-reported outcomes, and "tumor-centered endpoints" such as local control. The remainder were "other endpoints" including toxicity or observer-rated symptoms. We applied descriptive statistics to summarize data and logistic regression to assess if year of publication predicted the choice of primary endpoints.

Of 7379 records screened, 292 were eligible. Trials were characterized by small sample sizes and use of external beam radiotherapy for metastases or thoracic primaries. Median patient age was 64 and median ECOG was 1. Only 64.4%(145/225) of published trials clearly stated their primary endpoint. Published trials employed a "patient-centered primary endpoint" in 45.5%(66/145) and a "tumor-centered primary endpoint" in 17.3%(25/145) of the cases. There was no statistically significant trend in time for the use of "patient-centered primary endpoints". Registered ongoing trials used a "patient-centered primary endpoint" in 32.8%(22/67) and a "tumor-centered primary endpoint" in 26.9%(18/67) of the cases.

Although "patient-centered primary endpoints" appear relatively prevalent in published trials of palliative radiotherapy, their use is still suboptimal and appears to be lower in currently ongoing trials.