Many negative trials presented with ‘not-negative conclusions’ at oncology meetings
Increased attention should be paid to the conclusions of formally negative trials discussed during oral presentations at oncology meetings, as many convey a not-negative message, according to a research letter published in JAMA Oncology.
“During [the 2019] ASCO Annual Meeting, several trials presented were considered formally negative according to their primary analysis, but they were presented with ambiguous and not-negative conclusions,” Massimo Di Maio, MD, researcher in the department of oncology at University of Turin in Torino, Italy, told Healio. “My colleagues and I later decided to perform a systemic review to better understand the real magnitude of this phenomenon. We started with ASCO 2019, but later found that the proportion of those clinical trials was not negligible. We then extended the analysis to the [European Society for Medical Oncology] meeting, covering 3 years in total.”
Di Maio and colleagues sought to describe the frequency and type of not-negative conclusions used by presenters to discuss the results of formally negative trials.
The researchers reviewed the oral presentations of 208 randomized phase 3 trials delivered between 2017 and 2019 during ASCO and ESMO annual meetings and classified trials as not negative or negative in terms of formal results and conclusions. They considered conclusions to be not negative when presenters considered the possibility of using the experimental treatment in that setting with no clear conclusions about the study negativity.
Researchers categorized formally negative trials according to the reasons for not-negative conclusions, including:
- numerically better outcome in the experimental group despite a nonsignificant P value;
- emphasis on positive subgroup or subgroups;
- emphasis on positive secondary endpoint or endpoints; and
- noninferiority interpretation of a negative superiority trial.
“Our aim was to be provocative,” Di Maio said. “We do not want to be considered fundamentalists of clinical research methodology, because we perfectly know that even in a formally negative trial there could be many potentially important findings and many ideas for further research. Oral presentations are not peer-reviewed, but we believe that the messages given in the conclusions of oral presentation are important and need to be balanced, considering the great diffusion of virtual meetings and the large coverage of important meetings like ASCO and ESMO on social media.”
Results showed 29% of the 91 formally negative studies had a not-negative conclusion. The proportion of negative studies with not-negative conclusions declined from 22% in 2017 to 13% in 2018 but increased to 47% in 2019. In addition, 30% of nonprofit studies and 26% of for-profit studies had not-negative conclusions.
Of the 26 studies with a negative primary analysis, study authors emphasized a numerically better outcome in the experimental group in half of cases, a positive result in one or more subgroups in 46% of cases, and a positive result in one or more secondary endpoints in 38% of cases.
Researchers noted that the word “negative” should be used explicitly by presenters when the study’s primary endpoint has not been met.
“It is important to note that the discussant of these presentations can be important to comment on the results and make things right after the presentation, but the conclusions of the authors are important in themselves,” Di Maio said. “We believe that our findings should be the subject of reflection for scientific societies that organize oncology meetings, but also for those who listen during the presentations. It would be interesting to compare the conclusions included in the oral presentation of these trials with the conclusions of their final publication in a peer-reviewed journal. It was too early to perform this analysis when we wrote this letter, but this could be an interesting follow-up.”
For more information:
Massimo Di Maio, MD, can be reached at Ordine Mauriziano Hospital, Via Magellano, Turin 10128, Italy; email: firstname.lastname@example.org.