CHICAGO — Late-phase trials that showed superiority for an experimental intervention on a surrogate endpoint did not necessarily lead to improvements in overall survival (OS) or quality of life (QOL), according to a meta-epidemiologic analysis.

Alexander Dean Sherry, MD of the MD Anderson Cancer Center in Houston, and colleagues evaluated 791 randomized controlled trials (RCTs) that employed a superiority design on a primary endpoint, mostly an endpoint other than OS, such as progression-free survival (PFS).

Some 53% of the RCTs were interpreted as showing primary endpoint superiority, but OS and QOL superiority were shown in just 28% and 11% of trials, respectively. In 32% of the trials, superiority of either OS or QOL was demonstrated, while superiority of both were shown in only 6%, Sherry’s group reported at the American Society of Clinical Oncology (ASCO) meeting. The findings were published simultaneously in JAMA Oncology.

“Even though the majority of RCTs are interpreted as positive, improvements in OS or QOL remain relatively uncommon and are exceedingly rare in combination,” the investigators wrote. “QOL data appear to be both underevaluated and underreported, a finding that is consistent with previous smaller studies.”

The findings suggest that “the heightened focus on alternative endpoints, now representing the most common primary endpoint, has overshadowed OS and QOL among modern phase 3 RCTs” and that “modern phase 3 oncology RCTs appear to be severely underachieving in creating improvements that directly affect how patients live and feel,” they said.

A meta-epidemiologic analysis examines how design, funding, and methodologic biases shape reported results. Sherry and colleagues used trial-level data and directed acyclic graphs to identify and control for confounders.

U.S.-based, phase III oncology RCTs published between 2002 and 2024 were found using the ClinicalTrials.gov registry. Of the 791 phase III RCTs identified, only 25% used OS as a primary or co-primary endpoint while 35% used an alternate endpoint and 23% used QOL or toxicity as the primary endpoint.

A total of 705 RCTs (89%) reported OS results. RCTs of metastatic tumors, industry-funded RCTs, and RCTs studying gastrointestinal malignancies were each more likely to demonstrate OS superiority.

Among the 434 RCTs with a positive alternative endpoint, only 43% showed OS superiority.

The investigators also found that global QOL superiority was seen in 10% of RCTs via statistical significance and 3% by minimum clinically important differences (MCID) criteria.

Global QOL was worsened in the experimental group versus the control group in 4% of trials according to statistical significance criteria and in 0.4% of trials RCTs when using MCID criteria. Industry-funded RCTs and metastatic solid tumor RCTs were more likely to demonstrate global QOL superiority, whereas cooperative group-sponsored RCTs and larger trials were less likely to find global QOL superiority.

In 31 RCTs, either OS or global QOL was inferior with the experimental intervention, including 16 RCTs in which the primary endpoint was considered superior.

The authors acknowledged that OS and QOL “for certain RCTs may be published in the future after additional follow-up.” Also, current results may not apply to studies done outside the U.S. or to studies with ≥3 groups as “these trials were not included in our study due to the challenge of interpreting multigroup results at scale,” they said.

Sherry’s group stated that the findings “argue strongly for renewed focus on OS and QOL in late-phase oncology trials and in the regulatory approval process.”

In an accompanying editor’s note, Fei Ye, PhD, MSPH, and Yu Shyr, PhD, both of Vanderbilt University in Nashville, pointed out that meta-epidemiology reduces bias, thereby fostering “more transparent and reliable oncology research, ultimately enhancing clinical decision-making and patient care.”

They said that the discrepancy between PFS and OS and QOL “underscores the broader question of whether alternative early clinical end points accurately reflect meaningful long-term clinical benefits.”

Please enable JavaScript to view the

comments powered by Disqus.