SAN DIEGO – Between 2020 and 2024, the number of unique sites in the United States where phase I clinical trials for non-small cell lung cancer (NSCLC) were conducted decreased by 44% and became increasingly concentrated at the top 20 highest‑volume clinical trial sites largely located in major cities, according to results presented at the American Association for Cancer Research (AACR) Annual Meeting 2026 , held April 17-22.
The U.S. Food and Drug Administration (FDA) has been working for decades to improve representation in clinical trials, and last year, the FDA’s Oncologic Drug Advisory Committee (ODAC) convened several meetings to address insufficient enrollment of U.S. patients in clinical trials and the underrepresentation of certain populations; the committee has voted against recommending drugs for approval partly due to questions over how relevant and representative the trial results were to the U.S. population, explained Brittany Avin McKelvey, PhD, senior director of regulatory policy at LUNGevity Foundation. To better understand the concerns being raised by ODAC, she and her colleagues examined the distribution of clinical trial sites in and outside of the United States.
“It is imperative that clinical trial populations reflect as closely as possible the end-user population to sufficiently determine a therapy’s safety and efficacy and inform its optimal use,” McKelvey said. “Therefore, a better understanding of where clinical trials are being conducted, and how the participation of trial sites changes over time, can ensure that we support policies that encourage representativeness and applicability of trial results to the U.S. population.”
McKelvey and her team focused on phase I clinical trials for NSCLC. They identified all interventional, industry-sponsored phase I trials that were listed on ClinicalTrials.gov between January 2020 and December 2024. Unique site locations were determined based on any address that had at least one trial tied to a unique National Clinical Trial (NCT) number. Trials were defined as open if the start date—indicating when the first patient was enrolled—fell within the study’s defined period. Trial instances were defined as the initiation of a unique trial at a unique site. Open trials could have several trial instances across different sites.
During the study period, overall, 555 trials opened for a total of 8,393 trial instances across 47 countries, with the majority located in the United States (45%), China (11%), Spain (8%), Korea (5%), France (4%), and Australia (4%).
McKelvey and colleagues found that while the number of trial instances in the United States increased from 819 in 2020 to 955 in 2022, they declined to 566 by the end of 2024. Similar trends were seen globally with trial instances declining in China (196 to 95), Spain (105 to 80), Korea (80 to 66), and France (48 to 47) over the study period. The only countries with considerable increases in trial instances were Australia (64 to 76), Japan (32 to 49), and Brazil (9 to 29).
Additionally, the number of unique trial sites in the United States decreased from 395 to 223 over the course of the study period. However, the number of trials at the top 20 sites with highest trial volume remained stable over this period with a median of seven to 11 trials opening annually at each of these sites across the five years. These top 20 sites were largely in big cities with an average population of more than 1.9 million.
“Our analysis showed a concerning trend towards trial consolidation at top-performing sites in the United States, further affirming concerns raised at the recent FDA ODAC meetings about site saturation and conflicting with the agency’s appeals to decentralize and move trials into the community,” McKelvey said. “This geographically concentrated clinical trial landscape may result in more limited access to clinical trials.”
McKelvey noted that several factors may have contributed to this decline, including the fact that phase I trials have increasingly complex protocols and the safety profiles of new oncology modalities create substantial infrastructure and expertise requirements that may be prohibitive for smaller sites. Increased regulatory and compliance burdens associated with initiating and conducting trials may have served as another barrier for smaller sites.
“The competitive oncology landscape for trial enrollment may favor established sites with proven track records, creating a self-reinforcing cycle where high-performing sites attract more trials, develop greater expertise, and become increasingly preferred by sponsors—while lower-volume sites struggle to maintain the infrastructure and expertise needed to compete for trials,” McKelvey said.
However, trial instances did not decline for all lower-volume sites across the study period, McKelvey noted. Low-volume sites that were part of the National Cancer Institute Community Oncology Research Program (NCORP) were able to largely maintain their trial instances over the observed time period compared with other low-volume sites, she explained.
“This could suggest that the infrastructure support and integration into research networks at NCORP sites may be able to buffer these community sites against the factors driving consolidation,” McKelvey said. “Overall, addressing this consolidation and ensuring identification of the appropriate sites for clinical research will require coordinated efforts to reduce site activation barriers, provide infrastructure support, and potentially implement policy incentives that encourage diversification of trial portfolios within the United States.”
Limitations of this study include the heterogeneity of ClinicalTrials.gov data, which made it difficult to determine if sites within the same zip code with relatively similar names should be treated as institutions that have satellite locations within communities or grouped together as a single site. Additionally, since patient enrollment numbers are not publicly available, the number of sites actively enrolling patients may be lower than the number of trial instances examined in this study. Finally, the focus on NSCLC and phase I trials may not fully represent patterns in other cancer types or late-phase trials. McKelvey added that her team is considering studies on later-phase clinical trials, which may allow for broader participation by less experienced sites given the known safety profiles, to compare analyses.
This study was funded by LUNGevity Foundation through unrestricted grants. McKelvey declares no conflicts of interest.