Characteristics of Clinical Trials in Community Oncology Practices Across the US

In this interview, Ivy Altomare, MD, senior medical director, Flatiron Health, shares the results from her study investigating the characteristics of clinical trials carried out in community oncology practices within the Flatiron Health network across the US.

Please introduce yourself by stating your name, title, organization, and relevant professional experience.

Hi, I'm Ivy Altomare, MD. I am a medical oncologist and one of the senior medical directors at Flatiron Health. I am the clinical lead for the Flatiron Research Network. I interface with investigators and research teams at community practices and academic centers that use Flatiron's technology to support their clinical research trial portfolio and their clinical research programs.

Please give some background about your study and what prompted you to undertake it.

Flatiron is a company that provides and maintains an oncology-specific electronic health record that is tailored to the specific needs of community practices. And so, we're always interested in research that is aimed at studying and supporting ways to deliver optimal cancer care, whether that's through practice management tools, clinical decision support, or studying clinical research infrastructure, and that's what I'm most interested in.

I believe that extending clinical trials to the community setting is critical for a number of reasons. It's the way that studies can meet their accrual goals. It ensures that the results of clinical research are generalizable since half of the patients get their care in the community. It can also facilitate optimal and equitable cancer care since access to clinical trials is such a critical component of providing quality cancer care.

There's a perception that research in the community is not well done, that community clinics have poor infrastructure for research, or they have low trial volume and can only handle a couple of trials at a time. So, we wanted to formally assess this. Also, the types of trials that are attractive to community practices and successful in the community are likely different than those that are run primarily at academic centers. It turns out that this has never been well studied.

The idea for this research project was to look at the characteristics of clinical trials specifically in a community setting. We did this in collaboration with the Association of Cancer Care Centers (ACCC), a well-established, huge advocacy and policy nonprofit that offers resources to community cancer clinics. They were interested in collaborating with us on this research project and studying this question as well.

Can you briefly describe how the study was conducted?

We were able to conduct this study because of a tool called OncoTrials. This is a tool that is available in the OncoEMR electronic health record (HER), and it allows research teams to screen patients for clinical trials that they have open at their practice and track accruals for patients that are enrolled in those trials. To use this tool, research teams at the community site enter the trial NCT number from clinicaltrials.gov. They can then match their patients to trials based on the clinical trial criteria, such as the diagnosis, whether or not the patient is metastatic, or whether the patient has the presence or absence of certain biomarkers.

Since sites used this tool to support their research programs, we were able to pull the NCT number of every clinical trial entered into OncoTrials. The observation period was 2 years, between October 2021 and October 2023. In that way, we were able to assess the types of trials that were open across a whole network of community practices in a 2-year period.

In order to meet criteria for our study, the trial must have been open at no less than one practice in the Flatiron Research Network and the study must have accrued at least one patient at that practice within the 2-year observation period. We tracked the studies that were open at these sites across the observation period. And then we used descriptive statistics to describe the characteristics of the trials, and we accessed those through an application programming interface (API) from clinicaltrials.gov.

What were the main findings of your study?

This abstract is now published on the American Society of Clinical Oncology (ASCO) website and in the Journal of Clinical Oncology in conjunction with this year's ASCO conference. In summary, there were 1310 clinical trials open at 64 practices during our observation period. It was a high number and we sort of debunked the myth that clinical research in the community can only support a low volume of trials. These were a lot of studies open at these practices. The mean number of studies open at a single site was 54, again large portfolio of trials. The median was 35 and the range was one study open at practice during this observation period to 264, so a wide range. So, those were the results in terms of trial volume.

In terms of the types of studies, 80% of the trials were industry sponsored. About 5% were sponsored by networks such as ASCO or accrue or Canadian cancer trial groups. About 6% were National Cancer Institute (NCI)-funded cooperative group studies, open through either NCI Community Oncology Research Program (NCORP) or a direct affiliation with a core operative group such as NRG Oncology or Alliance or the Southwest Oncology Group (SWOG). And about 9% of studies were investigator-initiated studies accessed through an academic center affiliation.

In terms of the study type, 83% of trials were interventional studies where treatment was offered or studied, and the vast majority of those were phase 2 to 3 clinical trials. The other 7% of studies opened during our observation period in the community were either observational, registry, or expanded access. So again, there's a perception that a lot of research in the community is just observational.

In fact, we found the opposite, that the vast majority are treatment trials. The most interesting finding from our study was that 40% of trials were open at only one practice in our cohort and only 2.6% of studies were open at 10 or more practices.

Another finding was that 48% of sites in our cohort had at least one phase 1 study open during our observation period. Of all of those 1310 studies, 15% of them (185) were phase 1.

Looking ahead, what potential impact do you hope your findings will have on the collaboration and network utilization among community oncology practices for expanding clinical trials in this space?

As I said earlier, this study debunks some of the common myths about cancer clinical research in the community, specifically that community sites can only handle a few studies at a time or that they can't do phase 1. The results clearly showed that community oncology practices do support a robust portfolio of clinical trials. Most of them are focused on the treatment of solid tumors.

The vast majority were interventional and mostly phase 2, phase 3, and industry sponsored. Half of the sites were able to run phase 1 studies in our cohort. Although phase 4 studies were uncommon, they tended to be popular and open at multiple sites at one time. The phase 4 studies did have broad participation even though there were an empirically low number of phase 4 studies open during our observation period in this study.

As I called out in the results, a large proportion of studies were only open at one single site in our cohort of 64 community oncology practices and we don't know the reason for this. We don't know whether this is because sponsors are hesitant to offer studies in the community or whether it's  because the sites are ultra selective about the studies that they open. But I think it's the former and  this highlights the opportunity for pharma sponsors to take advantage of the quality and the commitment that community clinics have to clinical research, especially community clinics that are tech enabled with tools that can support clinical research infrastructure. These include tools that enable screening, tools that facilitate automated transfer of clinical trial data, or tools that facilitate EHR to electronic data capture data connection.

It's an opportunity for sponsors to leverage community networks and consortia to ensure that appropriate studies are accessible among a broader range of community research sites.

What session studies stood out to you at this year's ASCO annual meeting that could have an impact on clinical trials or clinical pathways in the community oncology setting.

Treatment decision-making is becoming increasingly complex and that was evident this year and every year at ASCO, with such a broad and high volume of practice changing data across multiple tumor types being presented every year.

There are a couple of things to call out in terms of practice changing studies. There was the NADINA  trial calling for neoadjuvant doublet immunotherapy followed by response-driven adjuvant therapy for patients with early-stage melanoma that was in the plenary session. There was also the ADRIATIC Trial, which shows the benefit of escalation of adjuvant immunotherapy after definitive treatment for early-stage small cell lung cancer. So, there was a lot of data impacting treatment decision-making in early-stage settings.

There was also work demonstrating how treatments that were formally thought of as needing to be given inpatient or supported at academic centers to now being able to move into the community, such as CAR-T therapy or administration of BiTE therapy. Therefore, the need for accessible and interoperable point of care decision support for community practitioners has never been greater. In addition, there was a lot of data presented around new ways to incorporate diagnostic testing at various points in clinical care. There is a need for better guidance here as well, not only in terms of treatment, but in terms of the diagnostic tests that we order for cancer patients.

There's an ever-expanding role for circulating tumor DNA analysis, minimal residual disease  analysis, broad panel next-generation sequencing testing in earlier stage settings, and an increasing public interest in multicancer early detection tests. So, it's becoming much more complex for oncologists to know what type of molecular or biomarker testing to order, not only for diagnostic purposes, but also for tracking purposes as patients advance in their treatment journey. This area is becoming increasingly complex and decision support tools could be really useful here.

But the issue is that there's limited evidence that has prospectively evaluated whether and how those diagnostic tests should impact clinical decision-making. Not only do we need clinical decision support tools, but we need the evidence in order to develop those decision support tools. This is a whole area of high need and something that's going to be a focus of further research in the future.