Menu Back to Poster-Presentations-Details

P229: Using Structured Clinical Trial Information from the National Cancer Institute to Improve Cancer Clinical Trials Search

Poster Presenter

Rebecca Williams

Senior Clinical Trials Subject Matter Expert
Essex, Part of Emmes Group
United States

Objectives

To better support patients in finding relevant cancer clinical trials, we created a prototype search tool that integrates and adapts freely available, cancer-specific, structured content from the National Cancer Institute’s Clinical Trials Reporting Program (CTRP) Clinical Trials Search (CTS) API.

Method

From February 2021 to December 2023, we accessed active, treatment trials from the CTRP CTS API and integrated structured content into a web-based search tool. We also curated select eligibility criteria to further support patients in narrowing search results to a focused list of relevant trials.

Results

We integrated 4,151 active, treatment trials from the CTRP CTS API into the prototype search tool, emFACT (essex management Find A Clinical Trial). We used CTRP for its structured, cancer-focused terminology and frequently updated standardized site information. It covers about 85% of recruiting, cancer clinical trials in the United States (U.S.) listed on ClinicalTrials.gov. We integrated and augmented CTRP CTS API content that is most impactful in search: disease, lead disease, eligibility criteria (structured biomarkers and free text), and trial location information. CTRP structures and enriches disease information using ontological data from NCI Thesaurus (NCIt), a widely recognized standard for biomedical coding and reference. Tumor biomarkers (e.g., EGFR mutation) are also coded using NCIt and available in CTRP as structured eligibility criteria. We augmented CTRP CTS API content by structuring impactful and/or frequently occurring eligibility criteria (e.g., brain metastases, prior therapy, HIV, performance status, platelet count) for 3,913 trials. As a case study, in March 2024, there were 650 small cell lung carcinoma (SCLC) and non-SCLC (NSCLC) trials open to enrollment in the CTRP CTS API. An emFACT search identified 91 SCLC trials (86% decrease), with no trials of NSCLC (unless trial is both SCLC and NSCLC), and a focused SCLC lead disease search (excludes general solid tumor trials) identified 52 SCLC trials (92% decrease). Applying relevant emFACT curated eligibility criteria further focused search results (about a 95% decrease), for example, by applying brain metastasis, prior anti-PD-L1 agent treatment (prior therapy), HIV infection, and performance status level two (work activity restricted but ambulatory and capable of self-care). Using geographic location further focuses search to allow for critical review of trials most relevant to potential participation.

Conclusion

Our experience using CTRP CTS API content demonstrated it is a rich resource of structured information tailored to cancer clinical trials conducted in the U.S.. We effectively integrated its content into a web-based search tool to support patients in retrieving a list of trials most likely to be relevant to their cancer diagnosis and practical considerations, such as trial location. Disease information is most impactful in improving search and CTRP’s structured disease information using NCIt helps to eliminate irrelevant trials (for example, excluding NSCLC trials when searching for SCLC). We also demonstrated that techniques for normalizing and structuring free text eligibility criteria can be integrated to improve search. The CTRP CTS API continues to improve and evolve. After we initiated our project, prior therapy structured eligibility criteria content was added to the API. This addition is notable as patients may turn to clinical trials after first and second-line therapies have been exhausted and, as in our example, being able to include or exclude trials based on prior treatment leads to meaningful search result improvements. CTRP CTS API limitations include not covering all cancer trials in the U.S., however third-party applications may augment content by including other sources (such as ClinicalTrials.gov). The CTRP CTS API contains structured and frequently updated site information, but its focus on trials with NCI direct or indirect funding means other site content may not be as up to date. The emFACT prototype search tool is being further developed to apply these methods to structure additional eligibility criteria that are useful for pediatric populations, hematopoietic cancers, and treatment-naïve patients. The approach used in the prototype to structure eligibility criteria primarily relied on manual curation, but additional work focuses on automated approaches that leverage large language models and artificial intelligence.