Introduction Clinical trials are the gold standard for advancing the forefront of medical knowledge and rely on consistent patient accrual for success. However, patient screening for clinical trials is complex and resource intensive. There is a need to increase the scalability of trial recruitment while maintaining or improving upon the sensitivity of the current process. We hypothesized that we could use a state-of-the-art large language model, prompt engineering, and publicly available clinical trial data to predict patient eligibility for clinical trials from clinic notes. Here, we present pilot data demonstrating the accuracy of our tool in a cohort of patients being evaluated for pancreas cancer. Methods We used a cohort of patients who were screened for clinical trials at a single institution for this study. We developed an LLM application using LangChain and the GPT-4o model to assist in clinical trial screening. Deidentified patient data from clinical notes and trial eligibility criteria from ClinicalTrials.gov were used as inputs. For each patient and trial, the model determined inclusion or exclusion with selected eligibility criteria as well as the trial overall. Model responses were graded programmatically against a human rater standard. We recorded time elapsed and cost of running each analysis. Results Of the 24 patients in the test set, 19 were eligible for at least one trial. There were 43 eligible patient-trial matches in our data set. Our model correctly predicted 39 out of 42 (90.7%) of these matches. There were 105 individual eligibility criteria evaluated per patient for a total of 2520 binary criteria. GPT-4o agreed with the raters for 2438 out of 2520 (96.7%) binary eligibility criteria. Sensitivity to overall trial eligibility by trial ranged from 87.5% to 100% for 8 out of 9 trials. The application incorrectly screened the single patient in the CA-4948 trial test set, leading to a sensitivity of 0% for this trial. Specificity ranged from 73.3% to 100% over all nine trials. The median cost for screening a patient was 0.67 USD (0.63-0.74). Median time elapsed was 137.66 seconds (130.04-146.04). Median total token usage across three assistants was 112,266.5 tokens (102982.0-122174.2). Conclusion Overall, our model showed high efficiency and accuracy in selecting patients for appropriate clinical trials. Our results showed promise with a small cohort and future studies are needed to assess its accuracy with a larger sample of patients and trials. This model can be applied systematically to widen the surface area covered by the screening process as well as adapted to any set of clinical trials, opening the door for expansion to other surgical and medical disciplines.

The authors have declared no competing interest.

This study did not receive any funding

I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained.

Yes

The details of the IRB/oversight body that provided approval or exemption for the research described are given below:

IRB of Columbia University gave ethical approval for this work.

I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals.

Yes

I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance).

Yes

I have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable.

Yes

All data produced in the present study are available upon reasonable request to the authors