Netherlands-based myTomorrows has launched a beta version of TrialSearch AI to help physicians connect their patients with suitable clinical trials and expanded access programmes (EAP).

The artificial intelligence (AI) tool will broaden physician insights into available studies that are curated in myTomorrows’s database, which is sourced from public registries. The aim of TrialSearch AI is to simplify the search process and help physicians identify opportunities for suitable pre-approval treatments based on their patient’s profile.

CEO Dr Michel van Harten said: “As a physician who cared for patients diagnosed with cancer, I know first-hand the importance of providing actionable information that can help physicians, patients, and caregivers discover and access potentially life-saving treatments.”

TrialSearch AI operates on the GPT series of Large Language Model (LLM) AI models that are developed by OpenAI. The LLM model will comb through public trial registries, such as and the EU’s Clinical Trials Register, to provide an unbiased and comprehensive breakdown of pre-approval options.

Physicians will have to provide de-identified patient profile information, such as their condition, sex, age, and other relevant medical information in a free-text form. This input will be used to identify eligibility criteria and provide a list of suitable clinical trials and EAPs.

Earlier this month, myTomorrows’s investigators conducted a test run on how the tool operates by creating 10 fictitious patient profiles across 10 different diseases. The results showed that the LLM tool was able to correctly identify the screenability of 72% of the criteria and reduce pre-screening checking time for physicians by 90%.

Researchers concluded that by forcing instruction-tuned LLMs to produce chain-oh thought responses, the decision process can be amended by physicians and reasoning can be made transparent. As such, the tool can be feasible in real-world scenarios. The company is currently accepting requests from physicians to trial the tool.

Clinical Trials Arena has previously investigated the benefits of AI to identify cohort selection and patient recruitment. Yet, experts believe that data management and inherent bias in datasets can hinder its application. AI also has the potential to assist with patient adherence and retention, while pointing to clinical trial sponsor blind spots, and improve data analysis that is generated from digital biomarkers and wearables.

Other AI-powered models, such as BioGPT, are entering the clinical research industry. This generative language model is trained on millions of previously published biomedical research articles and has demonstrated human parity in answering biomedical research questions.