🤖 AI Summary
To address the critical challenges of insufficient patient recruitment in oncology clinical trials and low efficiency of manual eligibility screening, this paper proposes TrialSpace: an interpretable, AI-assisted matching framework grounded in semantic embeddings. Methodologically, it introduces a novel decoupled matching paradigm—jointly modeling patient electronic health records and trial protocols within a disease-specific semantic space. Clinical BERT is fine-tuned to extract domain-aware features, followed by vector-based candidate retrieval and a lightweight binary classifier for eligibility verification. Key contributions include: (1) the first open-source clinical trial matching toolkit and a synthetically generated benchmark dataset; (2) strong performance (Top-5 recall >92% on synthetic data), model interpretability, and clinical deployability; and (3) a publicly available interactive demo, along with full source code and pre-trained model weights. TrialSpace significantly enhances physician screening efficiency while supporting evidence-based decision-making—not replacing clinicians, but augmenting their expertise.
📝 Abstract
Clinical trials drive improvements in cancer treatments and outcomes. However, most adults with cancer do not participate in trials, and trials often fail to enroll enough patients to answer their scientific questions. Artificial intelligence could accelerate matching of patients to appropriate clinical trials. Here, we describe the development and evaluation of the MatchMiner-AI pipeline for clinical trial searching and ranking. MatchMiner-AI focuses on matching patients to potential trials based on core criteria describing clinical"spaces,"or disease contexts, targeted by a trial. It aims to accelerate the human work of identifying potential matches, not to fully automate trial screening. The pipeline includes modules for extraction of key information from a patient's longitudinal electronic health record; rapid ranking of candidate trial-patient matches based on embeddings in vector space; and classification of whether a candidate match represents a reasonable clinical consideration. Code and synthetic data are available at https://huggingface.co/ksg-dfci/MatchMiner-AI . Model weights based on synthetic data are available at https://huggingface.co/ksg-dfci/TrialSpace and https://huggingface.co/ksg-dfci/TrialChecker . A simple cancer clinical trial search engine to demonstrate pipeline components is available at https://huggingface.co/spaces/ksg-dfci/trial_search_alpha .