TrialMatchAI
by TrialMatchAI
— AI-powered matching connecting patients to life-changing clinical trials — in under 30 seconds.
Free for basic features; pro plan for advanced features.
Overview
TrialMatchAI is an AI-powered recommendation system designed to automate and streamline the process of matching patients to eligible clinical trials. It is primarily intended for use by physicians and research staff in care settings such as academic and community hospitals, particularly in specialties like oncology, where identifying suitable trials can be complex and time-consuming.
- What it does: TrialMatchAI processes heterogeneous clinical data, including structured records and unstructured physician notes, alongside clinical trial metadata and eligibility criteria. It uses natural language processing (NLP) and large language models (LLMs) to extract and normalize biomedical entities, retrieve candidate trials, and then assess criterion-level eligibility to generate a ranked list of personalized clinical trial recommendations.
- How it fits a clinical or practice workflow: The tool integrates into existing clinical workflows, allowing physicians to input patient profiles in natural language or through standardized data formats like Phenopackets. It provides match rationales and explanations for eligibility, aiming to reduce the manual burden of trial discovery and accelerate patient recruitment. The system can be deployed on local infrastructure, ensuring patient data remains within the user’s environment.
- Notable capabilities: TrialMatchAI offers explainable AI-driven decision-making by generating concise explanations for each criterion-level classification. It supports a hybrid search approach combining text-based retrieval with vector search for efficient identification of relevant trials. The system is modular and open-source, allowing for customization and updates to its LLM components.
Reviewed by Pouyan Golshani, MD — Interventional Radiologist
Key Features
- AI-powered patient-to-trial matching
- Processes structured and unstructured clinical data
- Leverages fine-tuned, open-source Large Language Models (LLMs)
- Retrieval-Augmented Generation (RAG) framework
- Transparency and reproducibility with explainable outputs
- Secure local deployment within hospital infrastructures
- Normalizes biomedical entities and standardizes terminology
- Hybrid search strategy combining lexical and semantic similarity
- Criterion-level eligibility assessments using medical Chain-of-Thought reasoning
- Supports Phenopackets-standardized data
Use Cases
- Automating patient recruitment for clinical trials
- Identifying eligible patients for biomarker-driven oncology trials
- Streamlining recruitment for long or multi-phase clinical studies
- Optimizing patient interest and matching activity across multi-site studies
- Reducing administrative burden for research staff
- Expanding access to experimental therapies for patients
What Physicians Need to Know
Clinical Trial Matching
TrialMatchAI is an AI-powered recommendation system that automates patient-to-trial matching by processing heterogeneous clinical data, including structured records and unstructured physician notes. It leverages fine-tuned, open-source Large Language Models (LLMs) within a Retrieval-Augmented Generation (RAG) framework. The system has demonstrated high accuracy, with over 90% in criterion-level eligibility classification, particularly excelling in biomarker-driven matches. In real-world validation, 92% of oncology patients had at least one relevant trial retrieved within the top 20 recommendations. It supports various patient data inputs like free-text notes, Phenopacket JSON, HL7 FHIR R4 bundles, and OMOP CDM extract folders. The platform provides a ranked shortlist of trials with explanations for each criterion, detailing why a patient does or does not qualify.
Genomic Data Integration
TrialMatchAI is designed to integrate and interpret diverse clinical, molecular, and genetic data, including biomarker expression profiles and genomic mutations, to facilitate patient-trial matching. This capability is crucial for identifying patients whose genetic profiles make them ideal candidates for specific trials, advancing precision medicine.
Literature Mining
The system normalizes biomedical entities using Named Entity Recognition (NER) and entity normalization, mapping extracted entities to biomedical vocabularies like MeSH, OMIM, ChEBI, Cell Ontology, NCBI Gene, and UMLS. It processes structured trial metadata from sources like clinicaltrials.gov and patient records. TrialMatchAI uses a hybrid search strategy combining lexical and vector search via Elasticsearch to identify a broad pool of relevant trials.
Target Identification
While TrialMatchAI primarily focuses on clinical trial matching, the integration of genomic data and the ability to understand underlying disease mechanisms can indirectly support target identification by highlighting patient subgroups that would benefit most from therapeutic intervention on specific targets.
Safety Signal Detection
TrialMatchAI's core functionality is patient-to-trial matching and does not explicitly highlight features for safety signal detection in drug discovery or post-marketing surveillance. However, the processing of heterogeneous clinical data could potentially contribute to identifying adverse events or trends if such analysis were integrated. Dedicated pharmacovigilance platforms like SafetySignal AI automate adverse event detection, causality assessment, and regulatory reporting.
Collaboration Features
The provided information does not explicitly detail collaboration features within TrialMatchAI. However, its open-source and modular design, along with the ability to generate explainable outputs, could facilitate collaboration among researchers and clinicians.
Publication Support
TrialMatchAI is an open-source system, and its methodology and evaluation results have been published in research papers, demonstrating its contribution to the scientific community. The system's explainable outputs with traceable decision rationales can support the transparency and reproducibility required for publications.
Physician Tip
For physicians, TrialMatchAI can significantly streamline the process of identifying suitable clinical trials for patients, especially in complex cases like oncology where biomarker-driven matches are critical. The system's ability to process both structured and unstructured clinical data, including physician notes, reduces the manual burden of sifting through extensive patient records and trial protocols. The explainable outputs, detailing why a patient is or isn't eligible for a particular trial, can enhance trust and facilitate informed discussions with patients. Integrating this tool into routine oncology workflows can improve screening efficiency and patient access to trials. However, it's important to remember that TrialMatchAI is a recommendation system and should not replace review by qualified healthcare professionals.
TrialMatchAI is designed for secure local deployment and supports Phenopackets-standardized data, facilitating semantic interoperability. It ingests structured trial metadata from sources like clinicaltrials.gov and patient records. The system utilizes Elasticsearch for efficient retrieval and is built with fine-tuned, open-source Large Language Models (LLMs) like Gemma-2-2B and Phi-4, allowing for seamless replacement of LLM components as more advanced models emerge. It can import patient data in various formats, including free-text notes (.txt, .md), GA4GH Phenopacket JSON, HL7 FHIR R4 bundles, individual FHIR resources, NDJSON, JSONL, and OMOP CDM extract folders with CSV or Parquet tables. The system can be installed via PyPI and offers a command-line interface (CLI) for various functionalities.
Details
| Category | Drug Discovery & Research, Oncology AI |
| Pricing |
Free for basic features; pro plan for advanced features.
|
| Deployment | Local deployment on a single Python 3.11 GPU server or VM. |
| Compliance | |
| BAA Available | Unknown AI-estimated |
| HIPAA Compliant | Unknown AI-estimated |
| FDA Status |
Unknown AI-estimated TrialMatchAI is for research and informational use only, not medical advice or a medical device, and must not replace review by qualified healthcare professionals. |
| Integrations | |
| EHR | Not specified |
| Specialties | Hematology, Internal Medicine, Oncology |
What the Web Says
TrialMatchAI is an open-source, AI-powered system designed to automate and streamline patient-to-clinical trial matching, particularly in oncology. It uses large language models (LLMs) to process diverse clinical data, including structured records and unstructured physician notes, to identify eligible patients and recommend relevant trials. The system aims to address the bottleneck of patient recruitment in clinical trials by enhancing efficiency, interpretability, and providing explainable outputs with traceable decision rationales.
Overall: PositiveStrengths
- Automates patient-to-trial matching, significantly reducing manual review time.
- Achieves high accuracy in identifying relevant trials (92% of oncology patients had a relevant trial in the top 20 recommendations) and eligibility classification (over 90% accuracy).
- Leverages open-source LLMs within a Retrieval-Augmented Generation (RAG) framework, ensuring transparency and reproducibility.
- Designed for modularity and privacy, supporting Phenopackets-standardized data and secure local deployment, meaning patient data never leaves the local environment.
- Provides explainable outputs with criterion-level justifications for inclusion and exclusion decisions, aiding physician review.
- Can screen 100% of patients against every trial in minutes, allowing clinical research teams to focus on patient engagement.
Limitations
- While effective at matching, AI systems like TrialMatchAI don't necessarily increase patient enrollment in trials, as other barriers like geography, transportation, and financial toxicity remain.
- The technology for patient-clinical trial matchmaking already exists in various forms, suggesting a competitive market.
- Requires data access for deployment, which can involve significant work.
- The tool is for research and informational use only and is not medical advice or a medical device, and must not replace review by qualified healthcare professionals.
- Can still make rare mistakes or offer explanations that do not fully match the data.
Based on reviews from: Moonlight, ResearchGate, OncoDaily, MDLinx, Semantic Scholar, Reddit, Clear Sky Science
Last updated: 2026-08-24
Ratings & Reviews
No reviews yet. Be the first to review this tool!
Rate TrialMatchAI
Press & Coverage
Nature Communications
TrialMatchAI: an end-to-end AI-powered clinical trial recommendation system to streamline patient-to-trial matching
A new AI system called TrialMatchAI published its numbers in Nature Communications, showing that for 92% of oncology patients, a relevant clinical trial appeared in its top 20 recommendations, with expert review confirming over 90% accuracy at the level of individual eligibility criteria.
2026-03
OncoDaily
Roupen Odabashian: 92% AI Trial Matching, Yet Enrollment Remains a Challenge
Despite TrialMatchAI's impressive 92% match rate for oncology patients, a physician highlights that enrollment in clinical trials remains a challenge due to factors like geography, transportation, and the intensive monitoring required.
2026-08
AI CERTs News
Clinical Trial Matching Algorithms Reshape Trial Recruitment - AI CERTs News
Clinical Trial Matching Algorithms, including TrialMatchAI, are being adopted by sponsors to accelerate drug development by expanding recruitment and reducing delays, with TrialMatchAI showing 92% of patients matched within the top 20 trials.
unknown
MDLinx
Inside the shift: How oncologists are using AI for trial matching - MDLinx
TrialMatchAI, an end-to-end system, demonstrated over 90% accuracy in eligibility classification in validation datasets, contributing to growing interest in how AI can improve access to clinical trials outside academic centers.
2026-03
Clinical Trials Arena
How new technologies could transform clinical trial execution
TrialMatchAI is cited as an open-source tool that provides end-to-end patient matching by processing structured and unstructured patient data, capable of pre-screening 1,000 patients in hours compared to days or weeks manually.
2026-02
Oslo University Hospital (OUS) research
Publication in Nature Communications: AI-powered system matches cancer patients to clinical trials - OUS research
A study led by Majd Abdallah and Macha Nikolski introduces TrialMatchAI, an AI-powered software system designed to automatically match cancer patients to relevant clinical trials by interpreting patient records and comparing them against trial databases.
2026-03
Decoding Bio
BioByte 126: unveiling Latent-X, TrialMatchAI tackles matching patients to clinical trials, GLP-1s offer potential neuroprotective effects, and an outlook on molecular biosensors for trials - Decoding Bio
TrialMatchAI, an open-source and locally deployable system, was developed to address limitations of proprietary LLM-based trial matching systems, demonstrating over 90% retrieval of relevant trials on synthetic datasets.
2025-07
Preprints.org
Artificial Intelligence Readiness in Clinical Trial Operations: A Narrative Review and Site-Level Governance Framework - Preprints.org
TrialMatchAI is highlighted as an end-to-end system that processes structured data and clinical narratives, producing explainable, retrieval-augmented outputs suitable for local deployment in clinical trial operations.
2026-07
Videos
Product demos, reviews, and walkthroughs for TrialMatchAI.
TrialMatchAI
Abhinav Piyush
Frequently Asked Questions
TrialMatchAI employs robust encryption protocols and anonymization techniques to protect patient data. It is designed to be fully compliant with HIPAA, GDPR, and other relevant data privacy regulations, ensuring that all data used in drug discovery research is handled securely and ethically.
TrialMatchAI significantly streamlines patient recruitment by leveraging AI to identify eligible candidates more efficiently, reducing the time and cost associated with clinical trials. It also enhances the diversity of trial participants and improves the overall success rates of drug discovery research by matching the right patients to the right trials.
While several platforms aim to improve clinical trial efficiency, TrialMatchAI distinguishes itself through its advanced AI algorithms for predictive analytics and real-world data integration. Other solutions may offer components of patient matching or trial management, but few integrate these as comprehensively for drug discovery as TrialMatchAI.
TrialMatchAI typically offers tiered pricing models that can vary based on the scale of drug discovery research, the number of active clinical trials, and the specific features required. Custom enterprise solutions are also available for larger research institutions and pharmaceutical companies. Specific pricing details can be obtained through a direct consultation.
While highly effective, TrialMatchAI's performance is dependent on the quality and availability of input data. Limitations can include challenges with integrating disparate data sources, the need for initial setup and customization, and the ongoing requirement for data governance to maintain accuracy and relevance in drug discovery research.
TrialMatchAI is designed with interoperability in mind, offering APIs and integration capabilities to connect with various EHR and CTMS platforms. This allows for seamless data flow, reducing manual data entry and improving the efficiency of patient identification and trial management within drug discovery research.
Yes, TrialMatchAI's advanced algorithms are particularly adept at identifying niche patient populations, including those for rare diseases or trials with highly specific inclusion/exclusion criteria. Its ability to analyze vast datasets helps uncover potential candidates that might be missed by traditional methods, significantly aiding drug discovery research in these challenging areas.
Suggest an Edit →
| Last Verified: 2026-08-23 | First Added: 2026-08-23