AI-assisted workflow enables rapid, high-fidelity breast cancer clinical trial eligibility prescreening

2025-11-07

AI Summary

To address low pre-screening efficiency, time-intensive manual verification, and suboptimal patient enrollment rates in breast cancer clinical trials, this study proposes MSK-MATCH: an interpretable, traceable automated eligibility screening framework integrating large language models (LLMs), retrieval-augmented generation (RAG), and a structured oncology knowledge base. The system employs evidence-based, source-document-driven reasoning to ensure transparent decision-making and supports human-in-the-loop triage. Evaluated on 88,518 real-world clinical documents, it achieved fully automated accurate classification for 61.9% of cases (98.6% accuracy), reduced per-case manual review time from 20 minutes to 43 seconds, and lowered associated costs to $0.96. Its core innovation lies in deeply embedding domain-specific oncological knowledge into the LLM’s reasoning pipeline—thereby achieving high accuracy, strong interpretability, and practical clinical deployability.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval

Application Category

• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Clinical trials play an important role in cancer care and research, yet participation rates remain low. We developed MSK-MATCH (Memorial Sloan Kettering Multi-Agent Trial Coordination Hub), an AI system for automated eligibility screening from clinical text. MSK-MATCH integrates a large language model with a curated oncology trial knowledge base and retrieval-augmented architecture providing explanations for all AI predictions grounded in source text. In a retrospective dataset of 88,518 clinical documents from 731 patients across six breast cancer trials, MSK-MATCH automatically resolved 61.9% of cases and triaged 38.1% for human review. This AI-assisted workflow achieved 98.6% accuracy, 98.4% sensitivity, and 98.7% specificity for patient-level eligibility classification, matching or exceeding performance of the human-only and AI-only comparisons. For the triaged cases requiring manual review, prepopulating eligibility screens with AI-generated explanations reduced screening time from 20 minutes to 43 seconds at an average cost of $0.96 per patient-trial pair.

Problem

• Automating clinical trial eligibility screening using AI to improve efficiency
• Enhancing accuracy and speed of breast cancer patient-trial matching
• Reducing manual screening time through AI-generated explanations and predictions

Innovation

• AI system automates clinical trial eligibility screening
• Integrates language model with oncology knowledge base
• Uses retrieval-augmented architecture for explainable predictions