Jin Qiao, Wang Zifeng, Floudas Charalampos S, Chen Fangyuan, Gong Changlin, Bracken-Clarke Dara, Xue Elisabetta, Yang Yifan, Sun Jimeng, Lu Zhiyong
National Center for Biotechnology Information (NCBI), National Library of Medicine (NLM), National Institutes of Health (NIH).
Department of Computer Science, University of Illinois Urbana-Champaign.
ArXiv. 2024 Nov 18:arXiv:2307.15051v5.
Patient recruitment is challenging for clinical trials. We introduce TrialGPT, an end-to-end framework for zero-shot patient-to-trial matching with large language models. TrialGPT comprises three modules: it first performs large-scale filtering to retrieve candidate trials (TrialGPT-Retrieval); then predicts criterion-level patient eligibility (TrialGPT-Matching); and finally generates trial-level scores (TrialGPT-Ranking). We evaluate TrialGPT on three cohorts of 183 synthetic patients with over 75,000 trial annotations. TrialGPT-Retrieval can recall over 90% of relevant trials using less than 6% of the initial collection. Manual evaluations on 1,015 patient-criterion pairs show that TrialGPT-Matching achieves an accuracy of 87.3% with faithful explanations, close to the expert performance. The TrialGPT-Ranking scores are highly correlated with human judgments and outperform the best-competing models by 43.8% in ranking and excluding trials. Furthermore, our user study reveals that TrialGPT can reduce the screening time by 42.6% in patient recruitment. Overall, these results have demonstrated promising opportunities for patient-to-trial matching with TriaGPT.
患者招募是临床试验面临的一项挑战。我们引入了TrialGPT，这是一个用于通过大语言模型进行零样本患者与试验匹配的端到端框架。TrialGPT由三个模块组成：它首先进行大规模筛选以检索候选试验（TrialGPT检索）；然后预测标准层面的患者资格（TrialGPT匹配）；最后生成试验层面的分数（TrialGPT排名）。我们在三个包含183名合成患者及超过75000条试验注释的队列上对TrialGPT进行了评估。TrialGPT检索能够使用不到初始集合6%的内容召回超过90%的相关试验。对1015个患者-标准对的人工评估表明，TrialGPT匹配在给出可靠解释的情况下准确率达到87.3%，接近专家表现。TrialGPT排名分数与人类判断高度相关，在试验排名和排除方面比最佳竞争模型高出43.8%。此外，我们的用户研究表明，TrialGPT在患者招募中可将筛选时间减少42.6%。总体而言，这些结果证明了使用TrialGPT进行患者与试验匹配具有广阔前景。
