Metastatic vs. Localized Disease As Inclusion Criteria That Can Be Automatically Extracted From Randomized Controlled Trials Using Natural Language Processing

BackgroundExtracting inclusion and exclusion criteria in a structured, automated fashion remains a challenge to developing better search functionalities or automating systematic reviews of randomized controlled trials in oncology. The question "Did this trial enroll patients with localized disease, metastatic disease, or both?" could be used to narrow down the number of potentially relevant trials when conducting a search. Methods600 trials from high-impact medical journals were classified depending on whether they allowed for the inclusion of patients with localized and/or metastatic disease. 500 trials were used to develop and validate three different models with 100 trials being stored away for testing. ResultsOn the test set, a rule-based system using regular expressions achieved an F1-score of 0.72 (95% CI: 0.64 - 0.81) for the prediction of whether the trial allowed for the inclusion of patients with localized disease and 0.77 (95% CI: 0.69 - 0.85) for metastatic disease. A transformer-based machine learning model achieved F1 scores of 0.97 (95% CI: 0.93 - 1.00) and 0.88 (95% CI: 0.82 - 0.94), respectively. The best performance was achieved by a combined approach where the rule-based system was allowed to overrule the machine learning model with F1 scores of 0.97 (95% CI: 0.94 - 1.00) and 0.89 (95% CI: 0.83 - 0.95), respectively. ConclusionAutomatic classification of cancer trials with regard to the inclusion of patients with localized and or metastatic disease is feasible. Turning the extraction of trial criteria into classification problems could, in selected cases, improve text-mining approaches in evidence-based medicine.