肿瘤免疫治疗证据动态（CIEL）文库的原理与设计：持续更新的肿瘤免疫治疗临床试验数据库

We included planned, ongoing, and completed interventional trials of immunotherapies for cancer, regardless of trial design (e.g., randomization, blinding, and type of comparator). We systematically searched PubMed (for published reports) and ClinicalTrials.gov (for registered clinical trials). PubMed-retrieved records were screened using the AI-assisted software ASReview and manually extracted and curated. We imported data from ClinicalTrials.gov using the Clinical Trials Transformation Initiative database, which then requires further curation. The CIEL-Library was implemented as a web application. It also contained the "Match My Patient" feature, a patient-centered clinical decision support system, aiming to filter planned, ongoing, or completed trials based on four patient characteristics (disease staging, previous treatments, performance status, and location). We piloted our database with one type of cancer immunotherapy, the tumor-infiltrating lymphocytes (TIL) transfer. The CIEL-Library was a prototype and no further developments are planned.

The CIEL-Library offers a blueprint for a dynamic evidence synthesis infrastructure by providing a collection of clinical trials with curated trial characteristics and results. This blueprint may be applied across fields, specialties, and topics. The main challenges to making a database of clinical trials are the time and resources needed to populate it with curated and updated data. The CIEL-Library project highlights the potential and the main limitations of designing trial databases intended to be used in routine care.