Castellano T, Lara O D, McCormick C, Chase D, BaeJump V, Jackson A L, Peppin J T, Ghamande S, Moore K N, Pothuri B, Herzog T J, Myers T
Louisisana State University HSC-New Orleans, Division of Gynecologic Oncology, LA, United States of America.
University of North Carolina, Division of Gynecologic Oncology, Chapel Hill, NC, United States of America.
Gynecol Oncol. 2025 Jan;192:111-119. doi: 10.1016/j.ygyno.2024.11.009. Epub 2024 Dec 6.
Evidence is limited in gynecologic cancers on best practices for clinical trial screening, but the risk of ineffective screening processes and subsequent under-enrollment introduces significant cost to patient, healthcare systems, and scientific advancement. Absence of a defined screening process makes determination of who and when to screen potential patients inconsistent allowing inefficiency and potential introduction of biases. This is especially germane as generative artificial intelligence (AI), and electronic health record (EHR) integration is applied to trial screening. Though often a requirement of cooperative groups such as the Cancer therapy Evaluation Program (CTEP), and/or the Commission on Cancer (CoC), there are no standard practice guidelines on best practices regarding screening and how best to track screening data.
The authors provided a review of current clinical trial screening practices and the effect on enrollment and trial activation across a variety of disease and practice sites. Established clinical trial screening practices and evidence supporting emerging strategies were reviewed and reported. Due to lack of published literature in gynecologic oncology, authors sought to survey the members of current rostered GOG sites to provide perspectives on clinical trial screening practices. Survey results showed a variety of screening practices. Most respondents participate in some type of manual screening process, where approximately 13 % also report incorporating AI or EHR integration. Over half (60 %) of sites track screening data to use for feasibility when opening new trials. The rapid increase in generative AI, EHR integration, and site agnostic screening initiatives could provide a significant opportunity to improve screening efficiency, translating to improved enrollment, but limitations and barriers remain.
关于妇科癌症临床试验筛查的最佳实践，证据有限，但筛查流程无效及后续入组不足的风险给患者、医疗系统和科学进步带来了巨大成本。缺乏明确的筛查流程使得确定哪些潜在患者以及何时进行筛查不一致，导致效率低下并可能引入偏差。随着生成式人工智能（AI）和电子健康记录（EHR）集成应用于试验筛查，这一点尤为重要。尽管这通常是癌症治疗评估计划（CTEP）和/或癌症委员会（CoC）等合作组织的要求，但关于筛查的最佳实践以及如何最好地跟踪筛查数据，尚无标准操作指南。
作者回顾了当前临床试验筛查实践及其对各种疾病和实践场所的入组及试验启动的影响。对既定的临床试验筛查实践以及支持新兴策略的证据进行了回顾和报告。由于妇科肿瘤学领域缺乏已发表的文献，作者试图对当前GOG登记站点的成员进行调查，以了解临床试验筛查实践的情况。调查结果显示了多种筛查实践。大多数受访者参与某种类型的手动筛查流程，其中约13%的人还报告纳入了AI或EHR集成。超过一半（60%）的站点跟踪筛查数据，以便在开展新试验时用于评估可行性。生成式AI、EHR集成和无站点筛查计划的迅速增加可能为提高筛查效率提供重大机遇，从而转化为更高的入组率，但限制和障碍仍然存在。
