指南与共识头颈与中枢神经肿瘤

人工智能用于儿童神经肿瘤疗效评估:挑战、机遇与临床转化建议

Artificial Intelligence for Response Assessment in Pediatric Neuro-Oncology (AI-RAPNO), part 2: challenges, opportunities, and recommendations for clinical translation.
指南 / 共识 · Lancet Oncol · 2025-01-01

本文为AI-RAPNO小组发布的政策综述,旨在探讨将人工智能整合入RAPNO疗效评估框架的挑战与机遇。文章指出影像协议差异、标注数据稀缺及监管伦理是主要障碍,并提出标准化影像协议、稳健验证框架及基础设施建设等建议,以提升AI可信度与泛化性,推动其在儿童神经肿瘤临床试验中的临床转化。

要点

  • AI整合入RAPNO框架可增强疗效评估的定量化与数据驱动能力。
  • 临床转化面临影像协议差异、标注数据稀缺及监管伦理等挑战。
  • 建议标准化影像协议、建立验证框架及基础设施以支持AI就绪。

核心结论

  • AI有望增强儿童神经肿瘤疗效评估的定量化与数据驱动方法,但需解决影像协议差异、数据稀缺及监管伦理问题以实现可信临床转化。

研究设计与样本

  • 由AI-RAPNO小组主导的政策综述,基于RAPNO框架讨论AI整合的挑战与建议。

主要发现

  • AI在疗效评估、多模态整合及临床试验合成对照组方面具有潜力。
  • 影像协议差异、标注数据集稀缺及监管伦理考量是主要实施障碍。

关键数值与推荐

  • 建议标准化影像协议、建立稳健验证框架及支持AI就绪的基础设施。

临床怎么用

  • 临床研究中应优先采用标准化影像协议并构建验证框架,以提升AI工具的可信度与泛化性。

以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。

原文

The Response Assessment in Pediatric Neuro-Oncology (RAPNO) criteria provide an important framework for evaluating treatment efficacy and tumour progression in clinical studies of paediatric brain tumours. As artificial intelligence (AI) rapidly transforms clinical practice, integrating AI into the RAPNO framework presents a unique opportunity to enhance quantitative, data-driven approaches for response assessment. However, successful clinical implementation faces challenges, including variability in imaging protocols, scarce annotated datasets, and regulatory and ethical considerations. To address these barriers, this Policy Review, led by the AI for Assessment in Pediatric Neuro-Oncology (AI-RAPNO) subcommittee, outlines key challenges and proposes recommendations to improve AI trustworthiness, generalisability, and implementation in paediatric neuro-oncology. We highlight the potential of AI for response assessment, multimodal integration, and synthetic control groups in clinical trials. Our recommendations emphasise the need for standardised imaging protocols, robust validation frameworks, and infrastructure to support AI readiness in clinical studies. By addressing these needs, AI-RAPNO aims to bridge the gap between AI research and clinical application, ensuring reliable and actionable AI-driven tools for paediatric neuro-oncology.

来源:Lancet Oncol,查看原文。版权归原作者所有。

引用

  1. Kazerooni AF, Familiar AM, Aboian M, et al. Artificial Intelligence for Response Assessment in Pediatric Neuro-Oncology (AI-RAPNO), part 2: challenges, opportunities, and recommendations for clinical translation. Lancet Oncol. 2025. PMID: 41167228. doi:10.1016/S1470-2045(25)00489-9

供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。