文献速递头颈外科

影像组学在喉癌管理中的应用:文献范围综述

The Application of Radiomics in Laryngeal Cancer Management: A Scoping Review of the Literature.
PubMed 文献 · Laryngoscope · 2025-10-16

影像组学模型在预测总生存、无进展生存、局部控制、治疗反应和分期方面优于临床或传统影像模型,联合临床病理变量时准确性最高。但影像协议、特征提取软件和建模算法异质性大,且缺乏外部验证,阻碍常规临床采用。

要点

  • 影像组学模型在喉癌生存、局控、治疗反应和分期预测中优于临床或传统影像模型。
  • 影像组学特征联合临床病理变量可获得最高预测准确性,深度学习进一步提升性能。
  • 影像协议、特征提取软件和建模算法异质性大且缺乏外部验证,限制临床常规应用。

核心结论

  • 影像组学在提高喉癌分期准确性、预后判断和个体化管理方面展现显著潜力,但当前局限性阻碍其常规临床采用。

主要发现

  • 主要目标是开发总生存、无进展生存、局部控制、治疗反应和分期的预测模型。
  • 影像组学模型在预测关键结局方面持续优于临床或传统影像模型。
  • 影像组学特征联合临床病理变量的模型准确性最高。
  • 深度学习较传统影像组学进一步改善预测性能。

关键数值与推荐

  • 检索Medline和PubMed至2025年5月。
  • 采用RQS框架评估方法学质量。

临床怎么用

  • 影像组学模型可作为喉癌分期、预后和治疗反应预测的辅助工具,联合临床病理变量可提高准确性。
  • 因异质性和缺乏外部验证,目前尚不能替代常规临床决策,需谨慎解读。

局限性

  • 影像协议、特征提取软件和模型开发算法存在异质性。
  • 缺乏外部验证。
  • 这些局限性目前阻碍影像组学的常规临床采用。

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

OBJECTIVE: Laryngeal cancer is a significant head and neck malignancy, whose prevalence is increasing. Radiomics consists of high-dimensional and reproducible quantitative data which can serve as biomarkers with strong correlations to tumor staging, prognosis, and survival outcomes. The aim of this scoping review is to summarize the knowledge from the existing literature regarding the value of radiomics in laryngeal cancer management. DATA SOURCES: Eligible articles were identified by a search of the Medline and PubMed bibliographical databases for the period up to May 2025. REVIEW METHODS: This scoping review was performed in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. The methodological quality of the included studies was assessed using the Radiomics Quality Score (RQS) framework. RESULTS: Thirty studies were deemed eligible, including a total of 3503 patients with laryngeal cancer. Primary objectives were the development of predictive models for overall survival, progression-free survival, local control, treatment response, and staging. Across studies, radiomics-based models consistently outperformed clinical or conventional imaging-based models in predicting key outcomes. Models combining radiomic features with clinicopathological variables achieved the highest accuracy. Recent advances include the use of deep learning, which further improved predictive performance compared to traditional radiomics. However, common limitations include heterogeneity in imaging protocols, feature extraction software, and model development algorithms, as well as a lack of external validation. CONCLUSION: Despite several limitations that currently hinder routine clinical adoption of radiomics, this rapidly evolving field demonstrates significant potential to improve staging accuracy, prognostication, and personalized management in laryngeal cancer.

来源:Laryngoscope,查看原文。版权归原作者所有。

引用

  1. Papadopoulou AM, Matsopoulos GK, Kouloulias V, et al. The Application of Radiomics in Laryngeal Cancer Management: A Scoping Review of the Literature. Laryngoscope. 2025. PMID: 41099387. doi:10.1002/lary.70197

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。