文献速递肺结节与肺癌

人工智能在肺磨玻璃结节鉴别与诊断中的研究进展

[Artificial intelligence research advances in discrimination and diagnosis of pulmonary ground-glass nodules].
PubMed 文献 · Zhonghua Jie He He Hu Xi Za Zhi · 2024-01-01

本文综述人工智能(AI)特别是深度学习算法在肺磨玻璃结节(GGN)评估中的应用进展。AI通过分析复杂影像数据并结合临床信息,可辅助鉴别GGN良恶性、预测病理亚型及EGFR突变。但AI临床应用仍需广泛验证、影像协议标准化及算法可解释性提升。

要点

  • AI可辅助鉴别GGN良恶性并预测病理亚型,提高评估准确性
  • AI在预测GGN相关EGFR突变方面展现潜力,助力个体化治疗
  • AI临床落地仍需广泛验证、影像协议标准化及可解释性改进

核心结论

  • AI有潜力通过为临床医生提供更准确及时的诊断治疗信息来革新GGN管理,但尚需进一步研究与验证。

研究设计与样本

  • 本文为综述,汇总AI在GGN良恶性、病理学亚型和EGFR突变鉴别诊断中的研究进展。

主要发现

  • 深度学习算法可分析复杂影像数据并预测GGN良恶性、病理亚型及EGFR等基因突变。
  • 整合影像特征与临床数据的AI模型在鉴别良恶性GGN和预测特定病理亚型方面表现出较高准确性。
  • AI在预测EGFR突变方面展现潜力,该突变对肺癌个体化治疗决策至关重要。

关键数值与推荐

  • 肺癌约占全球所有癌症相关死亡的18%,5年生存率低于20%。
  • 早期肺癌(TNM分期IA1、IA2、IA3和IB期)生存率显著更高,凸显早期发现、诊断和治疗的重要性。

临床怎么用

  • 临床可探索将AI辅助诊断模型用于GGN影像检测及性质预测,以提高评估准确性与效率。
  • AI预测的病理亚型和EGFR突变信息可辅助制定个体化诊疗方案及随访管理策略。

局限性

  • AI应用仍需广泛验证研究。
  • 影像协议尚需标准化。
  • AI算法可解释性有待提高。

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

Lung cancer, which accounts for about 18% of all cancer-related deaths worldwide, has a dismal 5-year survival rate of less than 20%. Survival rates for early-stage lung cancers (stages IA1, IA2, IA3, and IB, according to the TNM staging system) are significantly higher, underscoring the critical importance of early detection, diagnosis, and treatment. Ground-glass nodules (GGNs), which are commonly seen on lung imaging, can be indicative of both benign and malignant lesions. For clinicians, accurately characterizing GGNs and choosing the right management strategies present significant challenges. Artificial intelligence (AI), specifically deep learning algorithms, has shown promise in the evaluation of GGNs by analyzing complex imaging data and predicting the nature of GGNs, including their benign or malignant status, pathological subtypes, and genetic mutations such as epidermal growth factor receptor (EGFR) mutations. By integrating imaging features and clinical data, AI models have demonstrated high accuracy in distinguishing between benign and malignant GGNs and in predicting specific pathological subtypes. In addition, AI has shown promise in predicting genetic mutations such as EGFR mutations, which are critical for personalized treatment decisions in lung cancer. While AI offers significant potential to improve the accuracy and efficiency of GGN assessment, challenges remain, such as the need for extensive validation studies, standardization of imaging protocols, and improving the interpretability of AI algorithms. In summary, AI has the potential to revolutionise the management of GGNs by providing clinicians with more accurate and timely information for diagnosis and treatment decisions. However, further research and validation are needed to fully realize the benefits of AI in clinical practice. 肺磨玻璃结节(GGN)作为常见的肺部影像学表现,可见于多种病理状态,如炎症、出血、肺腺癌等。临床诊疗中,主要通过分析其影像学特征,初步预测GGN的性质,进而选择最佳的个性化诊疗方案及随访管理策略。随着信息学技术的迅速发展与普及,人工智能(AI)通过深度学习、多维度分析病例样本,构建医学影像辅助诊断模型,目前正逐步应用于GGN影像检测及预测。AI的临床应用省时省力,同时也提高了GGN预测的准确性。本文对AI在GGN良恶性、病理学亚型和EGFR突变鉴别诊断中的研究进展作一综述。.

来源:Zhonghua Jie He He Hu Xi Za Zhi,查看原文。版权归原作者所有。

引用

  1. Li YJ, Wang Y, Qiu ZX, et al. [Artificial intelligence research advances in discrimination and diagnosis of pulmonary ground-glass nodules]. Zhonghua Jie He He Hu Xi Za Zhi. 2024. PMID: 38858209. doi:10.3760/cma.j.cn112147-20231214-00370

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。