文献速递胸部技术与安全CT

用于术前预测cT1期肺腺癌淋巴结转移的深度学习模型:一项多中心外部验证研究

A Deep Learning Model for Preoperative Prediction of Lymph Node Metastasis in cT1-Stage Lung Adenocarcinoma: A Multicenter External Validation Study.
PubMed 文献 · Acad Radiol · 2025-07-11

这项多中心回顾性研究纳入8家机构2503例cT1期肺腺癌患者共2568个结节,开发并验证了基于残差网络的ResLNM深度学习模型用于术前预测淋巴结转移。ResLNM与影像组学模型均显著优于语义模型,ResLNM在外部验证集-1中准确率优于影像组学模型。将影像学纵隔肿大淋巴结状态(短轴直径≥10mm)整合后预测性能进一步提升。

要点

  • ResLNM深度学习模型在内部测试集AUC为0.85,外部验证集-1为0.82、集-2为0.71。
  • ResLNM与影像组学模型AUC范围分别为0.71-0.85和0.72-0.84,均显著优于语义模型的0.58-0.74。
  • 整合影像学纵隔肿大淋巴结状态(短轴直径≥10mm)后,测试集及两个验证队列AUC分别达0.88、0.89和0.76。

核心结论

  • ResLNM模型为cT1期肺腺癌术前淋巴结转移预测提供了临床可行的工具,性能优于传统语义和影像组学方法,整合常规可得的淋巴结大小标准后可进一步优化,有望改善手术决策并减少过度治疗。

研究设计与样本

  • 多中心回顾性研究,纳入8家机构2503例患者共2568个经病理证实的cT1期肺腺癌结节;6个中心1994例患者2059个结节随机分为训练队列(1600例患者1664个结节)和内部测试队列(394例患者395个结节),另设两个独立外部验证队列(集-1:283例患者283个结节;集-2:226例患者226个结节)。开发了语义模型(含毛刺征、胸膜牵拉、空气支气管征和空泡征)、影像组学模型和基于残差网络的ResLNM深度学习模型,采用受试者工作特征曲线下面积(AUC)和决策曲线分析(DCA)评估诊断性能。

关键影像表现/征象

  • 语义模型纳入的征象包括毛刺征、胸膜牵拉、空气支气管征和空泡征。
  • ResLNM与影像组学模型预测淋巴结转移的AUC范围分别为0.71-0.85和0.72-0.84,语义模型为0.58-0.74。

诊断标准与关键数值

  • ResLNM在内部测试集AUC为0.85(95%CI:0.81-0.89),影像组学模型为0.84(95%CI:0.80-0.88),两者差异无统计学意义(P=0.624)。
  • ResLNM在外部验证集-1的AUC为0.82(95%CI:0.76-0.88),优于影像组学模型的0.77(95%CI:0.71-0.83),P=0.039。
  • ResLNM在外部验证集-2的AUC为0.71(95%CI:0.63-0.79),影像组学模型为0.72(95%CI:0.64-0.80),两者差异无统计学意义(P=0.472)。
  • 将ResLNM与影像学纵隔肿大淋巴结状态(短轴直径≥10mm)整合后,测试集AUC为0.88(95%CI:0.85-0.92),外部验证集-1为0.89(95%CI:0.84-0.94),外部验证集-2为0.76(95%CI:0.68-0.83)。

临床应用与阅片要点

  • ResLNM模型基于术前影像预测cT1期肺腺癌淋巴结转移,可作为临床可行工具,性能优于传统语义和影像组学方法。
  • 将影像学纵隔肿大淋巴结状态(短轴直径≥10mm)与ResLNM整合可显著提升预测性能。
  • ResLNM与语义或影像组学特征整合未带来增量价值(P>0.05)。
  • 决策曲线分析证实ResLNM具有临床优越性。

局限性

  • 本文为多中心回顾性研究。
  • ResLNM与影像组学模型在内部测试集和外部验证集-2中性能相当,差异无统计学意义(P分别为0.624和0.472)。
  • 整合ResLNM与语义或影像组学特征未提供增量价值(P>0.05)。
  • ResLNM整合淋巴结大小标准后的性能提升仍需进一步研究。

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)。

英文原文(摘要)· 点击展开

PURPOSE: To develop and validate a deep learning (DL) model for preoperative prediction of lymph node metastasis (LNM) in clinical T1-stage lung adenocarcinoma (LUAD), and to compare its performance with conventional semantic and radiomics signatures.

METHODS: This multicenter retrospective study enrolled 2503 patients with 2568 pathologically confirmed cT1-stage LUAD nodules from eight institutions. Data from six centers (1994 patients/2059 nodules) were randomly divided into training (1600 patients/1664 nodules) and internal test (394 patients/395 nodules) cohorts. Two independent external validation cohorts (Set-1: 283 patients/nodules; Set-2: 226 patients/nodules) were included. Three predictive models were developed as follows: 1) a semantic model incorporating spiculation, pleural traction, air bronchogram, and vacuole signs; 2) a radiomics model; and 3) ResLNM-a residual network-based DL model. Diagnostic performance was evaluated using the area under the receiver operating characteristic curve (AUC) and decision curve analysis (DCA).

RESULTS: Both ResLNM and radiomics models significantly outperformed the semantic model in predicting LNM (AUC range: 0.71-0.85 and 0.72-0.84 vs. 0.58-0.74, respectively; P<0.05). While ResLNM demonstrated comparable performance to the radiomics model in the internal test set (AUC: 0.85, 0.81-0.89 vs. 0.84, 0.80-0.88; P=0.624) and external validation set-2 (0.71, 0.63-0.79 vs. 0.72, 0.64-0.80; P=0.472), it achieved superior accuracy in external validation set-1 (0.82, 0.76-0.88 vs. 0.77, 0.71-0.83; P=0.039). DCA confirmed the clinical superiority of ResLNM. Notably, integrating ResLNM with either semantic or radiomics signatures provided no incremental value (P>0.05), whereas incorporating radiologically mediastinal enlarged lymph node status (short-axis diameter ≥10mm) significantly enhanced predictive performance, achieving AUCs of 0.88 (95%CI:0.85-0.92), 0.89 (95%CI:0.84-0.94), and 0.76 (95%CI:0.68-0.83) in the test set and two validation cohorts, respectively.

CONCLUSION: The ResLNM model provides a clinically feasible tool for preoperative LNM prediction in cT1-stage LUAD, outperforming conventional semantic and radiomics approaches. Its performance can be further optimized by integrating routinely available lymph node size criteria, offering potential to refine surgical decision-making and reduce overtreatment.

来源:Acad Radiol,查看原文。版权归原作者所有。

引用

  1. Yan Q, Wei Y, Cheng Z, et al. A Deep Learning Model for Preoperative Prediction of Lymph Node Metastasis in cT1-Stage Lung Adenocarcinoma: A Multicenter External Validation Study. Acad Radiol. 2025. PMID: 40651922. doi:10.1016/j.acra.2025.06.001

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。