文献速递肿瘤 PET/CT

18F-FDG PET/CT影像组学预测极早期及早期肝细胞癌微血管侵犯与预后

Radiomics analysis of [18F]FDG PET/CT for microvascular invasion and prognosis prediction in very-early- and early-stage hepatocellular carcinoma.
PubMed 文献 · Eur J Nucl Med Mol Imaging · 2021-01-08

该回顾性研究纳入80例BCLC 0-A期肝细胞癌患者,术前接受18F-FDG PET/CT检查,随机分为训练集和验证集,利用LASSO回归筛选纹理特征构建影像组学列线图。预测微血管侵犯的模型AUC为0.891(95%CI: 0.799-0.984),预测无病生存的D-PET/CT模型C-index为0.831(95%CI: 0.761-0.900),优于简单临床模型。

要点

  • 18F-FDG PET/CT影像组学特征与MVI状态显著相关(训练集P=0.001)。
  • M-PET/CT模型预测MVI的AUC达0.891,判别与校准良好。
  • D-PET/CT模型整合影像组学特征与INR、TB,预测DFS的C-index为0.831。

核心结论

  • 新构建的18F-FDG PET/CT影像组学特征是完全极早期及早期HCC患者MVI和DFS的独立生物标志物。
  • 整合影像组学特征与临床危险因素的PET/CT列线图对个体化DFS估计表现更优。

研究设计与样本

  • 回顾性研究纳入80例BCLC 0-A期HCC患者,术前接受18F-FDG PET/CT,随机分为训练集和验证集。
  • 使用Lifex软件提取纹理特征,LASSO回归筛选MVI和DFS的预测特征,构建影像组学列线图并验证。

主要发现

  • 预测MVI的影像组学特征由PET的5个纹理特征和CT的6个纹理特征组成。
  • 该特征与训练集MVI状态显著相关(P=0.001),临床特征均非MVI的独立预测因子(P>0.05)。
  • M-PET/CT模型训练集AUC为0.891(95%CI: 0.799-0.984),判别与校准良好。
  • D-PET/CT模型整合影像组学特征与INR和TB,C-index为0.831(95%CI: 0.761-0.900)。
  • 决策曲线分析显示D-PET/CT模型较简单临床模型提供更大净获益。

关键数值与推荐

  • MVI预测模型M-PET/CT的AUC为0.891(95%CI: 0.799-0.984)。
  • DFS预测模型D-PET/CT的C-index为0.831(95%CI: 0.761-0.900)。
  • 影像组学特征与MVI状态在训练集显著相关(P=0.001)。
  • 临床特征对MVI均无独立预测价值(P>0.05)。

临床怎么用

  • 对于极早期及早期HCC患者,术前18F-FDG PET/CT影像组学分析可辅助无创预测MVI状态。
  • 整合影像组学特征与INR、TB的列线图可用于个体化无病生存估计,辅助精准医学决策。

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

UNLABELLED: As a reliable preoperative predictor for microvascular invasion (MVI) and disease-free survival (DFS) is lacking, we developed a radiomics nomogram of [18F]FDG PET/CT to predict MVI status and DFS in patients with very-early- and early-stage (BCLC 0, BCLC A) hepatocellular carcinoma (HCC). METHODS: Patients (N = 80) with BCLC0-A HCC who underwent [18F]FDG PET/CT before surgery were enrolled in this retrospective study and were randomized to a training cohort and a validation cohort. Texture features from patients obtained using Lifex software in the training cohort were subjected to LASSO regression to select the most useful predictive features of MVI and DFS. Then, the radiomics nomogram was constructed using the radiomics signature and clinical features and further validated. RESULTS: To predict MVI, the [18F]FDG PET/CT radiomics signature consisted of five texture features from the PET and six texture features from CT. The signature was significantly associated with MVI status in the training cohort (P = 0.001). None of the clinical features was independent predictors for MVI status (P > 0.05). The area under the curve value of the M-PET/CT model was 0.891 (95% CI: 0.799-0.984) in the training cohort and showed good discrimination and calibration. To predict DFS, the [18F]FDG PET/CT radiomics nomogram (D-PET/CT model) and a clinicopathologic nomogram were built in the training cohort. The D-PET/CT model, which integrated the D-PET/CT radiomics signature with INR and TB, provided better predictive performance (C-index: 0.831, 95% CI: 0.761-0.900) and larger net benefits than the simple clinical model, as determined by decision curve analyses. CONCLUSION: The newly developed [18F]FDG PET/CT radiomics signature was an independent biomarker for the estimation of MVI and DFS in patients with very-early- and early-stage HCC. Moreover, PET/CT nomogram, which incorporated the radiomics signature of [18F]FDG PET/CT and clinical risk factors in patients with very-early- and early-stage HCC, performed better for individualized DFS estimation, which might enable a step forward in precise medicine.

来源:Eur J Nucl Med Mol Imaging,查看原文。版权归原作者所有。

引用

  1. Li Y, Zhang Y, Fang Q, et al. Radiomics analysis of [18F]FDG PET/CT for microvascular invasion and prognosis prediction in very-early- and early-stage hepatocellular carcinoma. Eur J Nucl Med Mol Imaging. 2021. PMID: 33416951. doi:10.1007/s00259-020-05119-9

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。