文献速递腹部MR

机器学习预测肝硬化患者MRI上HCC发生部位的概念验证研究

Machine learning modeling to predict HCC locations in cirrhotic patients undergoing MRI - a proof-of-concept study.
PubMed 文献 · Abdom Radiol (NY) · 2026-07-13

本研究回顾性纳入121例肝硬化且后续经MRI确诊HCC的患者,利用HCC确诊前18个月内的筛查MRI提取影像组学特征。研究训练并验证了逻辑回归、随机森林和XGBoost模型,以预测未来发生HCC的肝实质区域。结果显示机器学习模型在手动标注数据上AUC最高可达0.80,且在可见前驱病变区域预测性能更优,提示此类方法具有提前识别高风险肝组织的可行性。

要点

  • 对LR-3等可见前驱病变区域,随机森林模型预测未来HCC的AUC可达0.86,阅片时应重点关注此类区域。
  • 约30%患者的热图在后续确诊HCC部位显示概率增高信号,但尚未达到可替代常规判读的程度。
  • 模型在手动标注数据上表现最佳,提示高质量标注对影像组学研究至关重要。

核心结论

  • 机器学习基于MRI影像组学可在HCC影像学确诊前识别肝硬化肝实质中未来发生HCC的高风险区域,尤其在存在前驱病变时预测性能更优。

研究设计与样本

  • 回顾性研究,纳入121例2008至2023年间行MRI且有HCC诊断的肝硬化患者,平均年龄65±9.56岁,其中99例为男性。
  • 分析对象为确诊HCC前18个月内阴性筛查MRI及随后确诊LR-5病灶的阳性筛查MRI,在筛查MRI上手动或自动标注肝组织VOI。
  • 从VOI中利用PyRadiomics提取影像组学特征,训练并验证逻辑回归、随机森林和XGBoost三种模型,采用四种手动/自动标注组合。
  • 后续随访确认为LR-5的HCC病灶位置被映射回筛查MRI作为恶性组织标注,其余为非恶性肝组织。

关键影像表现/征象

  • 在36/121例患者(30%)中,探索性体素级热图在随访影像上后续出现HCC的区域显示视觉上增高的概率信号。
  • 具有可见前驱病变区域(LR-3病灶)的预测性能优于无可见前驱区域的预测性能,提示前驱病变本身是重要影像征象。

诊断标准与关键数值

  • 在手动标注训练和验证条件下,逻辑回归、随机森林和XGBoost模型的最佳AUC分别为0.75(95%CI 0.72-0.78)、0.80(95%CI 0.78-0.81)和0.79(95%CI 0.77-0.81)。
  • 对LR-3病灶区域的预测中,随机森林模型AUC为0.86(95%CI 0.84-0.88),而无可见前驱区域时AUC为0.66(95%CI 0.63-0.70)。
  • 纳入标准为肝硬化患者在18个月内由阴性筛查MRI进展为阳性筛查MRI并确诊LR-5病灶。

临床应用与阅片要点

  • 对存在LI-RADS LR-3类结节或前驱病变的肝硬化患者,阅片时可意识到该区域未来发展为HCC的风险可能升高。
  • 当前机器学习热图仅为概念验证,不能作为临床诊断依据,实际报告仍应依据LI-RADS标准进行分类。
  • 影像组学模型的预测效能受标注方式影响,手动标注优于自动标注,提示研究应用中需保证标注质量。

局限性

  • 回顾性单中心研究设计,样本量为121例,模型性能的泛化能力未经外部验证。
  • 仅30%患者的热图在后续HCC部位显示视觉上增高的概率信号,提示模型空间定位能力仍较有限。

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: Current hepatocellular carcinoma (HCC) surveillance guidelines rely on manually defined LI-RADS (Liver Imaging Reporting and Data System) features rather than imaging data analysis. This study evaluates the feasibility of machine-learning (ML)-based image analysis frameworks to identify and localize hepatic parenchyma at elevated risk for HCC.

METHODS: In this retrospective study, cirrhotic patients with HCC diagnosis undergoing MRI between 2008 and 2023 were included. The analysis included negative screening MRI preceding a positive screening MRI confirming a LR-5 lesion within 18 months. Volume-of-interest (VOI) annotations of 'non-malignant' and 'malignant' liver tissue on the screening MRI were manually or automatically placed, mapped from future HCC lesion on positive screening MRI. Radiomics were extracted from these VOIs using PyRadiomics. Logistic regression (LR), random forest (RF), and eXtreme Gradient Boosting (XGB) models were trained and validated across four manual/automatic annotation combinations. Exploratory voxel-level heatmaps were generated to visualize high-risk HCC areas. Model performance was summarized using median values and 95% non-parametric confidence intervals.

RESULTS: 121 patients (65 ± 9.56; 99 men) were included. The best model performances for LR, RF, and XGB were achieved when training and validating on manual annotations: (AUC [95% CI]: 0.75 [0.72-0.78]; 0.80 [0.78-0.81]; 0.79 [0.77-0.81], respectively). Predictions from LR-3 lesions outperformed regions without visible precursor territory (RF AUC [95% CI]: 0.86 [0.84-0.88] vs. 0.66 [0.63-0.70]). In 36 out of 121 patients (30%), heatmaps showed visually increased probability signals in the region where HCC subsequently became visible on follow-up imaging.

CONCLUSION: The results demonstrate the feasibility of ML-based MRI analysis for identifying liver regions at increased risk for HCC development prior to radiologic diagnosis, particularly in the presence of precursor lesions.

来源:Abdom Radiol (NY),查看原文。版权归原作者所有。

引用

  1. Gaddum, Zeevi, Dai, et al. Machine learning modeling to predict HCC locations in cirrhotic patients undergoing MRI - a proof-of-concept study. Abdom Radiol (NY). 2026. PMID: 42440113. doi:10.1007/s00261-026-05653-4

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。