Scholarlink

<p>Recurrence rate remains unsatisfactory among surgical resected hepatocellular carcinoma (HCC) patients with radical resection intention, and effective surveillance methods are lacking for post-operative recurrence. 436 HCC patients were selected for ultimate analyses. Significant features were extracted on favorable regions of interest (ROIs) of contrast-enhanced CT (CECT) and selected by least absolute shrinkage and selection operator (LASSO) method to construct radiomics signature. A novel CT radiomics-clinicopathological prediction model was constructed to evaluate 2-year recurrence-free survival (RFS) of resected HCC. Model discrimination was evaluated by area under the receiver operating characteristic (ROC) curve, the calibrated curves and decision curve analysis. Gene expressions of CENPA, FAM83D etc. were assessed by quantitative real-time PCR (q-PCR) for 48 randomly selected HCC patients, and correlation between radiomics signature and genomic characteristics was investigated. Radiomics signature was established based on the 20 early recurrence-related CT features. Microvascular invasion (MVI), alpha-fetoprotein (AFP), gamma-glutamyl transpeptidase to lymphocyte ratio (GLR) and radiomics signature were selected as independent predictors of early recurrence of HCC, and thereafter used to construct the novel nomogram which predicted 2-year RFS. The radiomics-clinicopathological combined model revealed favorable prediction ability with area under the ROC curve (AUC) of 0.744 and 0.821 to predict 2-year RFS in training and validation cohort. HCC patients were divided into three risk groups, with RFS difference between groups statistically significant. Similar results were observed in AFP-negative HCC cohort. Pearson correlation analyses revealed favorable relationship between radiomics signature and BCAT1 and CENPA. This study constructed a noninvasive and simple prediction model based on CT radiomics and clinical features. The novel radiomics-clinicopathological predictive model could evaluate post-operative early recurrence for HCC, and its significance remained in AFP-negative HCC.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Development of a CT radiomics and clinical feature combined model for predicting early recurrence of surgical resected hepatocellular carcinoma

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年2月25日

查看全文
  • Minjun Liao,
  • Naying Liao,
  • Shengjun Huo,
  • Xiaoxiao Wang,
  • Zilong Wang,
  • Baiyi Liu,
  • Xin Ai,
  • Kai Wang,
  • Feng Liu,
  • Yuanping Zhou,
  • Huiying Rao

摘要

Recurrence rate remains unsatisfactory among surgical resected hepatocellular carcinoma (HCC) patients with radical resection intention, and effective surveillance methods are lacking for post-operative recurrence. 436 HCC patients were selected for ultimate analyses. Significant features were extracted on favorable regions of interest (ROIs) of contrast-enhanced CT (CECT) and selected by least absolute shrinkage and selection operator (LASSO) method to construct radiomics signature. A novel CT radiomics-clinicopathological prediction model was constructed to evaluate 2-year recurrence-free survival (RFS) of resected HCC. Model discrimination was evaluated by area under the receiver operating characteristic (ROC) curve, the calibrated curves and decision curve analysis. Gene expressions of CENPA, FAM83D etc. were assessed by quantitative real-time PCR (q-PCR) for 48 randomly selected HCC patients, and correlation between radiomics signature and genomic characteristics was investigated. Radiomics signature was established based on the 20 early recurrence-related CT features. Microvascular invasion (MVI), alpha-fetoprotein (AFP), gamma-glutamyl transpeptidase to lymphocyte ratio (GLR) and radiomics signature were selected as independent predictors of early recurrence of HCC, and thereafter used to construct the novel nomogram which predicted 2-year RFS. The radiomics-clinicopathological combined model revealed favorable prediction ability with area under the ROC curve (AUC) of 0.744 and 0.821 to predict 2-year RFS in training and validation cohort. HCC patients were divided into three risk groups, with RFS difference between groups statistically significant. Similar results were observed in AFP-negative HCC cohort. Pearson correlation analyses revealed favorable relationship between radiomics signature and BCAT1 and CENPA. This study constructed a noninvasive and simple prediction model based on CT radiomics and clinical features. The novel radiomics-clinicopathological predictive model could evaluate post-operative early recurrence for HCC, and its significance remained in AFP-negative HCC.