Scholarlink

Background <p>Accurate preoperative identification of muscle-invasive bladder cancer (MIBC) is critical for clinical decision-making but remains challenging. In this study, we aimed to develop and validate a novel predictive model that integrates contrast-enhanced ultrasound (CEUS) radiomics features with the Vesical Imaging Reporting and Data System (VI-RADS) score to improve the preoperative discrimination of MIBC.</p> Methods <p>In this retrospective, single-center study, we enrolled 116 consecutive patients with pathologically confirmed bladder tumor who underwent pre-operative CEUS between May 2015 and October 2024. Patients were randomly allocated to training and validation cohorts in a 7:3 ratio. The CEUS frame at peak tumor enhancement was selected for analysis. Tumors were manually segmented on the largest cross-sectional plane using 3D-Slicer to generate regions of interest. Radiomic features were extracted using PyRadiomics. After sequential feature reduction and model selection, radiomic predictors were integrated with clinically significant variables to construct a combined model.</p> Results <p>Three CEUS radiomic features were identified as relevant to MIBC. Among clinical variables, the VI-RADS score (<i>P</i>=0.04, odds ratio 1.89) was independently associated with MIBC. Of the five evaluated classifiers, XGBoost achieved the best performance for both the radiomic (area under the curve 0.86) and the combined (area under the curve 0.96) models. Finally, SHapley Additive exPlanations analysis was performed to interpret the model, and calibration curves and decision curve analysis were employed to evaluate its accuracy and clinical utility.</p> Conclusion <p>This study developed and conducted preliminary validation of a model integrating CEUS radiomics and the VI-RADS score, demonstrating its feasibility and potential for preoperative MIBC prediction.</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

A Novel Model for Muscle-Invasive Bladder Cancer Diagnosis Using Contrast-Enhanced Ultrasound

  • 期刊论文

  • 发表日期：

    2026年2月4日

查看全文
  • Qiping Liu,
  • Xuwei Lu,
  • Chang He,
  • Jiangang Chen,
  • Minke He,
  • Shiyuan Zhang,
  • Ying Zhu,
  • Jiawen Wu,
  • Bin Hu,
  • Jiaqi Huang

摘要

Background

Accurate preoperative identification of muscle-invasive bladder cancer (MIBC) is critical for clinical decision-making but remains challenging. In this study, we aimed to develop and validate a novel predictive model that integrates contrast-enhanced ultrasound (CEUS) radiomics features with the Vesical Imaging Reporting and Data System (VI-RADS) score to improve the preoperative discrimination of MIBC.

Methods

In this retrospective, single-center study, we enrolled 116 consecutive patients with pathologically confirmed bladder tumor who underwent pre-operative CEUS between May 2015 and October 2024. Patients were randomly allocated to training and validation cohorts in a 7:3 ratio. The CEUS frame at peak tumor enhancement was selected for analysis. Tumors were manually segmented on the largest cross-sectional plane using 3D-Slicer to generate regions of interest. Radiomic features were extracted using PyRadiomics. After sequential feature reduction and model selection, radiomic predictors were integrated with clinically significant variables to construct a combined model.

Results

Three CEUS radiomic features were identified as relevant to MIBC. Among clinical variables, the VI-RADS score (P=0.04, odds ratio 1.89) was independently associated with MIBC. Of the five evaluated classifiers, XGBoost achieved the best performance for both the radiomic (area under the curve 0.86) and the combined (area under the curve 0.96) models. Finally, SHapley Additive exPlanations analysis was performed to interpret the model, and calibration curves and decision curve analysis were employed to evaluate its accuracy and clinical utility.

Conclusion

This study developed and conducted preliminary validation of a model integrating CEUS radiomics and the VI-RADS score, demonstrating its feasibility and potential for preoperative MIBC prediction.