Scholarlink

Background <p>This study aimed to establish and validate a predictive model for microsatellite instability (MSI) in colorectal cancer (CRC) patients by integrating CT-based extracellular volume (ECV) fraction, clinical parameters, and conventional imaging signs.</p> Methods <p>A retrospective cohort of CRC patients was divided into a training set (<i>n</i> = 155) and a validation set (<i>n</i> = 67). The ECV fraction was derived by incorporating patients’ hematocrit values and applying subtraction algorithms to pre-contrast and equilibrium phase images. The ECV fraction was then combined with clinical parameters and conventional imaging signs to construct two predictive models: a clinical imaging sign model and a hybrid model. Model performance was assessed using the area under the curve (AUC), net reclassification index (NRI), and integrated discrimination improvement (IDI).</p> Results <p>In both sets, the clinical imaging sign model demonstrated robust MSI prediction, with AUCs of 0.829 (training) and 0.792 (validation). Incorporating ECV fraction showed superior performance with corresponding AUCs of 0.867 and 0.848. Significant improvements in both NRI and IDI were observed when comparing the hybrid model to the clinical imaging sign model.</p> Conclusion <p>The combination of ECV fraction, clinical parameters, and conventional imaging signs of tumor lesions, when analyzed using logistic regression, enables accurate prediction of MSI status in CRC. This approach demonstrates the incremental predictive value contributed by ECV fraction in MSI assessment.</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

Incremental value of extracellular volume fraction based on CT for microsatellite status in colorectal cancer

  • 期刊论文

  • 发表日期：

    2025年7月4日

查看全文
  • Minhong Wang,
  • Lifang Fan,
  • Lixiang Zhou

摘要

Background

This study aimed to establish and validate a predictive model for microsatellite instability (MSI) in colorectal cancer (CRC) patients by integrating CT-based extracellular volume (ECV) fraction, clinical parameters, and conventional imaging signs.

Methods

A retrospective cohort of CRC patients was divided into a training set (n = 155) and a validation set (n = 67). The ECV fraction was derived by incorporating patients’ hematocrit values and applying subtraction algorithms to pre-contrast and equilibrium phase images. The ECV fraction was then combined with clinical parameters and conventional imaging signs to construct two predictive models: a clinical imaging sign model and a hybrid model. Model performance was assessed using the area under the curve (AUC), net reclassification index (NRI), and integrated discrimination improvement (IDI).

Results

In both sets, the clinical imaging sign model demonstrated robust MSI prediction, with AUCs of 0.829 (training) and 0.792 (validation). Incorporating ECV fraction showed superior performance with corresponding AUCs of 0.867 and 0.848. Significant improvements in both NRI and IDI were observed when comparing the hybrid model to the clinical imaging sign model.

Conclusion

The combination of ECV fraction, clinical parameters, and conventional imaging signs of tumor lesions, when analyzed using logistic regression, enables accurate prediction of MSI status in CRC. This approach demonstrates the incremental predictive value contributed by ECV fraction in MSI assessment.