Scholarlink

Purpose <p>To evaluate the value of the whole-tumor histogram model based on intravoxel incoherent motion diffusion-weighted imaging (IVIM-DWI) in diagnosing tumor deposits (TDs) in rectal cancer (RC) patients.</p> Methods <p>A total of 101 RC patients, 39 TD-positive and 62 TD-negative cases, were enrolled. The whole-tumor volume was obtained by manually outlining the lesion on IVIM-DWI slices where the tumor was visible. Eighteen histogram features were extracted from the ADC, D, D*, and f maps derived from IVIM-DWI. Multivariate binary logistic regression analysis was used to develop a combined model for predicting TD status. Diagnostic performance was evaluated using receiver operating characteristic (ROC) curve analysis and the area under the ROC curve (AUC). Internal validation was performed to evaluate model performance.</p> Results <p>ADC energy, ADC total energy, D energy, D kurtosis, D maximum, D range, D total energy, D* maximum, D* mean, D* range, D* mean absolute deviation, D* root mean squared, D* variance, f energy, and f total energy differed significantly between the TD-positive and TD-negative groups. The combined model incorporating D range, D* mean absolute deviation, magnetic resonance T stage (mrT), and tumor thickness showed superior diagnostic performance for TD prediction, with an AUC of 0.854 (95% CI, 0.769–0.916), sensitivity of 0.923 (95% CI, 0.75–0.978), specificity of 0.645 (95% CI, 0.519–0.819), and accuracy of 0.752 (95% CI, 0.663–0.861). The results of internal validation indicated that the logistic regression model demonstrated good predictive performance, with an AUC of 0.841 (95% CI: 0.821–0.853), an accuracy of 0.743 (95% CI: 0.713–0.772), and sensitivity and specificity of 0.650 (95% CI: 0.487–0.795) and 0.802 (95% CI: 0.694–0.903), respectively.</p> Conclusion <p>Whole-tumor histogram analysis derived from IVIM-DWI provides a novel method for diagnosing TDs. The combined histogram-based model may aid the preoperative evaluation of TDs in RC patients.</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 and validation: a whole-tumor histogram model based on intravoxel incoherent motion diffusion-weighted imaging for diagnosing tumor deposits in rectal cancer

  • 期刊论文

  • 发表日期：

    2025年12月4日

查看全文
  • Hongyu Zhao,
  • Yantong Sun,
  • Longxia Xu,
  • Haojie Sun,
  • Dmytro Pylypenko,
  • Zhengjun Dai,
  • Aiyin Li,
  • Gesheng Song

摘要

Purpose

To evaluate the value of the whole-tumor histogram model based on intravoxel incoherent motion diffusion-weighted imaging (IVIM-DWI) in diagnosing tumor deposits (TDs) in rectal cancer (RC) patients.

Methods

A total of 101 RC patients, 39 TD-positive and 62 TD-negative cases, were enrolled. The whole-tumor volume was obtained by manually outlining the lesion on IVIM-DWI slices where the tumor was visible. Eighteen histogram features were extracted from the ADC, D, D*, and f maps derived from IVIM-DWI. Multivariate binary logistic regression analysis was used to develop a combined model for predicting TD status. Diagnostic performance was evaluated using receiver operating characteristic (ROC) curve analysis and the area under the ROC curve (AUC). Internal validation was performed to evaluate model performance.

Results

ADC energy, ADC total energy, D energy, D kurtosis, D maximum, D range, D total energy, D* maximum, D* mean, D* range, D* mean absolute deviation, D* root mean squared, D* variance, f energy, and f total energy differed significantly between the TD-positive and TD-negative groups. The combined model incorporating D range, D* mean absolute deviation, magnetic resonance T stage (mrT), and tumor thickness showed superior diagnostic performance for TD prediction, with an AUC of 0.854 (95% CI, 0.769–0.916), sensitivity of 0.923 (95% CI, 0.75–0.978), specificity of 0.645 (95% CI, 0.519–0.819), and accuracy of 0.752 (95% CI, 0.663–0.861). The results of internal validation indicated that the logistic regression model demonstrated good predictive performance, with an AUC of 0.841 (95% CI: 0.821–0.853), an accuracy of 0.743 (95% CI: 0.713–0.772), and sensitivity and specificity of 0.650 (95% CI: 0.487–0.795) and 0.802 (95% CI: 0.694–0.903), respectively.

Conclusion

Whole-tumor histogram analysis derived from IVIM-DWI provides a novel method for diagnosing TDs. The combined histogram-based model may aid the preoperative evaluation of TDs in RC patients.