Scholarlink

Objective <p>This study aimed to investigate the imaging characteristics of Microphthalmia Transcription Factor (MIT) Family Translocation Renal Cell Carcinoma (MIT-RCC), assess the relationship between CT-based radiomics and MIT-RCC, and develop a predictive model to improve preoperative non-invasive diagnosis.</p> Materials and methods <p>A retrospective analysis was conducted using preoperative computed tomography (CT) images and clinical data from 746 RCC patients (77 MIT-RCC and 669 other RCC subtypes) across multiple centers. Regions of interest (ROIs) in both tumor and normal renal parenchymal during non-contrast and arterial phases were manually segmented using ITK-SNAP. Radiomics features were extracted via Python, and the Least Absolute Shrinkage and Selection Operator (LASSO) method was applied for feature selection. A combined model incorporating radiomics scores and clinical predictors (age, stage) was developed using logistic regression and visualized as a nomogram. Patients from two centers formed the training cohort (<i>n</i> = 539), and three other centers served as the validation cohort (<i>n</i> = 207).</p> Results <p>From 1316 features per ROI, 22 tumor and 24 normal renal parenchymal features were selected. The nomogram integrating Tumor_Radscores, Kidney_Radscores, age, and clinical stage showed strong predictive performance in the training cohort (area under the curve [AUC] = 0.952, 95% confidence interval [CI]: 0.927–0.971) and validation cohort (AUC = 0.914, 95% CI: 0.879–0.942), with good calibration and clinical utility.</p> Conclusion <p>By integrating radiomics and clinical data, the proposed nomogram achieved accurate, non-invasive prediction of MIT-RCC, offering valuable support for preoperative diagnosis and personalized treatment planning.</p> Graphical Abstract <p></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

Radiomics analysis of CT imaging for predicting MIT family translocation renal cell carcinoma: a multicenter retrospective clinical study

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年6月20日

查看全文
  • Shi-Wei Lin,
  • Xiao-Hui Wu,
  • Wen-Qi Liu,
  • Dong-Ning Chen,
  • Yan Lin,
  • Zheng-Sheng Liu,
  • Ji-yuan Wang,
  • Hu-bin Yin,
  • Xue-Yi Xue,
  • Ning Xu,
  • Shao-Hao Chen

摘要

Objective

This study aimed to investigate the imaging characteristics of Microphthalmia Transcription Factor (MIT) Family Translocation Renal Cell Carcinoma (MIT-RCC), assess the relationship between CT-based radiomics and MIT-RCC, and develop a predictive model to improve preoperative non-invasive diagnosis.

Materials and methods

A retrospective analysis was conducted using preoperative computed tomography (CT) images and clinical data from 746 RCC patients (77 MIT-RCC and 669 other RCC subtypes) across multiple centers. Regions of interest (ROIs) in both tumor and normal renal parenchymal during non-contrast and arterial phases were manually segmented using ITK-SNAP. Radiomics features were extracted via Python, and the Least Absolute Shrinkage and Selection Operator (LASSO) method was applied for feature selection. A combined model incorporating radiomics scores and clinical predictors (age, stage) was developed using logistic regression and visualized as a nomogram. Patients from two centers formed the training cohort (n = 539), and three other centers served as the validation cohort (n = 207).

Results

From 1316 features per ROI, 22 tumor and 24 normal renal parenchymal features were selected. The nomogram integrating Tumor_Radscores, Kidney_Radscores, age, and clinical stage showed strong predictive performance in the training cohort (area under the curve [AUC] = 0.952, 95% confidence interval [CI]: 0.927–0.971) and validation cohort (AUC = 0.914, 95% CI: 0.879–0.942), with good calibration and clinical utility.

Conclusion

By integrating radiomics and clinical data, the proposed nomogram achieved accurate, non-invasive prediction of MIT-RCC, offering valuable support for preoperative diagnosis and personalized treatment planning.

Graphical Abstract