Radiomics-based interpretable machine learning model from multiphasic CT imaging for predicting pathological grade in upper tract urothelial carcinoma: a multicenter study - Report - MDSpire

Radiomics-based interpretable machine learning model from multiphasic CT imaging for predicting pathological grade in upper tract urothelial carcinoma: a multicenter study

  • By

  • Zhanpeng Yuan

  • Yuhua Mei

  • Xiang Peng

  • Zongjie Wei

  • Yingjie Xv

  • Bangxin Xiao

  • Mingzhao Xiao

  • June 23, 2026

  • 0 min

Share

Clinical Report: Development of an Interpretable Machine Learning Model Utilizing Radiomics

Overview

This study developed a machine learning model using radiomic features from CT imaging to predict the pathological grade of upper tract urothelial carcinoma (UTUC).

Background

Upper tract urothelial carcinoma (UTUC) is a rare but aggressive cancer, accounting for 5%-10% of urothelial tumors. Accurate preoperative assessment of tumor grade is crucial for treatment planning.

Data Highlights

ModelAUC (Training)Sensitivity (Training)Specificity (Training)AUC (Testing)Sensitivity (Testing)Specificity (Testing)
LGBM0.94584.5%91.2%0.82971.8%77.4%

Key Findings

  • The LGBM model achieved the highest AUC of 0.945 in the training dataset.
  • In the testing dataset, the LGBM model maintained an AUC of 0.829.
  • Key predictive features were derived from the venous and arterial phases of CTU.
  • SHAP analysis enhanced the interpretability of the model.
  • Feature selection involved various statistical methods including LASSO regression.

Clinical Implications

The radiomics-driven machine learning model provides a non-invasive method for predicting the pathological grade of UTUC.

Conclusion

The developed machine learning model offers a tool for improving diagnostic precision and treatment planning.

Related Resources & Content

  1. European Radiology, 2023 -- An interpretable radiomics model for predicting the pathological grading of pancreatic neuroendocrine tumors
  2. Frontiers in Oncology, 2026 -- Radiomics and deep learning in upper tract urothelial carcinoma: advancing preoperative risk stratification and clinical decision-making
  3. conexiant -- PET/CT and MRI Model May Help Stratify Prostate Cancer Risk
  4. European Radiology, 2024 -- Utilizing Radiomics and Machine Learning for the Evaluation of Renal Tumor Subtypes via Multiphase CT in a Multicenter Study
  5. EAU Guidelines on Upper Urinary Tract Urothelial Carcinoma, 2025
  6. NCCN Guidelines, 2026
  7. Diagnostic performance of radiomics for detecting and characterising upper tract urothelial carcinoma (UTUC): a systematic review
  8. https://d56bochluxqnz.cloudfront.net/documents/full-guideline/EAU-Guidelines-on-Upper-Urinary-Tract-Urothelial-Carcinoma-2025_2025-06-02-054038_pezz.pdf
  9. https://urology.wiki/Guidelines/Cancers/NCCN/2026/%EF%BC%882026.V1%EF%BC%89NCCN%20%E4%B8%B4%E5%BA%8A%E5%AE%9E%E8%B7%B5%E6%8C%87%E5%8D%97%EF%BC%9A%E8%86%80%E8%83%B1%E7%99%8C.pdf

Original Source(s)

Radiomics-based interpretable machine learning model from multiphasic CT imaging for predicting pathological grade in upper tract urothelial carcinoma: a multicenter study

  • by Zhanpeng Yuan, Yuhua Mei, Xiang Peng, Zongjie Wei, Yingjie Xv, Bangxin Xiao, Mingzhao Xiao

  • June 23, 2026

Related Content

Early Local Recurrence of Well-Differentiated Liposarcoma of the Hypopharynx in a Young Adult: A Case Report and Literature Review

Association between radiotherapy timing and intracranial outcomes in EGFR-mutant NSCLC with brain metastases treated with EGFR-TKIs

Efficacy and safety of reduced-dose daratumumab plus bortezomib and dexamethasone (DVd-lite) in newly diagnosed MGRS patients