JoVE - Journal of Visualized Experiments
JoVE Visualize
Contact Us

Development and validation of multi-sequence magnetic resonance imaging radiomics models for predicting tumor response to radiotherapy in locally advanced non-small cell lung cancer

  • 0Department of Radiation Oncology, Harbin Medical University Cancer Hospital, Harbin, China.
Journal of Thoracic Disease +

|

April 14, 2025

|

April 14, 2025

View abstract on PubMed

Summary

This summary is machine-generated.

This study developed a multisequence MRI radiomics model to predict radiotherapy response in locally advanced non-small cell lung cancer (LA-NSCLC). The combined sequence model showed high accuracy, offering a promising tool for personalized LA-NSCLC treatment.

Area Of Science

  • Radiomics and Medical Imaging
  • Oncology
  • Radiotherapy

Background

  • Locally advanced non-small cell lung cancer (LA-NSCLC) presents significant heterogeneity, complicating treatment decisions.
  • Current methods for predicting radiotherapy response in LA-NSCLC lack accuracy and comprehensiveness.
  • There is a need for a robust model to assess LA-NSCLC patient response to radiotherapy.

Purpose Of The Study

  • To develop and evaluate a multisequence magnetic resonance imaging (MRI) radiomics model for predicting tumor response to radiotherapy in LA-NSCLC patients.
  • To assess the clinical utility of the developed radiomics model.

Main Methods

  • Retrospective collection of MRI data from Stage III NSCLC patients treated with radiotherapy.
  • Extraction and integration of 3,045 radiomic features from T1WI, T2WI, and DWI sequences.
  • Feature selection using mRMR and LASSO, followed by model construction with LR, SVM, KNN, and random forest algorithms.
  • Evaluation of models using ROC curves, calibration curves, and DCA.

Main Results

  • A combined sequence logistic regression (LR) model demonstrated the highest predictive performance.
  • The training set achieved an AUC of 0.888 with 83.3% accuracy, 92.9% sensitivity, and 75% specificity.
  • The testing set achieved an AUC of 0.815 with 73.1% accuracy, 91.7% sensitivity, and 57.1% specificity.
  • Calibration curves and DCA confirmed the model's good predictive performance and clinical utility.

Conclusions

  • A multisequence MRI radiomics model shows significant potential for predicting radiotherapy response in LA-NSCLC.
  • The combined sequence model outperformed individual sequences in predictive accuracy and clinical applicability.
  • This model could aid in optimizing radiotherapy strategies for LA-NSCLC patients.

Keywords: