Related Experiment Video
Updated: Jul 2, 2025
Ultrasound Imaging-guided Intracardiac Injection to Develop a Mouse Model of Breast Cancer Brain Metastases Followed by Longitudinal MRI
Published on: March 6, 2014
Using MRI radiomics to predict the efficacy of immunotherapy for brain metastasis in patients with small cell lung
Xiaonan Shi1, Peiliang Wang1,2, Yikun Li1
1Department of Radiation Oncology, Shandong Cancer Hospital and Institute, Shandong First Medical University and Shandong Academy of Medical Sciences, Jinan, China.
This study developed a radiomics nomogram to predict immune checkpoint inhibitor (ICI) efficacy in small cell lung cancer (SCLC) patients with brain metastases (BMs). The model accurately predicts intracranial efficacy, aiding personalized treatment strategies for SCLC with BMs.
Area of Science:
- Radiomics and Medical Imaging
- Oncology
- Neuro-oncology
Background:
- Brain metastases (BMs) are frequent in small cell lung cancer (SCLC).
- The effectiveness of immune checkpoint inhibitors (ICIs) for SCLC patients with BMs remains uncertain.
- Predictive biomarkers for treatment response are needed.
Purpose of the Study:
- To develop and validate a radiomics nomogram using magnetic resonance imaging (MRI).
- To predict the intracranial efficacy of ICIs in SCLC patients with BMs.
- To aid in personalized treatment selection.
Main Methods:
- A training and validation cohort of 101 SCLC patients with BMs treated with ICIs.
- Radiomic features were selected using interclass correlation coefficient (ICC), univariate logistic regression, and random forest.
- A nomogram was constructed combining a radiomics score (Rad-score), treatment lines, and neutrophil-to-lymphocyte ratio (NLR).
- Model performance was assessed using discrimination, calibration, and clinical utility, with Kaplan-Meier curves for survival analysis.
Main Results:
- Ten radiomic features formed the Rad-score, differentiating intracranial efficacy (AUC training: 0.759, AUC validation: 0.667).
- The combined nomogram (Rad-score, treatment lines, NLR) showed high predictive performance (AUC training: 0.878, AUC validation: 0.875).
- The nomogram significantly correlated with progression-free survival (PFS) and intracranial progression-free survival (iPFS), but not overall survival (OS).
Conclusions:
- A validated radiomics nomogram can predict intracranial efficacy of ICIs in SCLC patients with BMs.
- This tool can assist in tailoring individual-based treatment strategies.
- Further research may refine predictive models for SCLC brain metastases.
