Related Experiment Video
Updated: Aug 5, 2026
10:17
Guidelines and Experience Using Imaging Biomarker Explorer (IBEX) for Radiomics
Published on: January 8, 2018
Radiomics-Based Prediction of Treatment Response in Non-Small Cell Lung Cancer Using Pre-Treatment CT Imaging
Lama Almudaimeegh1, Noman Nazeer2, Zuhal Y Hamd3
1Department of Internal Medicine, College of Medicine, Princess Nourah bint Abdulrahman University, P.O. Box 84428, Riyadh 11671, Saudi Arabia.
Diagnostics (Basel, Switzerland)
|July 28, 2026
Summary
This study developed a CT-based radiomics signature to predict chemotherapy response in non-small cell lung cancer (NSCLC) patients. The signature shows promise as a decision-support tool for personalized treatment strategies.
Area of Science:
- Radiology
- Oncology
- Medical Imaging
Background:
- Non-small cell lung cancer (NSCLC) is the most prevalent form of lung cancer globally.
- Predicting patient response to chemotherapy remains a significant clinical challenge.
- Radiomics offers a non-invasive method to extract quantitative imaging data reflecting tumor heterogeneity.
Purpose of the Study:
- To develop and externally validate a CT-derived radiomic signature for predicting treatment outcomes in advanced NSCLC patients receiving first-line platinum-based chemotherapy.
- To assess the signature's ability to differentiate patients based on their likelihood of therapeutic response.
Main Methods:
- Utilized baseline contrast-enhanced CT scans from NSCLC patients across three centers.
- Extracted 851 quantitative imaging biomarkers using PyRadiomics, focusing on histogram, morphology, texture, and wavelet features.
- Applied LASSO regression and stability selection to identify predictive parameters, and tested predictive capability using five machine learning models.
- Validated the radiomics signature in an independent external cohort, with treatment response assessed by RECIST 1.1.
Main Results:
- After rigorous selection, 14 radiomic parameters were identified.
- The integrated model, combining radiomic and clinical features, demonstrated strong predictive power with AUC values of 0.876 (training), 0.849 (internal validation), and 0.831 (external validation).
- The radiomic signature effectively distinguished patients by their probability of therapeutic response, and the combined nomogram offered greater clinical utility than clinical factors alone.
Conclusions:
- A pre-treatment CT-based radiomics signature shows potential as a decision-support tool for predicting chemotherapy response in NSCLC.
- Further prospective and multi-ethnic validation is necessary before widespread clinical application.
- Future research should focus on validating these findings in larger, diverse patient cohorts.
