Related Experiment Video
Updated: Sep 16, 2025
Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model
Published on: August 16, 2020
Radiomics-Based Machine Learning for Determining MYCN Amplification Status in Childhood Neuroblastoma: A Systematic
Haoru Wang1, Yi Ji2,3, Xin Chen1
1Department of Radiology, Children's Hospital of Chongqing Medical University, National Clinical Research Center for Child Health and Disorders, Ministry of Education Key Laboratory of Child Development and Disorders, Chongqing Key Laboratory of Pediatrics, Chongqing, China.
Radiomics shows promise for non-invasively detecting MYCN amplification in neuroblastoma, a marker of poor prognosis. While study quality varies, pooled analysis indicates good diagnostic accuracy, suggesting potential for clinical use.
Area of Science:
- Oncology
- Radiology
- Machine Learning
- Medical Imaging
Background:
- MYCN oncogene amplification in neuroblastoma is a key driver of tumor cell proliferation and a marker of poor prognosis.
- Non-invasive methods to determine MYCN amplification status are crucial for treatment planning and patient management.
- Radiomics, extracting quantitative features from medical images, has emerged as a promising approach for this non-invasive assessment.
Purpose of the Study:
- To quantitatively evaluate the diagnostic accuracy of radiomics-based machine learning models for determining MYCN amplification in neuroblastoma.
- To critically assess the methodological quality of existing studies using radiomics for MYCN amplification detection.
Main Methods:
- Systematic literature search across major databases (PubMed, Embase, Web of Science, Cochrane Library) for studies published between January 1, 2000, and June 30, 2024.
- Inclusion of studies focusing on radiomics for MYCN amplification in neuroblastoma, with methodological quality assessed using RQS, METRICS, and QUADAS-2 tools.
- Meta-analysis of validation performance for eligible studies, focusing on pooled diagnostic accuracy metrics.
Main Results:
- Nine studies involving 851 patients were included; seven studies with 217 patients were eligible for meta-analysis.
- Methodological quality assessments revealed variability, with most studies having a low or unclear risk of bias.
- Pooled analysis demonstrated high diagnostic accuracy: sensitivity 0.78, specificity 0.92, AUC 0.94, indicating significant potential.
Conclusions:
- Radiomics-based machine learning models show considerable promise as a non-invasive tool for detecting MYCN amplification in neuroblastoma.
- Despite variability in study design and quality, the pooled diagnostic performance is encouraging.
- Further validation in larger, multicenter studies is essential to improve and confirm the clinical applicability of these radiomics models.
