MRI-based radiomics nomogram may predict the response to induction chemotherapy and survival in locally advanced nasopharyngeal carcinoma
To establish and validate a radiomics nomogram for prediction of induction chemotherapy (IC) response and survival in nasopharyngeal carcinoma (NPC) patients.
One hundred twenty-three NPC patients (100 in training and 23 in validation cohort) with multi-MR images were enrolled. A radiomics nomogram was established by integrating the clinical data and radiomics signature generated by support vector machine.
The radiomics signature consisting of 19 selected features from the joint T1-weighted (T1-WI), T2-weighted (T2-WI), and contrast-enhanced T1-weighted MRI images (T1-C) showed good prognostic performance in terms of evaluating IC response in two cohorts. The radiomics nomogram established by integrating the radiomics signature with clinical data outperformed clinical nomogram alone (C-index in validation cohort, 0.863 vs 0.549; p < 0.01). Decision curve analysis demonstrated the clinical utility of the radiomics nomogram. Survival analysis showed that IC responders had significant better PFS (progression-free survival) than non-responders (3-year PFS 84.81% vs 39.75%, p < 0.001). Low-risk groups defined by radiomics signature had significant better PFS than high-risk groups (3-year PFS 76.24% vs 48.04%, p < 0.05).
Multiparametric MRI-based radiomics could be helpful for personalized risk stratification and treatment in NPC patients receiving IC.
• MRI Radiomics can predict IC response and survival in non-endemic NPC.
• Radiomics signature in combination with clinical data showed excellent predictive performance.
• Radiomics signature could separate patients into two groups with different prognosis.
KeywordsNasopharyngeal carcinoma Magnetic resonance imaging Radiomics Machine learning Induction chemotherapy
Least absolute shrinkage and selection operator
Support vector machine
This study has received funding by the National Natural Science Foundation of China Grants 81872699 and Key project of Shanxi Province 2017ZDXM-SF-043.
Compliance with ethical standards
The scientific guarantor of this publication is Lina Zhao.
Conflict of interest
The authors of this manuscript declare no relationships with any companies, whose products or services may be related to the subject matter of the article.
Statistics and biometry
Yutian Yin, one of the authors, has significant statistical expertise.
Written informed consent was not required for this study because the retrospective nature of the study.
Institutional Review Board approval was obtained.
• diagnostic or prognostic study
• performed at one institution
- 13.Verma V, Simone CB 2nd, Krishnan S, Lin SH, Yang J, Hahn SM (2017) The rise of radiomics and implications for oncologic management. J Natl Cancer Inst 109Google Scholar
- 29.Zhou M, Leung A, Echegaray S et al (2017) Non-small cell lung cancer radiogenomics map identifies relationships between molecular and imaging phenotypes with prognostic implications. Radiology. https://doi.org/10.1148/radiol.2017161845:161845