Radiomic Machine-Learning Classifiers for Prognostic Biomarkers of Head and Neck Cancer

Introduction “Radiomics” extracts and mines a large number of medical imaging features in a non-invasive and cost-effective way. The underlying assumption of radiomics is that these imaging features quantify phenotypic characteristics of an entire tumor. In order to enhance applicability of radiomics in clinical oncology, highly accurate and reliable machine-learning approaches are required. In this radiomic study, 13 feature selection methods and 11 machine-learning classification methods were evaluated in terms of their performance and stability for predicting overall survival in head and neck cancer patients. Methods Two independent head and neck cancer cohorts were investigated. Training cohort HN1 consisted of 101 head and neck cancer patients. Cohort HN2 (n = 95) was used for validation. A total of 440 radiomic features were extracted from the segmented tumor regions in CT images. Feature selection and classification methods were compared using an unbiased evaluation framework. Results We observed that the three feature selection methods minimum redundancy maximum relevance (AUC = 0.69, Stability = 0.66), mutual information feature selection (AUC = 0.66, Stability = 0.69), and conditional infomax feature extraction (AUC = 0.68, Stability = 0.7) had high prognostic performance and stability. The three classifiers BY (AUC = 0.67, RSD = 11.28), RF (AUC = 0.61, RSD = 7.36), and NN (AUC = 0.62, RSD = 10.52) also showed high prognostic performance and stability. Analysis investigating performance variability indicated that the choice of classification method is the major factor driving the performance variation (29.02% of total variance). Conclusion Our study identified prognostic and reliable machine-learning methods for the prediction of overall survival of head and neck cancer patients. Identification of optimal machine-learning methods for radiomics-based prognostic analyses could broaden the scope of radiomics in precision oncology and cancer care.

[1]  Robert M. Haralick,et al.  Textural Features for Image Classification , 1973, IEEE Trans. Syst. Man Cybern..

[2]  Mary M. Galloway,et al.  Texture analysis using gray level run lengths , 1974 .

[3]  Isabelle Guyon,et al.  An Introduction to Variable and Feature Selection , 2003, J. Mach. Learn. Res..

[4]  Howard Y. Chang,et al.  Decoding global gene expression programs in liver cancer by noninvasive imaging , 2007, Nature Biotechnology.

[5]  Max Kuhn,et al.  Building Predictive Models in R Using the caret Package , 2008 .

[6]  L. Turnbull,et al.  Prognostic value of pre-treatment DCE-MRI parameters in predicting disease free and overall survival for breast cancer patients undergoing neoadjuvant chemotherapy. , 2009, European journal of radiology.

[7]  Huan Liu,et al.  Advancing feature selection research , 2010 .

[8]  Balaji Ganeshan,et al.  Texture analysis of non-small cell lung cancer on unenhanced computed tomography: initial evidence for a relationship with tumour glucose metabolism and stage , 2010, Cancer imaging : the official publication of the International Cancer Imaging Society.

[9]  Huan Liu,et al.  Advancing Feature Selection Research − ASU Feature Selection Repository , 2010 .

[10]  R. Jeraj,et al.  Variability of textural features in FDG PET images due to different acquisition modes and reconstruction parameters , 2010, Acta oncologica.

[11]  Jean-Philippe Vert,et al.  The Influence of Feature Selection Methods on Accuracy, Stability and Interpretability of Molecular Signatures , 2011, PloS one.

[12]  Robert J. Gillies,et al.  Developing a classifier model for lung tumors in CT-scan images , 2011, 2011 IEEE International Conference on Systems, Man, and Cybernetics.

[13]  Gavin Brown,et al.  Conditional Likelihood Maximisation: A Unifying Framework for Information Theoretic Feature Selection , 2012, J. Mach. Learn. Res..

[14]  Andre Dekker,et al.  Radiomics: the process and the challenges. , 2012, Magnetic resonance imaging.

[15]  Patrick Granton,et al.  Radiomics: extracting more information from medical images using advanced feature analysis. , 2012, European journal of cancer.

[16]  Debyani Chakravarty,et al.  Intratumoral heterogeneity of receptor tyrosine kinases EGFR and PDGFRA amplification in glioblastoma defines subpopulations with distinct growth factor response , 2012, Proceedings of the National Academy of Sciences.

[17]  Ameet Talwalkar,et al.  Foundations of Machine Learning , 2012, Adaptive computation and machine learning.

[18]  D. Longo,et al.  Tumor heterogeneity and personalized medicine. , 2012, The New England journal of medicine.

[19]  J. Reis-Filho,et al.  Breast cancer intratumor genetic heterogeneity: causes and implications , 2012, Expert review of anticancer therapy.

[20]  P. Lambin,et al.  Stability of FDG-PET Radiomics features: An integrated analysis of test-retest and inter-observer variability , 2013, Acta oncologica.

[21]  Vicky Goh,et al.  Are Pretreatment 18F-FDG PET Tumor Textural Features in Non–Small Cell Lung Cancer Associated with Response and Survival After Chemoradiotherapy? , 2013, The Journal of Nuclear Medicine.

[22]  M. Martel,et al.  High quality machine-robust image features: identification in nonsmall cell lung cancer computed tomography images. , 2013, Medical physics.

[23]  L. Pusztai,et al.  Cancer heterogeneity: implications for targeted therapeutics , 2013, British Journal of Cancer.

[24]  V. Goh,et al.  Non-small cell lung cancer: histopathologic correlates for texture parameters at CT. , 2013, Radiology.

[25]  M. Hatt,et al.  Robustness of intratumour 18F-FDG PET uptake heterogeneity quantification for therapy response prediction in oesophageal carcinoma , 2013, European Journal of Nuclear Medicine and Molecular Imaging.

[26]  Ron Kikinis,et al.  Volumetric CT-based segmentation of NSCLC using 3D-Slicer , 2013, Scientific Reports.

[27]  U. Ficola,et al.  Prediction of 2 years-survival in patients with stage I and II non-small cell lung cancer utilizing 18F-FDG PET/CT SUV quantification , 2013, Radiology and oncology.

[28]  P. Lambin,et al.  Predicting outcomes in radiation oncology—multifactorial decision support systems , 2013, Nature Reviews Clinical Oncology.

[29]  Andre Dekker,et al.  Prognostic value of metabolic metrics extracted from baseline positron emission tomography images in non-small cell lung cancer , 2013, Acta oncologica.

[30]  W. Niessen,et al.  Quantification of Heterogeneity as a Biomarker in Tumor Imaging: A Systematic Review , 2014, PloS one.

[31]  P. Lambin,et al.  Robust Radiomics Feature Quantification Using Semiautomatic Volumetric Segmentation , 2014, PloS one.

[32]  Senén Barro,et al.  Do we need hundreds of classifiers to solve real world classification problems? , 2014, J. Mach. Learn. Res..

[33]  W. Tsai,et al.  Exploring Variability in CT Characterization of Tumors: A Preliminary Phantom Study. , 2014, Translational oncology.

[34]  S. Plevritis,et al.  Glioblastoma multiforme: exploratory radiogenomic analysis by using quantitative image features. , 2014, Radiology.

[35]  P. Lambin,et al.  A prospective study comparing the predictions of doctors versus models for treatment outcome of lung cancer patients: a step toward individualized care and shared decision making. , 2014, Radiotherapy and oncology : journal of the European Society for Therapeutic Radiology and Oncology.

[36]  P. Lambin,et al.  Decoding tumour phenotype by noninvasive imaging using a quantitative radiomics approach , 2014, Nature Communications.

[37]  Scott N. Hwang,et al.  Outcome prediction in patients with glioblastoma by using imaging, clinical, and genomic biomarkers: focus on the nonenhancing component of the tumor. , 2014, Radiology.

[38]  Robert J. Gillies,et al.  Predicting Outcomes of Nonsmall Cell Lung Cancer Using CT Image Features , 2014, IEEE Access.

[39]  Robert J. Gillies,et al.  The effect of SUV discretization in quantitative FDG-PET Radiomics: the need for standardized methodology in tumor texture analysis , 2015, Scientific Reports.

[40]  Olivier Gevaert,et al.  Addition of MR imaging features and genetic biomarkers strengthens glioblastoma survival prediction in TCGA patients. , 2015, Journal of neuroradiology. Journal de neuroradiologie.

[41]  P. Lambin,et al.  CT-based radiomic signature predicts distant metastasis in lung adenocarcinoma. , 2015, Radiotherapy and oncology : journal of the European Society for Therapeutic Radiology and Oncology.

[42]  Benjamin Haibe-Kains,et al.  Radiomic feature clusters and Prognostic Signatures specific for Lung and Head & Neck cancer , 2015, Scientific Reports.

[43]  Jinzhong Yang,et al.  Measuring Computed Tomography Scanner Variability of Radiomics Features , 2015, Investigative radiology.

[44]  P. Lambin,et al.  Machine Learning methods for Quantitative Radiomic Biomarkers , 2015, Scientific Reports.