Machine Learning methods for Quantitative Radiomic Biomarkers

By Chintan Parmar, Patrick Grossmann, J. Bussink, Philippe Lambin, Hugo J. W. L. Aerts. Scientific Reports, 2015

956 citations recorded by the source corpus.

Radiomics extracts and mines large number of medical imaging features quantifying tumor phenotypic characteristics. Highly accurate and reliable machine-learning approaches can drive the success of radiomic applications in clinical care. In this radiomic study, fourteen feature selection methods and twelve classification methods were examined in terms of their performance and stability for predicting overall survival. A total of 440 radiomic features were extracted from pre-treatment computed tomography (CT) images of 464 lung cancer patients. To ensure the unbiased evaluation of different machine-learning methods, publicly available implementations along with reported parameter configurations were used. Furthermore, we used two independent radiomic cohorts for training (n = 310 patients) and validation (n = 154 patients). We identified that Wilcoxon test based feature selection method WLCX (stability = 0.84 ± 0.05, AUC = 0.65 ± 0.02) and a classification method random forest RF (RSD = 3.52%, AUC = 0.66 ± 0.03) had highest prognostic performance with high stability against data perturbation. Our variability analysis indicated that the choice of classification method is the most dominant source of performance variation (34.21% of total variance). Identification of optimal machine-learning methods for radiomic applications is a crucial step towards stable and clinically relevant radiomic biomarkers, providing a non-invasive way of quantifying and monitoring tumor-phenotypic characteristics in clinical practice.

Subfields: Health Informatics, Pulmonary and Respiratory Medicine, Radiology, Nuclear Medicine and Imaging

View Paper: https://www.semanticscholar.org/paper/ebe14ab38c7bc8187737c8aed7fef3d7cd2becf7

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Health Informatics

In Pulmonary and Respiratory Medicine

In Radiology, Nuclear Medicine and Imaging

Citations in the corpus

Deep Learning Predicts Lung Cancer Treatment Response from Serial Medical Imaging (2019, background): “Indeed, several studies have demonstrated the ability to noninvasively describe tumor phenotypes with more predictive power than routine clinical measures (7 – 10).”
Artificial intelligence in radiology (2018, positive): “For example, studies in non-small-cell lung cancer (NSCLC) used radiomics to predict distant metastasis in lung adenocarcinoma 17 and tumour histological subtypes 18 as well as disease recurrence 19 , somatic mutations 20 , gene-expression profiles 21 and overall survival 22 .”
Radiomics: the bridge between medical imaging and personalized medicine (2017, positive): “The choice of modelling technique has been shown to affect prediction performance in radiomics 30 .”
Data Analysis Strategies in Medical Imaging (2018, methodology): “Predictive and prognostic models with high accuracy, reliability, and efficiency are vital factors driving the success of quantitative imaging approaches (10).”
Diagnostic Accuracy of a Machine-Learning Approach to Coronary Computed Tomographic Angiography–Based Fractional Flow Reserve: Result From the MACHINE Consortium (2018, background): “CMIV (15)† September 2009–March 2012 120 d 21/23 Ostial stenosis 35 (63%): complicated invasive FFR pressure wire position matching (21), insufficient CTA image quality (5), absence of thin-slice CTA reconstructions (4).”
Applications and limitations of radiomics (2016, background): “Despite the potential impact of these factors on quantification, strong prognostic signals of the features could still be found (Cheng et al 2013a, 2014, Cook et al 2013, Aerts et al 2014, Coroller et al 2015, Leijenaar et al 2015a, Parmar et al 2015b).”
Machine Learning in Radiation Oncology: Opportunities, Requirements, and Needs (2018, background): “Due in part to the small datasets typically encountered in radiotherapy, the choice of algorithm in a specific application can produce differences of up to 32% in predicted outcome (22).”
Residual Convolutional Neural Network for Determination of IDH Status in Low- and High-grade Gliomas from MR Imaging (2017, positive): “Radiomic approaches rely on multistep pipelines that include generation of numerous preengineered features, selection of features, and application of traditional machine learning techniques (27).”
Radiomics: the facts and the challenges of image analysis (2018, methodology): “The two most commonly used unsupervised approaches are cluster analysis [7, 14, 15] and principal component analysis (PCA) [13, 16].”
Characterization of PET/CT images using texture analysis: the past, the present… any future? (2016, methodology): “A recent study focused on this topic and compared 14 feature selection methods and 12 classifiers using large CT datasets, ranking relative performance in terms of the accuracy and stability of each approach [77].”
Radiomic Machine-Learning Classifiers for Prognostic Biomarkers of Head and Neck Cancer (2015, methodology): “We compared different feature selection and classification meth-ods using the experimental design defined by an earlier study of NSCLC radiomic cohorts (30).”
Radiomics in Brain Tumor: Image Assessment, Quantitative Feature Descriptors and Machine-learning Approaches (2017, background): “49 Anexampleofthisapproachisastudy 50 that showed the selection of machine-learning classifiers for supervised radiomic applications in detecting radiomic biomarkers.”
Quantitative radiomics studies for tissue characterization: a review of technology and methodological procedures (2017, methodology): “In a study comparing 14 filterbased feature selection techniques in combination with 18 machine-learning (ML) classification procedures in radiomics cohorts, Wilcoxon test-based feature selection showed the highest stability against data perturbation in combination with most classifiers.(81) Wrapper”
Introduction to Radiomics (2020, methodology): “(40), including mutual information–based methods and 12 machine learning classifiers; a similar approach was also used by Leger et al.”
Promises and challenges for the implementation of computational medical imaging (radiomics) in oncology (2017, background): “Machine learning approaches [40] such as decision trees and random forests [41–46], support vector machines [39, 47, 48] and more recently deep neural networks [49] appear to be promising in the domain of medical image computing.”