Skip to main content

Table 1 Summary of literature for magnetic resonance imaging radiomics feature robustness

From: Robustness of radiomic features in magnetic resonance imaging: review and a phantom study

ReferenceDisease / phantomMR sequences# featuresFeature classesParameters evaluatedStatistical analysisRobustness evaluation
Baessler et al. [26], 2019Vegetable/fruit phantomFLAIR, T1w, T2w45Intensity, shape, textureMR sequence, resolutionCCC, DR, Bland-Altman analyses, ICCTest-retest robustness, intraobserver and interobserver reproducibility
Traverso et al. [50], 2019Locally advanced rectal cancerDWI (ADC map)70Intensity, shape, texturePre-processing filter, re-binning and resamplingCCC, ICC, Spearman correlationInter-observer dependence
Duron et al. [39], 2019Lacrymal gland tumor and breast lesionT1w, DWI (ADC map), DIXON, DISCO69/57 (2 softwares)TextureDiscretization method, bin width and bin numberCCC, ICC(2,1)Intra- and inter-observer reproducibility
Lecler et al. [37], 2019Lacrimal gland tumorT1w, DWI (ADC map), DIXON85Intensity, shape, textureMR sequence, metric thresholdCCC, ICC(2,1), Spearman correlationIntra- and inter-observer reproducibility, non-redundancy
Um et al. [51], 2019Glioblastoma multiformeFLAIR, T1w, post-contrast T1w420Intensity, shape, texture, filter-basedPreprocessing technique on multi-scanner datasets, bin numberTwo-sided Wilcoxon testsFeature variability
Schwier et al. [24], 2019Prostate cancerT2w, DWI (ADC map)NAIntensity, shape, texture, filter-basedImage normalization, 2D/3D texture computation, bin widths, and image pre-filteringICC(1,1)Test-retest repeatability
Fiset et al. [38], 2019Cervical cancerT2w1761Intensity, shape, texture, filter-basedQuantization method, LoG kernel sizes,ICC(1,1), ICC(2,1), Pearson correlation, Krippendorff’s alphaTest-retest repeatability, cross-scanner reproducibility, inter-observer reproducibility
Peerlings et al. [33], 2019Ovarian, lung and colorectal liver metastasis cancerDWI (ADC map)1322Intensity, shape, texture, filter-basedCenter and vendorCCCFeature stability
Buch et al. [52], 2018Nonanatomic Gd-DTPA phantomT1w41Intensity, texture, filter-based (Laws)Magnet strength, flip-angle, number of excitations, scanner platformQ valuesFeature variability
Yang et al. [53], 2018Simulated data from digital phantom and gliomaT1w, T2w26TextureNoise level, acceleration factor, and image reconstruction algorithmStudent’s t-test, CVFeature variance
Bologna et al. [32], 2018Soft tissue sarcoma and oropharyngeal cancerDWI (ADC map)69Intensity, textureROI transformation and bin numberAbsolute percentage variation, two-way mixed effect ICCFeature stability and discrimination
Chirra et al. [40], 2018Prostate cancerT2w406Intensity, texture, filter-basedDifferent sitesMultivariate CV and Instability ScoreCross-site reproducibility
Saha et al. [31], 2018Breast cancerDCE-MRI (first postcontrast, PE, SER, washing rate maps)529Intensity, shape, textureScanner, contrast agentICC(3,1), Pearson correlation, average DSCInter-reader stability, inter-relations within feature groups, pairwise reader variability
Molina et al. [27], 2017GlioblastomaT1w16TextureSpatial resolution and bin numberCVFeature variation
Brynolfsson et al. [54], 2017Glioma and prostate cancerDWI (ADC map)19Texturenoise level, resolution, ADC map construction, quantization method, and bin numberTwo-sample Kolmogorov-Smirnov testsFeature distribution variation
Gourtsoyianni et al. [41], 2017Primary rectal cancerT2w46Intensity, texture, filter-based2 baseline examinationswCVTest-retest repeatability
Guan et al. [55], 2016Cervical cancerDWI (ADC map)8Intensity, textureGLCM directionICC, Wilcoxon test, Kruskal-Wallis test, and ROC curveInter- and intra-observer agreement
Molina et al. [56], 2016GlioblastomaT1w16TextureMatrix size and bin numberCVFeature variation
Savio et al. [57], 2010Multiple sclerosisT1w264Intensity, texture, filter-basedGlobal, regional and local featuresWilcoxon’s signed ranks testFeature variation
Mayerhoefer et al. [58], 2009PSAG phantomT2wNATexture, filter-basedSpatial resolution, NAs, TR, TE, and SBWLDA and k-NN classifierAbility to distinguish between different patterns
Collewet et al. [59], 2004Cheese phantomT2w, PDW90Texture, filter-basedMRI acquisition protocol and quantization methodPOE, ACC, 1-NN classifierClassification
  1. MR Magnetic resonance, FLAIR Fluid-attenuated inversion recovery, DWI Diffusion-weighted imaging, ADC Apparent diffusion coefficient, DISCO Differential subsampling with cartesian ordering, DCE-MRI Dynamic contrast-enhanced magnetic resonance imaging, PE Peak enhancement, SER Signal enhancement ratio, PDW Proton density weighted, LoG Laplacian of Gaussian, NAs Number of acquisitions, TR Repetition time, TE Echo time, SBW Sampling bandwidth, CCC Concordance correlation coefficient, DR Dynamic range, ICC Intraclass correlation coefficient, wCV Within-subject coefficient of variation, ROC Receiver operating characteristic, CV Coefficient of variation, DSC Dice similarity coefficients, LDA Linear discriminant analysis, k-NN k nearest neighbor, POE Probability of error, ACC Average correlation coefficient, 1-NN 1-nearest neighbor