Dataset Information

A downsampling strategy to assess the predictive value of radiomic features.

ABSTRACT: Many studies are devoted to the design of radiomic models for a prediction task. When no effective model is found, it is often difficult to know whether the radiomic features do not include information relevant to the task or because of insufficient data. We propose a downsampling method to answer that question when considering a classification task into two groups. Using two large patient cohorts, several experimental configurations involving different numbers of patients were created. Univariate or multivariate radiomic models were designed from each configuration. Their performance as reflected by the Youden index (YI) and Area Under the receiver operating characteristic Curve (AUC) was compared to the stable performance obtained with the highest number of patients. A downsampling method is described to predict the YI and AUC achievable with a large number of patients. Using the multivariate models involving machine learning, YI and AUC increased with the number of patients while they decreased for univariate models. The downsampling method better estimated YI and AUC obtained with the largest number of patients than the YI and AUC obtained using the number of available patients and identifies the lack of information relevant to the classification task when no such information exists.

SUBMITTER: Dirand AS

PROVIDER: S-EPMC6883070 | biostudies-literature | 2019 Nov

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

A downsampling strategy to assess the predictive value of radiomic features.

Dirand Anne-Sophie AS Frouin Frédérique F Buvat Irène I

Scientific reports 20191128 1

Many studies are devoted to the design of radiomic models for a prediction task. When no effective model is found, it is often difficult to know whether the radiomic features do not include information relevant to the task or because of insufficient data. We propose a downsampling method to answer that question when considering a classification task into two groups. Using two large patient cohorts, several experimental configurations involving different numbers of patients were created. Univaria ...[more]

PMID: 31780708

Similar Datasets

Project description:ObjectiveTo investigate the effects of Image Biomarker Standardisation Initiative (IBSI) compliance, harmonisation of calculation settings and platform version on the statistical reliability of radiomic features and their corresponding ability to predict clinical outcome.MethodsThe statistical reliability of radiomic features was assessed retrospectively in three clinical datasets (patient numbers: 108 head and neck cancer, 37 small-cell lung cancer, 47 non-small-cell lung cancer). Features were calculated using four platforms (PyRadiomics, LIFEx, CERR and IBEX). PyRadiomics, LIFEx and CERR are IBSI-compliant, whereas IBEX is not. The effects of IBSI compliance, user-defined calculation settings and platform version were assessed by calculating intraclass correlation coefficients and confidence intervals. The influence of platform choice on the relationship between radiomic biomarkers and survival was evaluated using univariable cox regression in the largest dataset.ResultsThe reliability of radiomic features calculated by the different software platforms was only excellent (ICC > 0.9) for 4/17 radiomic features when comparing all four platforms. Reliability improved to ICC > 0.9 for 15/17 radiomic features when analysis was restricted to the three IBSI-compliant platforms. Failure to harmonise calculation settings resulted in poor reliability, even across the IBSI-compliant platforms. Software platform version also had a marked effect on feature reliability in CERR and LIFEx. Features identified as having significant relationship to survival varied between platforms, as did the direction of hazard ratios.ConclusionIBSI compliance, user-defined calculation settings and choice of platform version all influence the statistical reliability and corresponding performance of prognostic models in radiomics.Key points• Reliability of radiomic features varies between feature calculation platforms and with choice of software version. • Image Biomarker Standardisation Initiative (IBSI) compliance improves reliability of radiomic features across platforms, but only when calculation settings are harmonised. • IBSI compliance, user-defined calculation settings and choice of platform version collectively affect the prognostic value of features.

Project description:BackgroundTo assess the potential of radiomic features to quantify components of blood in intraaortic vessels to non-invasively predict moderate-to-severe anemia in non-contrast enhanced CT scans.MethodsOne hundred patients (median age, 69 years; range, 19-94 years) who received CT scans of the thoracolumbar spine and blood-testing for hemoglobin and hematocrit levels ± 24 h between 08/2018 and 11/2019 were retrospectively included. Intraaortic blood was segmented using a spherical volume of interest of 1 cm diameter with consecutive radiomic analysis applying PyRadiomics software. Feature selection was performed applying analysis of correlation and collinearity. The final feature set was obtained to differentiate moderate-to-severe anemia. Random forest machine learning was applied and predictive performance was assessed. A decision-tree was obtained to propose a cut-off value of CT Hounsfield units (HU).ResultsHigh correlation with hemoglobin and hematocrit levels was shown for first-order radiomic features (p < 0.001 to p = 0.032). The top 3 features showed high correlation to hemoglobin values (p) and minimal collinearity (r) to the top ranked feature Median (p < 0.001), Energy (p = 0.002, r = 0.387), Minimum (p = 0.032, r = 0.437). Median (p < 0.001) and Minimum (p = 0.003) differed in moderate-to-severe anemia compared to non-anemic state. Median yielded superiority to the combination of Median and Minimum (p(AUC) = 0.015, p(precision) = 0.017, p(accuracy) = 0.612) in the predictive performance employing random forest analysis. A Median HU value ≤ 36.5 indicated moderate-to-severe anemia (accuracy = 0.90, precision = 0.80).ConclusionsFirst-order radiomic features correlate with hemoglobin levels and may be feasible for the prediction of moderate-to-severe anemia. High dimensional radiomic features did not aid augmenting the data in our exemplary use case of intraluminal blood component assessment. Trial registration Retrospectively registered.

Project description:BACKGROUND:Textural features extracted from MRI potentially provide prognostic information additional to volume for influencing surgical management of cervical cancer. PURPOSE:To identify textural features that differ between cervical tumors above and below the volume threshold of eligibility for trachelectomy and determine their value in predicting recurrence in patients with low-volume tumors. METHODS:Of 378 patients with Stage1-2 cervical cancer imaged prospectively (3T, endovaginal coil), 125 had well-defined, histologically-confirmed squamous or adenocarcinomas with >100 voxels (>0.07 cm3) suitable for radiomic analysis. Regions-of-interest outlined the whole tumor on T2-W images and apparent diffusion coefficient (ADC) maps. Textural features based on grey-level co-occurrence matrices were compared (Mann-Whitney test with Bonferroni correction) between tumors greater (n = 46) or less (n = 79) than 4.19 cm3. Clustering eliminated correlated variables. Significantly different features were used to predict recurrence (regression modelling) in surgically-treated patients with low-volume tumors and compared with a model using clinico-pathological features. RESULTS:Textural features (Dissimilarity, Energy, ClusterProminence, ClusterShade, InverseVariance, Autocorrelation) in 6 of 10 clusters from T2-W and ADC data differed between high-volume (mean ± SD 15.3 ± 11.7 cm3) and low-volume (mean ± SD 1.3 ± 1.2 cm3) tumors. (p < 0.02). In low-volume tumors, predicting recurrence was indicated by: Dissimilarity, Energy (ADC-radiomics, AUC = 0.864); Dissimilarity, ClusterProminence, InverseVariance (T2-W-radiomics, AUC = 0.808); Volume, Depth of Invasion, LymphoVascular Space Invasion (clinico-pathological features, AUC = 0.794). Combining ADC-radiomic (but not T2-radiomic) and clinico-pathological features improved prediction of recurrence compared to the clinico-pathological model (AUC = 0.916, p = 0.006). Findings were supported by bootstrap re-sampling (n = 1000). CONCLUSION:Textural features from ADC maps and T2-W images differ between high- and low-volume tumors and potentially predict recurrence in low-volume tumors.

Dataset Information

A downsampling strategy to assess the predictive value of radiomic features.

Publications

A downsampling strategy to assess the predictive value of radiomic features.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets