Quantitative radiomics: impact of stochastic effects on textural feature analysis implies the need for standards (original) (raw)
Related papers
PLOS ONE, 2020
Image texture is increasingly used to discriminate tissues and lesions in PET/CT. For quantification or in computer-aided diagnosis, textural feature analysis must produce robust and comparable values. Because statistical feature values depend on image count statistics, we investigated in depth the stability of Haralick features values as functions of acquisition duration, and for common image resolutions and reconstructions. Methods A homogeneous cylindrical phantom containing 9.6 kBq/ml Ge-68 was repeatedly imaged on a Siemens Biograph mCT, with acquisition durations ranging from three seconds to three hours. Images with 1.5, 2, and 4 mm isometrically spaced voxels were reconstructed with filtered back-projection (FBP), ordered subset expectation maximization (OSEM), and the Siemens TrueX algorithm. We analysed Haralick features derived from differently quantized (3 to 8-bit) grey level co-occurrence matrices (GLCMs) as functions of exposure E, which we defined as the product of activity concentration in a volume of interest (VOI) and acquisition duration. The VOI was a 50 mm wide cube at the centre of the phantom. Feature stability was defined for df/dE ! 0. Results The most stable feature values occurred in low resolution FBPs, whereas some feature values from 1.5 mm TrueX reconstructions ranged over two orders of magnitude. Within the same reconstructions, most feature value-exposure curves reached stable plateaus at similar exposures, regardless of GLCM quantization. With 8-bit GLCM, median time to stability was 16 s and 22 s for FBPs, 18 s and 125 s for OSEM, and 23 s, 45 s, and 76 s for PSF
Medical Physics, 2021
Background: Radiomics refers to the extraction of a large number of image biomarker describing the tumor phenotype displayed in a medical image. Extracted from positron emission tomography (PET) images, radiomics showed diagnostic and prognostic value for several cancer types. However, a large number of radiomic features are nonreproducible or highly correlated with conventional PET metrics. Moreover, radiomic features used in the clinic should yield relevant information about tumor texture. In this study, we propose a framework to identify technical and clinical meaningful features and exemplify our results using a PET non-small cell lung cancer (NSCLC) dataset. Materials and methods: The proposed selection procedure consists of several steps. A priori, we only include features that were found to be reproducible in a multicenter setting. Next, we apply a voxel randomization step to identify features that reflect actual textural information, that is, that yield in 90% of the patient scans a value significantly different from random texture. Finally, the remaining features were correlated with standard PET metrics to further remove redundancy with common PET metrics. The selection procedure was performed for different volume ranges, that is, excluding lesions with smaller volumes in order to assess the effect of tumor size on the results. To exemplify our procedure, the selected features were used to predict 1-yr survival in a dataset of 150 NSCLC patients. A predictive model was built using volume as predictive factor for smaller, and one of the selected features as predictive factor for bigger lesions. The prediction accuracy of the both models were compared with the prediction accuracy of volume. Results: The number of selected features depended on the lesion size included in the analysis. When including the whole dataset, from 19 features reflecting actual texture only two were found to be not strongly correlated with conventional PET metrics. When excluding lesions smaller than 11.49 and 33.10 mL (25 and 50 percentile of the dataset), four out of 27 features and 13 out of 29 features remained after eliminating features highly correlated with standard PET metrics. When excluding lesions smaller than 103.9 mL (75 percentile), 33 out of 53 features remained. For larger lesions,
Poisson Noise Obscures Hypometabolic Lesions in PET
The Yale Journal of Biology and Medicine, 2012
The technology of fluoro-deoxyglucose positron emission tomography (PET) has drastically increased our ability to visualize the metabolic process of numerous neurological diseases. The relationship between the methodological noise sources inherent to PET technology and the resulting noise in the reconstructed image is complex. In this study, we use Monte Carlo simulations to examine the effect of Poisson noise in the PET signal on the noise in reconstructed space for two pervasive reconstruction algorithms: the historical filtered back-projection (FBP) and the more modern expectation maximization (EM). We confirm previous observations that the image reconstructed with the FBP biases all intensity values toward the mean, likely due to spatial spreading of high intensity voxels. However, we demonstrate that in both algorithms the variance from high intensity voxels spreads to low intensity voxels and obliterates their signal to noise ratio. This finding has profound impacts on the clinical interpretation of hypometabolic lesions. Our results suggest that PET is relatively insensitive when it comes to detecting and quantifying changes in hypometabolic tissue. Further, the images reconstructed with EM visually match the original images more closely, but more detailed analysis reveals as much as a 40 percent decrease in the signal to noise ratio for high intensity voxels relative to the FBP. This suggests that even though the apparent spatial resolution of EM outperforms FBP, the signal to noise ratio of the intensity of each voxel may be higher in the FBP. Therefore, EM may be most appropriate for manual visualization of pathology, but FBP should be used when analyzing quantitative markers of the PET signal. This suggestion that different reconstruction algorithms should be used for quantification versus visualization represents a major paradigm shift in the analysis and interpretation of PET images.
2019 IEEE Nuclear Science Symposium and Medical Imaging Conference (NSS/MIC)
Medical imaging plays an essential role in the diagnosis and treatment of many types of cancer. Currently, medical images are assessed visually by radiologists and clinicians. However, the full utility of information contained within medical images has yet to be fully explored. One avenue for this exploration is the utilization of "radiomic features" through the application of texture analysis. The numerous radiomic features proposed may vary with confounding variables such as the time post injection of image acquisition and the accuracy of the delineation of the prescribed segmentation volume. To this avail, we propose using the determinant of the correlation matrix to analyze radiomic features robustness to confounding variables. For this purpose, dynamic pre-clinical positron emission tomography (PET) images of 8 mice with mammary carcinoma xenografts (4T1) were binned into 5 minutes intervals from 50 to 70 minutes post injection. The effect of variation in segmentation was also explored by incrementally increasing segmentation volume. From each image set, we extracted 78 Radiomic features for analysis. Analysis. The statistical association measured by the determinant of the correlation matrix when considering contour size was 0.02378; for acquisition time this value was 0.13296. From this analysis we conclude that both temporal variation and segmentation effect the measurement of temporal features and that texture features are less robust to varying acquisition time than to varying segmentation volume.
PloS one, 2017
Textural measures have been widely explored as imaging biomarkers in cancer. However, their robustness under dynamic range and spatial resolution changes in brain 3D magnetic resonance images (MRI) has not been assessed. The aim of this work was to study potential variations of textural measures due to changes in MRI protocols. Twenty patients harboring glioblastoma with pretreatment 3D T1-weighted MRIs were included in the study. Four different spatial resolution combinations and three dynamic ranges were studied for each patient. Sixteen three-dimensional textural heterogeneity measures were computed for each patient and configuration including co-occurrence matrices (CM) features and run-length matrices (RLM) features. The coefficient of variation was used to assess the robustness of the measures in two series of experiments corresponding to (i) changing the dynamic range and (ii) changing the matrix size. No textural measures were robust under dynamic range changes. Entropy was ...
Scientific Reports, 2015
FDG-PET-derived textural features describing intra-tumor heterogeneity are increasingly investigated as imaging biomarkers. As part of the process of quantifying heterogeneity, image intensities (SUVs) are typically resampled into a reduced number of discrete bins. We focused on the implications of the manner in which this discretization is implemented. Two methods were evaluated: (1) R D , dividing the SUV range into D equally spaced bins, where the intensity resolution (i.e. bin size) varies per image; and (2) R B , maintaining a constant intensity resolution B. Clinical feasibility was assessed on 35 lung cancer patients, imaged before and in the second week of radiotherapy. Forty-four textural features were determined for different D and B for both imaging time points. Feature values depended on the intensity resolution and out of both assessed methods, R B was shown to allow for a meaningful inter-and intra-patient comparison of feature values. Overall, patients ranked differently according to feature values-which was used as a surrogate for textural feature interpretation-between both discretization methods. Our study shows that the manner of SUV discretization has a crucial effect on the resulting textural features and the interpretation thereof, emphasizing the importance of standardized methodology in tumor texture analysis.
Connections Between Noise Equivalent Count Rate and Image Noise in Pet Medical Imaging
RAD Conference Proceedings, 2016
Assessing PET image quality is a challenge due to its clinical subjectivity and difficulties in standardisation. Methods to evaluate PET image quality include image noise and noise equivalent count rate (NECR), which are automatic and objective measurements determined from the reconstructed medical PET image or PET raw emission data from phantoms. Although manufacturers emphasize scanner performance by stating peak NECR, the peak NECR is often outside clinically limited by radiation safety factors, as opposed to scanner performance. instead, image noise in the expectation maximization (EM) algorithm inferred that this could also be true for three algorithm. One consequence is that for traditionally based on NECR, should be based on the true produced by increasing the injected activity to match the peak true count rate for 3D
Noise Equivalent Count Rate and Image in Pet Medical Imaging
2016
Assessing PET image quality is a challenge due to its clinical subjectivity and difficulties in standardisation. Methods to evaluate PET image quality include image noise and noise equivalent count rate (NECR), which are automatic and objective measurements determined from the reconstructed medical PET image or PET raw emission data from phantoms. Although manufacturers emphasize scanner performance by stating peak NECR, the peak NECR is often outside clinically limited by radiation safety factors, as opposed to scanner performance. instead, image noise in the expectation maximization (EM) algorithm inferred that this could also be true for three algorithm. One consequence is that for traditionally based on NECR, should be based on the true produced by increasing the injected activity to match the peak true count rate for 3D
Assessment of tumor heterogeneity: an emerging imaging tool for clinical practice
Insights into imaging, 2012
Background Tumor spatial heterogeneity is an important prognostic factor, which may be reflected in medical images Methods Image texture analysis is an approach of quantifying heterogeneity that may not be appreciated by the naked eye. Different methods can be applied including statistical-, model-, and transform-based methods. Results Early evidence suggests that texture analysis has the potential to augment diagnosis and characterization as well as improve tumor staging and therapy response assessment in oncological practice. Conclusion This review provides an overview of the application of texture analysis with different imaging modalities, CT, MRI, and PET, to date and describes the technical challenges that have limited its widespread clinical implementation so far. With further efforts to refine its application, image texture analysis has the potential to develop into a valuable clinical tool for oncologic imaging. Teaching Points • Tumor spatial heterogeneity is an important prognostic factor. • Image texture analysis is an approach of quantifying heterogeneity. • Different methods can be applied, including statistical-, model-, and transform-based methods. • Texture analysis could improve the diagnosis, tumor staging, and therapy response assessment.