Evaluation of radiomic texture feature error due to MRI acquisition and reconstruction: A simulation study utilizing ground truth

The purpose of this study was to examine the dependence of image texture features on MR acquisition parameters and reconstruction using a digital MR imaging phantom. MR signal was simulated in a parallel imaging radiofrequency coil setting as well as a single element volume coil setting, with varying levels of acquisition noise, three acceleration factors, and four image reconstruction algorithms. Twenty-six texture features were measured on the simulated images, ground truth images, and clinical brain images. Subtle algorithm-dependent errors were observed on reconstructed phantom images, even in the absence of added noise. Sources of image error include Gibbs ringing at image edge gradients (tissue interfaces) and well-known artifacts due to high acceleration; two of the iterative reconstruction algorithms studied were able to mitigate these image errors. The difference of the texture features from ground truth, and their variance over reconstruction algorithm and parallel imaging acceleration factor, were compared to the clinical “effect size”, i.e., the feature difference between high- and low-grade tumors on T1- and T2-weighted brain MR images of twenty glioma patients. The measured feature error (difference from ground truth) was small for some features, but substantial for others. The feature variance due to reconstruction algorithm and acceleration factor were generally smaller than the clinical effect size. Certain texture features may be preserved by MR imaging, but adequate precautions need to be taken regarding their validity and reliability. We present a general simulation framework for assessing the robustness and accuracy of radiomic textural features under various MR acquisition/reconstruction scenarios.     

Glioblastomas and brain metastases differentiation following an MRI texture analysis-based radiomics approach

PurposeTo evaluate the potential of 2D texture features extracted from magnetic resonance (MR) images for differentiating brain metastasis (BM) and glioblastomas (GBM) following a radiomics approach.MethodsThis retrospective study included 50 patients with BM and 50 with GBM who underwent T1-weighted MRI between December 2010 and January 2017. Eighty-eight rotation-invariant texture features were computed for each segmented lesion using six texture analysis methods. These features were also extracted from the four images obtained after applying the discrete wavelet transform (88 features × 4 images). Three feature selection methods and five predictive models were evaluated. A 5-fold cross-validation scheme was used to randomly split the study group into training (80 patients) and testing (20 patients), repeating the process ten times. Classification was evaluated computing the average area under the receiver operating characteristic curve. Sensibility, specificity and accuracy were also computed. The whole process was tested quantizing the images with different gray-level values to evaluate their influence in the final results.ResultsHighest classification accuracy was obtained using the original images quantized with 128 gray-levels and a feature selection method based on the p-value. The best overall performance was achieved using a support vector machine model with a subset of 32 features (AUC = 0.896 ± 0.067, sensitivity of 82% and specificity of 80%). Naïve Bayes and k-nearest neighbors models showed also valuable results (AUC ≈ 0.8) with a lower number of features (<13), thus suggesting that these models may be more generalizable when using external validations.ConclusionThe proposed radiomics MRI approach is able to discriminate between GBM and BM with high accuracy employing a set of 2D texture features, thus helping in the diagnosis of brain lesions in a fast and non-invasive way.     

Multi Texture Analysis of Colorectal Cancer Continuum Using Multispectral Imagery

Purpose

This paper proposes to characterize the continuum of colorectal cancer (CRC) using multiple texture features extracted from multispectral optical microscopy images. Three types of pathological tissues (PT) are considered: benign hyperplasia, intraepithelial neoplasia and carcinoma.

Materials and Methods

In the proposed approach, the region of interest containing PT is first extracted from multispectral images using active contour segmentation. This region is then encoded using texture features based on the Laplacian-of-Gaussian (LoG) filter, discrete wavelets (DW) and gray level co-occurrence matrices (GLCM). To assess the significance of textural differences between PT types, a statistical analysis based on the Kruskal-Wallis test is performed. The usefulness of texture features is then evaluated quantitatively in terms of their ability to predict PT types using various classifier models.

Results

Preliminary results show significant texture differences between PT types, for all texture features (p-value < 0.01). Individually, GLCM texture features outperform LoG and DW features in terms of PT type prediction. However, a higher performance can be achieved by combining all texture features, resulting in a mean classification accuracy of 98.92%, sensitivity of 98.12%, and specificity of 99.67%.

Conclusions

These results demonstrate the efficiency and effectiveness of combining multiple texture features for characterizing the continuum of CRC and discriminating between pathological tissues in multispectral images.     

超声图像处理中Snake模型研究

Snake模型是一种基于高层信息的有效目标轮廓提取算法,其优点是作用过程及最后结果的目标轮廓是一条完整的曲线,从而引起广泛的关注。鉴于医学超声图像的信噪比较低,用经典的边缘提取算法无法得到较好的结果,因此人们将Snake模型进行了各种各样的改进,并且越来越多地将它运用到医学超声图像处理中来。本文对乳腺超声图像进行阈值分割、形态滤波等一系列预处理后,将改进的Snake模型对乳腺超声图像进行肿瘤的边缘提取,得到了比较好的结果。     

Multiscale optical flow computation from the monogenic signal

We have developed an algorithm for the estimation of cardiac motion from medical images. The algorithm exploits monogenic signal theory, recently introduced as an N-dimensional generalization of the analytic signal. The displacement is computed locally by assuming the conservation of the monogenic phase over time. A local affine displacement model replaces the standard translation model to account for more complex motions as contraction/expansion and shear. A coarse-to-fine B-spline scheme allows a robust and effective computation of the models parameters and a pyramidal refinement scheme helps handle large motions. Robustness against noise is increased by replacing the standard pointwise computation of the monogenic orientation with a more robust least-squares orientation estimate. This paper reviews the results obtained on simulated cardiac images from different modalities, namely 2D and 3D cardiac ultrasound and tagged magnetic resonance. We also show how the proposed algorithm represents a valuable alternative to state-of-the-art algorithms in the respective fields.     

La géométrie fractale pour l’analyse de signaux médicaux : état de l’art

The diagnostic interpretation of medical images is a complex task aiming to detect potential abnormalities. One of the most used features in this process is texture which is a key component in the human understanding of images. Many studies were conducted to develop algorithms for texture quantification. The relevance of fractal geometry in medical image analysis is justified by the proven self-similarity of anatomical objects when imaged with a finite resolution. Over the last years, fractal geometry was applied extensively in many medical signal analysis applications. The use of these geometries relies heavily on estimation of the fractal features. Various methods were proposed to estimate the fractal dimension or multifractal spectrum of a signal. This article presents an overview of these algorithms, the way they work, their benefits and limits, and their application in the field of medical signal analysis.     

Human fetuses have nonlinear cardiac dynamics.

Approximate entropy (ApEn) is a statistic that quantifies regularity in time series data, and this parameter has several features that make it attractive for analyzing physiological systems. In this study, ApEn was used to detect nonlinearities in the heart rate (HR) patterns of 12 low-risk human fetuses between 38 and 40 wk of gestation. The fetal cardiac electrical signal was sampled at a rate of 1,024 Hz by using Ag-AgCl electrodes positioned across the mother's abdomen, and fetal R waves were extracted by using adaptive signal processing techniques. To test for nonlinearity, ApEn for the original HR time series was compared with ApEn for three dynamic models: temporally uncorrelated noise, linearly correlated noise, and linearly correlated noise with nonlinear distortion. Each model had the same mean and SD in HR as the original time series, and one model also preserved the Fourier power spectrum. We estimated that noise accounted for 17.2-44.5% of the total between-fetus variance in ApEn. Nevertheless, ApEn for the original time series data still differed significantly from ApEn for the three dynamic models for both group comparisons and individual fetuses. We concluded that the HR time series, in low-risk human fetuses, could not be modeled as temporally uncorrelated noise, linearly correlated noise, or static filtering of linearly correlated noise.     

Asymmetries in visual texture discrimination

Recent results have shown that texture discrimination is an asymmetrical process; texture A within texture B may be much easier to detect than texture B within texture A. Two questions regarding discrimination asymmetries are addressed: (i) what sorts of textural properties are associated with discrimination asymmetries; and (ii) what sort of architecture would yield asymmetries. Two experiments show that discrimination asymmetries obtain when textures comprise circles of different sizes (large circles are easier to detect in small than vice versa) and when circles differ only in the regularity of their placement (irregularly placed circles are easier to detect in a background of regularly placed circles than vice versa). A plausible account of texture discrimination would involve the decomposition of images via a set orientation and scale selective filters followed by a second layer of filtering to detect energy differences between adjacent regions in the original convolutions. Discrimination asymmetries provide prima facie evidence against such a model because it involves only local measurements and comparisons. We propose that discrimination asymmetries are elegantly explained if it is assumed that the responses of the orientation and scale selective filters are normalized by the degree to which similarly tuned operators are responding elsewhere in the image; viz., global normalization of filter responses. However, there are cases where such global normalization is not required to explain asymmetrical discrimination.     

Contourlet Textual Features: Improving the Diagnosis of Solitary Pulmonary Nodules in Two Dimensional CT Images

Objective

To determine the value of contourlet textural features obtained from solitary pulmonary nodules in two dimensional CT images used in diagnoses of lung cancer.

Materials and Methods

A total of 6,299 CT images were acquired from 336 patients, with 1,454 benign pulmonary nodule images from 84 patients (50 male, 34 female) and 4,845 malignant from 252 patients (150 male, 102 female). Further to this, nineteen patient information categories, which included seven demographic parameters and twelve morphological features, were also collected. A contourlet was used to extract fourteen types of textural features. These were then used to establish three support vector machine models. One comprised a database constructed of nineteen collected patient information categories, another included contourlet textural features and the third one contained both sets of information. Ten-fold cross-validation was used to evaluate the diagnosis results for the three databases, with sensitivity, specificity, accuracy, the area under the curve (AUC), precision, Youden index, and F-measure were used as the assessment criteria. In addition, the synthetic minority over-sampling technique (SMOTE) was used to preprocess the unbalanced data.

Results

Using a database containing textural features and patient information, sensitivity, specificity, accuracy, AUC, precision, Youden index, and F-measure were: 0.95, 0.71, 0.89, 0.89, 0.92, 0.66, and 0.93 respectively. These results were higher than results derived using the database without textural features (0.82, 0.47, 0.74, 0.67, 0.84, 0.29, and 0.83 respectively) as well as the database comprising only textural features (0.81, 0.64, 0.67, 0.72, 0.88, 0.44, and 0.85 respectively). Using the SMOTE as a pre-processing procedure, new balanced database generated, including observations of 5,816 benign ROIs and 5,815 malignant ROIs, and accuracy was 0.93.

Conclusion

Our results indicate that the combined contourlet textural features of solitary pulmonary nodules in CT images with patient profile information could potentially improve the diagnosis of lung cancer.     

Retrieval and Mapping of Soil Texture Based on Land Surface Diurnal Temperature Range Data from MODIS

Numerous studies have investigated the direct retrieval of soil properties, including soil texture, using remotely sensed images. However, few have considered how soil properties influence dynamic changes in remote images or how soil processes affect the characteristics of the spectrum. This study investigated a new method for mapping regional soil texture based on the hypothesis that the rate of change of land surface temperature is related to soil texture, given the assumption of similar starting soil moisture conditions. The study area was a typical flat area in the Yangtze-Huai River Plain, East China. We used the widely available land surface temperature product of MODIS as the main data source. We analyzed the relationships between the content of different particle soil size fractions at the soil surface and land surface day temperature, night temperature and diurnal temperature range (DTR) during three selected time periods. These periods occurred after rainfalls and between the previous harvest and the subsequent autumn sowing in 2004, 2007 and 2008. Then, linear regression models were developed between the land surface DTR and sand (> 0.05 mm), clay (< 0.001 mm) and physical clay (< 0.01 mm) contents. The models for each day were used to estimate soil texture. The spatial distribution of soil texture from the studied area was mapped based on the model with the minimum RMSE. A validation dataset produced error estimates for the predicted maps of sand, clay and physical clay, expressed as RMSE of 10.69%, 4.57%, and 12.99%, respectively. The absolute error of the predictions is largely influenced by variations in land cover. Additionally, the maps produced by the models illustrate the natural spatial continuity of soil texture. This study demonstrates the potential for digitally mapping regional soil texture variations in flat areas using readily available MODIS data.     

Ultrasonography to Quantify Hepatic Fat Content: Validation by 1H Magnetic Resonance Spectroscopy

An abundance of fat stored within the liver, or steatosis, is the beginning of a broad hepatological spectrum, usually referred to as fatty liver disease (FLD). For studies on FLD, quantitative hepatic fat ultrasonography would be an appealing study modality. Objective of this study was to develop a technique for quantifying hepatic fat content by ultrasonography and validate this using proton magnetic resonance spectroscopy (¹H MRS) as gold standard. Eighteen white volunteers (BMI range 21.0–42.9) were scanned by both ultrasonography and ¹H MRS. Altered ultrasound characteristics, present in the case of FLD, were assessed using a specially developed software program. Various attenuation and textural based indices of FLD were extracted from ultrasound images. Using linear regression analysis, the predictive power of several models (consisting of both attenuation and textural based measures) on log 10–transformed hepatic fat content by ¹H MRS were investigated. The best quantitative model was compared with a qualitative ultrasonography method, as used in clinical care. A model with four ultrasound characteristics could modestly predict the amount of liver fat (adjusted explained variance 43.2%, P = 0.021). Expanding the model to seven ultrasound characteristics increased adjusted explained variance to 60% (P = 0.015), with r = 0.789 (P < 0.001). Comparing this quantitative model with qualitative ultrasonography revealed a significant advantage of the quantitative model in predicting hepatic fat content (P < 0.001). This validation study shows that a combination of computer‐assessed ultrasound measures from routine ultrasound images can be used to quantitatively assess hepatic fat content.     

False Discovery Rates in PET and CT Studies with Texture Features: A Systematic Review

PurposeA number of recent publications have proposed that a family of image-derived indices, called texture features, can predict clinical outcome in patients with cancer. However, the investigation of multiple indices on a single data set can lead to significant inflation of type-I errors. We report a systematic review of the type-I error inflation in such studies and review the evidence regarding associations between patient outcome and texture features derived from positron emission tomography (PET) or computed tomography (CT) images.MethodsFor study identification PubMed and Scopus were searched (1/2000–9/2013) using combinations of the keywords texture, prognostic, predictive and cancer. Studies were divided into three categories according to the sources of the type-I error inflation and the use or not of an independent validation dataset. For each study, the true type-I error probability and the adjusted level of significance were estimated using the optimum cut-off approach correction, and the Benjamini-Hochberg method. To demonstrate explicitly the variable selection bias in these studies, we re-analyzed data from one of the published studies, but using 100 random variables substituted for the original image-derived indices. The significance of the random variables as potential predictors of outcome was examined using the analysis methods used in the identified studies.ResultsFifteen studies were identified. After applying appropriate statistical corrections, an average type-I error probability of 76% (range: 34–99%) was estimated with the majority of published results not reaching statistical significance. Only 3/15 studies used a validation dataset. For the 100 random variables examined, 10% proved to be significant predictors of survival when subjected to ROC and multiple hypothesis testing analysis.ConclusionsWe found insufficient evidence to support a relationship between PET or CT texture features and patient survival. Further fit for purpose validation of these image-derived biomarkers should be supported by appropriate biological and statistical evidence before their association with patient outcome is investigated in prospective studies.     

Nonmonotonic noise tuning of BOLD fMRI signal to natural images in the visual cortex of the anesthetized monkey

BACKGROUND: The perceptual ability of humans and monkeys to identify objects in the presence of noise varies systematically and monotonically as a function of how much noise is introduced to the visual display. That is, it becomes more and more difficult to identify an object with increasing noise. Here we examine whether the blood oxygen level-dependent functional magnetic resonance imaging (BOLD fMRI) signal in anesthetized monkeys also shows such monotonic tuning. We employed parametric stimulus sets containing natural images and noise patterns matched for spatial frequency and intensity as well as intermediate images generated by interpolation between natural images and noise patterns. Anesthetized monkeys provide us with the unique opportunity to examine visual processing largely in the absence of top-down cognitive modulations and can thus provide an important baseline against which work with awake monkeys and humans can be compared. RESULTS: We measured BOLD activity in occipital visual cortical areas as natural images and noise patterns, as well as intermediate interpolated patterns at three interpolation levels (25%, 50%, and 75%) were presented to anesthetized monkeys in a block paradigm. We observed reliable visual activity in occipital visual areas including V1, V2, V3, V3A, and V4 as well as the fundus and anterior bank of the superior temporal sulcus (STS). Natural images consistently elicited higher BOLD levels than noise patterns. For intermediate images, however, we did not observe monotonic tuning. Instead, we observed a characteristic V-shaped noise-tuning function in primary and extrastriate visual areas. BOLD signals initially decreased as noise was added to the stimulus but then increased again as the pure noise pattern was approached. We present a simple model based on the number of activated neurons and the strength of activation per neuron that can account for these results. CONCLUSIONS: We show that, for our parametric stimulus set, BOLD activity varied nonmonotonically as a function of how much noise was added to the visual stimuli, unlike the perceptual ability of humans and monkeys to identify such stimuli. This raises important caveats for interpreting fMRI data and demonstrates the importance of assessing not only which neural populations are activated by contrasting conditions during an fMRI study, but also the strength of this activation. This becomes particularly important when using the BOLD signal to make inferences about the relationship between neural activity and behavior.     

Comparative study on the performance of textural image features for active contour segmentation

We present a computerized method for the semi-automatic detection of contours in ultrasound images.The novelty of our study is the introduction of a fast and efficient image function relating to parametric active contour models.This new function is a combination of the gray-level information and first-order statistical features,called standard deviation parameters.In a comprehensive study,the developed algorithm and the efficiency of segmentation were first tested for synthetic images.Tests were also performed on breast and liver ultrasound images.The proposed method was compared with the watershed approach to show its efficiency.The performance of the segmentation was estimated using the area error rate.Using the standard deviation textural feature and a 5×5 kernel,our curve evolution was able to produce results close to the minimal area error rate(namely 8.88% for breast images and 10.82% for liver images).The image resolution was evaluated using the contrast-to-gradient method.The experiments showed promising segmentation results.     

Characterization of cervical lymph-nodes using a multi-parametric and multi-modal approach for an early prediction of tumor response to chemo-radiotherapy

PurposeIn the treatment of Head-and-Neck Squamous Cell Carcinoma (HNSCC), the early prediction of residual malignant lymph nodes (LNs) is currently required. Here, we investigated the potential of a multi-modal characterization (combination of CT, T2w-MRI and DW-MRI) at baseline and at mid-treatment, based on texture analysis (TA), for the early prediction of LNs response to chemo-radiotherapy (CRT).Methods30 patients with pathologically confirmed HNSCC treated with CRT were considered. All patients underwent a planning CT and two serial MR examinations (including T2w and DW images), one before and one at mid-CRT. For each patient the largest malignant LN was selected and within each LN, morphological and textural features were estimated from T2w-MRI and CT, besides a quantification of the apparent diffusion coefficient (ADC) from DW-MRI. After a median follow-up time of 26.6 months, 19 LNs showed regional control, while 11 LNs showed regional failure at a median time of 4.6 months. Linear discriminant analysis was used to test the accuracy of the image-based features in predicting the final response.ResultsPre-treatment features showed higher predictive power than mid-CRT features, the ADC having the highest accuracy (80%); CT-based indices were found not predictive. When ADC was combined with TA, the classification performance increased (accuracy = 82.8%). If only T2w-MRI features were considered, the best combination of pre-CRT indices and their variation reached an equivalent accuracy (81.8%).ConclusionOur results may suggest that TA on T2w-MRI and ADC can be combined together to obtain a more accurate prediction of response to CRT.     

Discrimination of complex textures by bees

A problem confronted by visual systems is that of discriminating textures. It appears that a recently described class of orientation-tuned neurones in the bee brain embody properties of mechanisms used by humans to discriminate complex textures. In particular these mechanisms would permit bees to discriminate a large range of textures by giving bees access to information related to higher-order correlations between texture elements. To determine if bees can exploit such textural information we have conducted behavioural experiments employing iso-dipole textures, that statistically speaking, differ from binary noise textures, and each other, only in their third-order correlation functions. While these textures are not themselves of any ethological significance their special properties permit us to show that bees can potentially use a very large palette of textures to classify textured objects. In electrophysiological experiments we demonstrate the requisite contrast sign invariance (rectification) of the orientation-selective neurones' responses and discuss other similarities of these neurones' responses to models accounting for human texture discrimination. Accepted: 7 October 1998     

Classification of Tree Species in Overstorey Canopy of Subtropical Forest Using QuickBird Images

This paper proposes a supervised classification scheme to identify 40 tree species (2 coniferous, 38 broadleaf) belonging to 22 families and 36 genera in high spatial resolution QuickBird multispectral images (HMS). Overall kappa coefficient (OKC) and species conditional kappa coefficients (SCKC) were used to evaluate classification performance in training samples and estimate accuracy and uncertainty in test samples. Baseline classification performance using HMS images and vegetation index (VI) images were evaluated with an OKC value of 0.58 and 0.48 respectively, but performance improved significantly (up to 0.99) when used in combination with an HMS spectral-spatial texture image (SpecTex). One of the 40 species had very high conditional kappa coefficient performance (SCKC ≥ 0.95) using 4-band HMS and 5-band VIs images, but, only five species had lower performance (0.68 ≤ SCKC ≤ 0.94) using the SpecTex images. When SpecTex images were combined with a Visible Atmospherically Resistant Index (VARI), there was a significant improvement in performance in the training samples. The same level of improvement could not be replicated in the test samples indicating that a high degree of uncertainty exists in species classification accuracy which may be due to individual tree crown density, leaf greenness (inter-canopy gaps), and noise in the background environment (intra-canopy gaps). These factors increase uncertainty in the spectral texture features and therefore represent potential problems when using pixel-based classification techniques for multi-species classification.     

Practical limits for reverse engineering of dynamical systems: a statistical analysis of sensitivity and parameter inferability in systems biology models

The size and complexity of cellular systems make building predictive models an extremely difficult task. In principle dynamical time-course data can be used to elucidate the structure of the underlying molecular mechanisms, but a central and recurring problem is that many and very different models can be fitted to experimental data, especially when the latter are limited and subject to noise. Even given a model, estimating its parameters remains challenging in real-world systems. Here we present a comprehensive analysis of 180 systems biology models, which allows us to classify the parameters with respect to their contribution to the overall dynamical behaviour of the different systems. Our results reveal candidate elements of control in biochemical pathways that differentially contribute to dynamics. We introduce sensitivity profiles that concisely characterize parameter sensitivity and demonstrate how this can be connected to variability in data. Systematically linking data and model sloppiness allows us to extract features of dynamical systems that determine how well parameters can be estimated from time-course measurements, and associates the extent of data required for parameter inference with the model structure, and also with the global dynamical state of the system. The comprehensive analysis of so many systems biology models reaffirms the inability to estimate precisely most model or kinetic parameters as a generic feature of dynamical systems, and provides safe guidelines for performing better inferences and model predictions in the context of reverse engineering of mathematical models for biological systems.