Comparison of automated and visual texture analysis in MRI: characterization of normal and diseased skeletal muscle

Automated magnetic resonance imaging (MRI) texture analysis was compared with visual MRI analysis for the diagnosis of skeletal muscle dystrophy in 14 healthy and 17 diseased subjects. MRI texture analysis was performed on 8 muscle regions of interest (ROI) using four statistical methods (histogram, co-occurrence matrix, gradient matrix, runlength matrix) and one structural (mathematical morphology) method. Nine senior radiologists assessed full leg transverse slice images and proposed a diagnosis. The 59 extracted texture parameters for each ROI were statistically analyzed by Correspondence Factorial Analysis. Non-parametric tests were used to compare diagnoses based on automated texture analysis and visual analysis. Texture analysis methods discriminated between healthy volunteers and patients with a sensitivity of 70%, and a specificity of 86%. Comparison with visual analysis of MR images suggests that texture analysis can provide useful information contributing to the diagnosis of skeletal muscle disease.     

Wavelet-based characterization of vertebral trabecular bone structure from magnetic resonance images at 3 T compared with micro-computed tomographic measurements

Trabecular bone structure and bone density contribute to the strength of bone and are important in the study of osteoporosis. Wavelets are a powerful tool in characterizing and quantifying texture in an image. The purpose of this study was to validate wavelets as a tool in computing trabecular bone thickness directly from gray-level images. To this end, eight cylindrical cores of vertebral trabecular bone were imaged using 3-T magnetic resonance imaging (MRI) and micro-computed tomography (microCT). Thickness measurements of the trabecular bone from the wavelet-based analysis were compared with standard 2D structural parameters analogous to bone histomorphometry (MR images) and direct 3D distance transformation methods (microCT images). Additionally, bone volume fraction was determined using each method. The average difference in trabecular thickness between the wavelet and standard methods was less than the size of 1 pixel size for both MRI and microCT analysis. A correlation (R) of .94 for microCT measurements and that of .52 for MRI were found for the bone volume fraction. Based on these results, we conclude that wavelet-based methods deliver results comparable with those from established MR histomorphometric measurements. Because the wavelet transform is more robust with respect to image noise and operates directly on gray-level images, it could be a powerful tool for computing structural bone parameters from MR images acquired using high resolution and thus limited signal scenarios.     

MRI texture analysis on texture test objects, normal brain and intracranial tumors

Texture analysis was performed in three different MRI units on T1 and T2-weighted MR images from 10 healthy volunteers and 63 patients with histologically confirmed intracranial tumors. The goal of this study was a multicenter evaluation of the usefulness of this quantitative approach for the characterization of healthy and pathologic human brain tissues (white matter, gray matter, cerebrospinal fluid, tumors and edema). Each selected brain region of interest was characterized with both its mean gray level values and several texture parameters. Multivariate statistical analyses were then applied in order to discriminate each brain tissue type represented by its own set of texture parameters. Texture analysis was previously performed on test objects to evaluate the method dependence on acquisition parameters and consequently the interest of a multicenter evaluation. Even obtained on different sites with their own acquisition routine protocol, MR brain images contain textural features that can reveal discriminant factors for tissue classification and image segmentation. It can also offer additional information in case of undetermined diagnosis or to develop a more accurate tumor grading.     

Effects of trabecular bone on marrow relaxation in the tibia

The effects of bone on marrow relaxation in the trabecular volume of the most proximal 3 cm in the left tibia were studied with a RF-spoiled gradient echo MRI protocol on a 1.0 T MR unit. The MR measurements were performed on six healthy volunteers, and repeated within one month in order to assess the precision of the method. In the same subjects, the area bone mineral density (bmd, g/cm²) was measured at the left proximal femur using dual-energy X-ray absorptiometry. The calcaneus of the same side was examined with quantitative ultrasound. The marrow T21 relaxation deviated from a mono-exponential decay, and resembled the decay of subcutaneous fat. The shape of the relaxation curve reflected the presence of several spectral components in bone marrow, and was further influenced by the amount and structure of the surrounding trabecular bone. The bone marrow decays showed substantially reduced inter-subject variability after normalisation of the marrow data fit parameters to corresponding values for s.c. fat. This suggests the use of an internal adipose tissue reference in order to correct for diet-related variations of marrow T21 estimates. The mean relative precision of the MR measurements was between 5% and 10% depending on the data fit model. Moderate-to-strong correlations between DXA bmd indices in the proximal femur and MR parameters were found (r_max = −0.96; p < 0.01), while ultrasound-derived measures of bone strength measured on the calcaneus demonstrated significantly weaker correlations to the MR parameters (r_max = −0.78; p > 0.05). The method employed in this study showed reasonable precision and a moderate to good correlation compared to other bone parameters derived at the same extremity, and is a promising tool for the use on patients.     

Characterization of trabecular bone structure from high-resolution magnetic resonance images using fuzzy logic

The purpose of this work was to apply fuzzy logic image processing techniques to characterize the trabecular bone structure with high-resolution magnetic resonance images. Fifteen ex vivo high-resolution magnetic resonance images of specimens of human radii at 1.5 T and 12 in vivo high-resolution magnetic resonance images of the calcanei of peri- and postmenopausal women at 3 T were obtained. Soft segmentation using fuzzy clustering was applied to MR data to obtain fuzzy bone volume fraction maps, which were then analyzed with three-dimensional (3D) fuzzy geometrical parameters and measures of fuzziness. Geometrical parameters included fuzzy perimeter and fuzzy compactness, while measures of fuzziness included linear index of fuzziness, quadratic index of fuzziness, logarithmic fuzzy entropy, and exponential fuzzy entropy. Fuzzy parameters were validated at 1.5 T with 3D structural parameters computed from microcomputed tomography images, which allow the observation of true trabecular bone structure and with apparent MR structural indexes at 1.5 T and 3 T. The validation was statistically performed with the Pearson correlation coefficient as well as with the Bland-Altman method. Bone volume fraction correlation values (r) were up to .99 (P<.001) with good agreements based on Bland-Altman analysis showing that fuzzy clustering is a valid technique to quantify this parameter. Measures of fuzziness also showed consistent correlations to trabecular number parameters (r>.85; P<.001) and good agreements based on Bland-Altman analysis, suggesting that the level of fuzziness in high-resolution magnetic resonance images could be related to the trabecular bone structure.     

In vivo multiple spin echoes imaging of trabecular bone on a clinical 1.5 T MR scanner

In vivo multiple spin echoes (MSE) images of bone marrow in trabecular bone were obtained for the first time on a clinical 1.5 T scanner. Despite of a reduced sensitivity of the MSE trabecular bone images with respect to the cerebral matter ones, it is possible to observe some features in the MSE trabecular bone images that may be useful in the diagnosis of osteopenic states. Two different CRAZED-type MSE imaging sequences based on spin-echo and EPI imaging modalities were applied in phantom and in vivo. Preliminary experimental results indicate that EPI imaging readout seems to conceal the MSE contrast correlated with pore dimension in porous media. However it is still possible to detect anisotropy effects related to the bone structure in MSE-EPI images. Some strategies are suggested to optimize the quality of MSE trabecular bone images.     

Comparison of high-resolution MRI, optical microscopy and SEM for quantitation of trabecular architecture in the rat femur

Magnetic resonance imaging (MRI) has been used to analyze trabecular bone architecture in femur heads taken from adult Wistar rats. The aim of this study was to validate the use of MRI in assessing trabecular structure and morphology by comparing standard measures of bone morphology in the rat femur obtained from high resolution MRI with those obtained by conventional optical microscopy and by scanning electron microscopy (SEM). MR images were obtained on a Bruker 4.7 T micro-imaging system using a three-dimensional spin echo sequence with spatial resolution of 23 microm in-plane and a slice thickness of 39 microm. Optical images were obtained by de-calcifying the bone in EDTA and then sectioning 5-microm-thick slices. SEM images were obtained from bone embedded in epoxy resin with surface preparation by diamond polishing. Values of standard bone morphological parameters were compared and correlation coefficients between the MRI and the optical- and SEM-derived measures of morphology were calculated. Partial volume effects in MRI were minimized in this study by the use of very thin slices, yielding better agreement with optical- and SEM-derived measures of trabecular bone morphology than have been obtained in previous studies. Correlations between the MRI and optical data were significantly lower than those between the MRI and SEM data. Effects of de-calcification were also investigated. The results indicate that comparison of MRI with thin (de-calcified) optical images may be inherently flawed due to the destructive de-calcification and sectioning process used to prepare samples for the optical imaging.     

Aortoiliac imaging by projective phase sensitive MR angiography: effects of triggering and timing of data acquisition on image quality

To assess the ability of projective phase sensitive magnetic resonance (MR) angiography to visualize the aortoiliac vascular segment, and to determine the effects of triggering and timing of data acquisition om image quality, we studied 18 healthy volunteers, mean age 33.3 +/- 11 years, by color Doppler imaging and by MR angiography. MR angiography was performed at 1.5 T using a flow-adjustable gradient-echo (FLAG) sequence operated in both ECG-triggered and non-triggered acquisition modes. The images were graded in a blinded fashion by two independent observers. The data were analyzed using Pearson's chi-square analysis. Eighteen triggered time-resolved and 17 non-triggered, time-averaged MR angiograms consisting of 252 and 17 angiographic images, (AI) respectively, were analyzed. In the triggered mode 69 (27.4%) AI and in the non-triggered mode 2 (11.8%) AI were diagnostic. At least one triggered diagnostic AI was obtained in each subject. The image grades were not statistically different between observers (kappa = 0.6686). In the triggered mode diagnostic images were acquired within +/- 90 msec of the peak systolic flow velocity determined by Doppler. The proportion of diagnostic images in the triggered mode was highest (73.3%) within a 30-msec interval before the peak flow. In healthy subjects the aortoiliac segment is reliably visualized by FLAG MR angiography. The optimum results are achieved using the triggered acquisition mode and timing acquisition to the initial 180 msec of the abdominal aortic systolic flow pulse.     

Trabecular bone volume fraction measurements of a large number of subjects using a compact MRI

Trabecular bone volume fraction (TBVF) and speed of sound (SOS) were measured for the right calcanei of 416 female volunteers. The TBVF was measured with a compact MRI developed in our laboratory. The SOS was measured with a commercial quantitative ultrasound system. It was observed that the correlation coefficient between TBVF and SOS and that between TBVF and age varied depending on the location of region of interest (ROI) in the calcaneus. As a result, an optimum circular ROI with a diameter of 20 mm was determined so that the correlation coefficients were maximized. In the optimum ROI, transverse relaxation time (T₂) of the bone marrow protons of the calcaneus was found to be concentrated in a narrow range over the subjects. This result suggested that a 50% scan time reduction in the TBVF measurements could be made by skipping the T₂ correction procedure.     

Reduced field-of-view DTI segmentation of cervical spine tissue

The number of diffusion tensor imaging (DTI) studies regarding the human spine has considerably increased and it is challenging because of the spine’s small size and artifacts associated with the most commonly used clinical imaging method. A novel segmentation method based on the reduced field-of-view (rFOV) DTI dataset is presented in cervical spinal canal cerebrospinal fluid, spinal cord grey matter and white matter classification in both healthy volunteers and patients with neuromyelitis optica (NMO) and multiple sclerosis (MS). Due to each channel based on high resolution rFOV DTI images providing complementary information on spinal tissue segmentation, we want to choose a different contribution map from multiple channel images. Via principal component analysis (PCA) and a hybrid diffusion filter with a continuous switch applied on fourteen channel features, eigen maps can be obtained and used for tissue segmentation based on the Bayesian discrimination method. Relative to segmentation by a pair of expert readers, all of the automated segmentation results in the experiment fall in the good segmentation area and performed well, giving an average segmentation accuracy of about 0.852 for cervical spinal cord grey matter in terms of volume overlap. Furthermore, this has important applications in defining more accurate human spinal cord tissue maps when fusing structural data with diffusion data. rFOV DTI and the proposed automatic segmentation outperform traditional manual segmentation methods in classifying MR cervical spinal images and might be potentially helpful for detecting cervical spine diseases in NMO and MS.     

应用高光谱图像光谱和纹理特征的番茄早疫病早期检测研究

提出了应用光谱和纹理特征的高光谱成像技术早期检测番茄叶片早疫病的方法。利用高光谱图像采集系统获取380～1 030 nm范围内71个染病和88个健康番茄叶片的高光谱图像,同时采用主成分分析法(PCA)对高光谱图像进行处理。选取染病和健康叶片感兴趣区域(region of interest, ROI)的光谱反射率值,同时分别从前8个主成分的每幅主成分图像的ROI中提取对比度(Contrast)、 相关性(Correlation)、 熵(Entropy)和同质性(Homogeneity)4个灰度共生矩阵的纹理特征值,再通过PCA和连续投影算法(SPA)结合最小二乘支持向量机(LS-SVM)构建番茄叶片早疫病的早期鉴别模型。建立的6个模型中,采用光谱反射率值的LS-SVM模型对番茄叶片早疫病的识别率最高,达到100%。结果表明,应用高光谱成像技术检测番茄叶片早疫病是可行的。     

Existence of contralateral abnormalities revealed by texture analysis in unilateral intractable hippocampal epilepsy

We selected 23 patients with unilateral temporal lobe epilepsy characterized by ipsilateral hippocampal sclerosis and an apparently normal contralateral hippocampus on MR imaging. Images were acquired on a 0.28 T MR scanner using a conventional Carr-Purcell Meiboom Gill sequence in all patients and in 9 healthy subjects. Texture analysis was applied to axial MR images of the first and tenth echoes. Texture analysis detects macroscopic lesions and microscopic abnormalities that can not be observed visually. The presence of texture differences in the between normal (controls) and sclerotic hippocampi was ascertained by statistical discriminant analysis. The apparently normal contralateral hippocampi can be classified into three categories in terms of texture: 4 apparently healthy, 8 similar to sclerosis, and 11 different from either healthy or sclerosis. These findings are related to a certain degree of hippocampal alteration, which further investigation might help better characterize.     

The TEM characterization of the lamellar structure of osteoporotic human trabecular bone

The lamellar structure of osteoporotic human trabecular bone was characterized experimentally by means of transmission electron microscopy (TEM). More specifically, the TEM was used to determine if trabecular bone exhibits similar lamellar structural motifs as cortical bone by analyzing unmineralized, mineralized and demineralized bone, and to study the influence of the osteocyte network on the lamellar structure of osteoporotic trabecular bone. Comparison with normal trabecular bone is included. This paper summarizes partial results of a larger study, which addressed the characterization of the hierarchical structure of normal versus osteoporotic human trabecular bone [Rubin, M.A., 2001. Multiscale characterization of the ultrastructure of trabecular bone in osteoporotic and normal humans and in two inbred strains of mice. MS Thesis, Georgia Institute of Technology.] at several structural scales.     

Detection of late epilepsy by the texture analysis of MR brain images in the lithium-pilocarpine rat model

The aim of the study was to detect by texture analysis non easily visible anomalies of magnetic resonance (MR) images of piriform and entorhinal cortices relevant to the lithium-pilocarpine (Li-Pilo) model of temporal lobe epilepsy in rats. Status epilepticus was induced by Li-Pilo in twenty male rats 21 day-old. T₂-weighted MR images of their brain, were obtained before injection of Li-Pilo and one day after status epilepticus. An hyperintense signal was found in the piriform and entorhinal cortices of six rats, which developed chronic epilepsy after a latent period of one to three months. Among the 14 other rats which displayed images similar to those obtained before injection, four remained healthy but 10 rats developed late epileptic symptoms, raising the problem of hidden cortical damage which may be too subtle to be detected by classic MRI examination. A numeric treatment of digital images was then undertaken by texture analysis, to derive image information from a purely computational point of view. The combined texture and discriminant analyses based on pixels pattern anomalies, selected 3 texture parameters derived from co-occurrence matrix which characterized structural abnormalities relevant to the hyperintense signal, not only in the modified images of 6 rats but also in images of 10 rats with apparently non modified images. These three texture’s parameters allowed to classify the twenty rats of our experiment as follows: sixteen epileptic rats were effectively classified with cortical lesions, two non epileptic were correctly classified with healthy cortex, but two healthy rats were not correctly classified. This misclassification is discussed on the basis of the time dependence of the onset of seizure in the Li-Pilo model. These promising results suggest to apply this method to MRI examinations for an improvement of the early diagnostic of human epilepsy.     

Reproducibility of the quantitative assessment of cartilage morphology and trabecular bone structure with magnetic resonance imaging at 7 T

To assess the reproducibility of quantitative measurements of cartilage morphology and trabecular bone structure of the knee at 7 T, high-resolution sagittal spoiled gradient-echo images and high-resolution axial fully refocused steady-state free-precession (SSFP) images from six healthy volunteers were acquired with a 7-T scanner. The subjects were repositioned between repeated scans to test the reproducibility of the measurements. The reproducibility of each measurement was evaluated using the coefficient(s) of variation (CV). The computed CV were 1.13% and 1.55% for cartilage thickness and cartilage volume, respectively, and were 2.86%, 1.07%, 2.27% and 3.30% for apparent bone volume over total volume fraction (app.BV/TV), apparent trabecular number (app.Tb.N), apparent trabecular separation (app.Tb.Sp) and apparent trabecular thickness (app.Tb.Th), respectively. The results demonstrate that quantitative assessment of cartilage morphology and trabecular bone structure is reproducible at 7 T and motivates future musculoskeletal applications seeking the high-field strength's superior signal-to-noise ratio.     

竞争性自适应重加权算法和相关系数法提取特征波长检测番茄叶片真菌病害

基于竞争性自适应重加权算法(CARS)和相关系数法(CA)特征波长选择方法,提出了利用可见-近红外高光谱成像技术检测番茄叶片灰霉病的方法。首先获取380～1 023 nm波段范围内80个染病和80个健康番茄叶片的高光谱图像,然后提取染病和健康叶片感兴趣区域(ROI)的光谱反射率值,作为番茄叶片灰霉病鉴别模型的输入来建立支持向量机(SVM)鉴别模型,训练集和验证集的鉴别率都是100%。研究进一步通过CARS和CA提取特征波长,分别得到5个(554, 694, 696, 738和880 nm)和4个(527, 555, 571和633 nm)特征波长,然后分别建立CARS-SVM和CA-SVM鉴别模型。结果显示,CARS-SVM模型中训练集和验证集的鉴别率都是100%,CA-SVM模型中训练集和验证集的鉴别率分别是91.59%和92.45%。以上结果说明了从可见-近红外高光谱图像中提取的光谱反射率值用于检测番茄叶片的灰霉病是可行的。     

Three-dimensional phase contrast MR cerebral venography with zero filling interpolation in the slice encoding direction.

The purpose of this study was to evaluate the magnetic resonance (MR) cerebral venography findings of a three-dimensional phase contrast MR sequence with zero filling interpolation of the data in the slice encoding direction. Fifty volunteers were enrolled in the study. Images were obtained on a 1.5 MR imaging system with acquisition time of 12 min. MIP images were reconstructed throughout the entire imaging volume. A grading scale system was used to assess dural venous sinuses, major deep veins, cortical, and cortical eponymic veins. Inferior group of dural venous sinuses, inferior sagittal sinus, and cortical eponymic veins were poorly demonstrated. Score of the superior sagittal sinus, the straight sinus, the confluence of the superior sinus group, the right transverse and sigmoid sinuses, the internal veins, and the vein of Galen was excellent. The score of the left transverse and sigmoid sinuses was good. In conclusion, when using zero filling interpolation of the data in a three-dimensional phase contrast MR cerebral venography sequence, the superior group of dural venous sinuses and main major deep veins are demonstrated with good conspicuity.     

Proton magnetic resonance microimaging of human trabecular bone

Proton magnetic resonance (1H magnetic resonance imaging (MRI)) images of human trabecular bone were acquired and discussed for two samples with different porosity. Three-dimensional 3D Spin Echo (3D SE) and Multi-Slice Multi-Echo (MSME) pulse sequences were examined. A very high slice resolution of (38 microm)2 was achieved (MSME). The intensity histograms were found useful for the characterization of the bone porosity. A spatial distribution of the spin-spin relaxation time T2 was monitored with the MSME pulse program. The work demonstrates the great potential of the proton MRI technique in the study of the trabecular bone morphology.     

Targeted-ROI imaging in electron paramagnetic resonance imaging

Electron paramagnetic resonance imaging (EPRI) is a technique that has been used for in vivo oxygen imaging of small animals. In continuous wave (CW) EPRI, the measurement can be interpreted as a sampled 4D Radon transform of the image function. The conventional filtered-backprojection (FBP) algorithm has been used widely for reconstructing images from full knowledge of the Radon transform acquired in CW EPRI. In practical applications of CW EPRI, one often is interested in information only in a region of interest (ROI) within the imaged subject. It is desirable to accurately reconstruct an ROI image only from partial knowledge of the Radon transform because acquisition of the partial data set can lead to considerable reduction of imaging time. The conventional FBP algorithm cannot, however, reconstruct accurate ROI images from partial knowledge of the Radon transform of even dimension. In this work, we describe two new algorithms, which are referred to as the backprojection filtration (BPF) and minimum-data filtered-backprojection (MDFBP) algorithms, for accurate ROI-image reconstruction from a partial Radon transform (or, truncated Radon transform) in CW EPRI. We have also performed numerical studies in the context of ROI-image reconstruction of a synthetic 2D image with density similar to that found in a small animal EPRI. This demonstrates both the inadequacy of the conventional FBP algorithm and the success of BPF and MDFBP algorithms in ROI reconstruction. The proposed ROI imaging approach promises a means to substantially reduce image acquisition time in CW EPRI.