Discriminative analysis of multivariate features from structural MRI and diffusion tensor images

Imaging markers derived from magnetic resonance images, together with machine learning techniques allow for the recognition of unique anatomical patterns and further differentiating Alzheimer's disease (AD) from normal states. T1-based imaging markers, especially volumetric patterns have demonstrated their discriminative potential, however, rely on the tissue abnormalities of gray matter alone. White matter abnormalities and their contribution to AD discrimination have been studied by measuring voxel-based intensities in diffusion tensor images (DTI); however, no systematic study has been done on the discriminative power of either region-of-interest (ROI)-based features from DTI or the combined features extracted from both T1 images and DTI. ROI-based analysis could potentially reduce the feature dimensionality of DTI indices, usually from more than 10e + 5, to 10–150 which is almost equal to the order of magnitude with respect to volumetric features from T1. Therefore it allows for straight forward combination of intensity based landmarks of DTI indices and volumetric features of T1. In the present study, the feasibility of tract-based features related to Alzheimer's disease was first evaluated by measuring its discriminative capability using support vector machine on fractional anisotropy (FA) maps collected from 21 subjects with Alzheimer's disease and 15 normal controls. Then the performance of the tract-based FA + gray matter volumes-combined feature was evaluated by cross-validation. The combined feature yielded good classification result with 94.3% accuracy, 95.0% sensitivity, 93.3% specificity, and 0.96 area under the receiver operating characteristic curve. The tract-based FA and the tract-based FA + gray matter volumes-combined features are certified their feasibilities for the recognition of anatomical features and may serve to complement classification methods based on other imaging markers.     

Microstructural characterization of materials by neural network technique

Ultrasonic signals received by pulse echo technique from plane parallel Zircaloy 2 samples of fixed thickness and of three different microstructures, were subjected to signal analysis, as conventional parameters like velocity and attenuation could not reliably discriminate them. The signals, obtained from these samples, were first sampled and digitized. Modified Karhunen Loeve Transform was used to reduce their dimensionality. A multilayered feed forward Artificial Neural Network was trained using a few signals in their reduced domain from the three different microstructures. The rest of the signals from the three samples with different microstructures were classified satisfactorily using this network.     

Unsupervised MRI segmentation of brain tissues using a local linear model and level set

Real-world magnetic resonance imaging of the brain is affected by intensity nonuniformity (INU) phenomena which makes it difficult to fully automate the segmentation process. This difficult task is accomplished in this work by using a new method with two original features: (1) each brain tissue class is locally modeled using a local linear region representative, which allows us to account for the INU in an implicit way and to more accurately position the region's boundaries; and (2) the region models are embedded in the level set framework, so that the spatial coherence of the segmentation can be controlled in a natural way. Our new method has been tested on the ground-truthed Internet Brain Segmentation Repository (IBSR) database and gave promising results, with Tanimoto indexes ranging from 0.61 to 0.79 for the classification of the white matter and from 0.72 to 0.84 for the gray matter. To our knowledge, this is the first time a region-based level set model has been used to perform the segmentation of real-world MRI brain scans with convincing results.     

Microstructural assessment of InN-on-GaN films grown by plasma-assisted MBE

The structural properties of InN thin films, grown by rf plasma-assisted molecular beam epitaxy on Ga-face GaN/Al₂O₃(0001) substrates, were investigated by means of conventional and high resolution electron microscopy. Our observations showed that a uniform InN film of total thickness up to 1 μm could be readily grown on GaN without any indication of columnar growth. A clear epitaxial orientation relationship of , was determined. The quality of the InN film was rather good, having threading dislocations as the dominant structural defect with a density in the range of 10⁹–10¹⁰ cm⁻². The crystal lattice parameters of wurtzite InN were estimated by electron diffraction analysis to be a=0.354 nm and c=0.569 nm, using Al₂O₃ as the reference crystal. Heteroepitaxial growth of InN on GaN was accomplished by the introduction of a network of three regularly spaced misfit dislocation arrays at the atomically flat interface plane. The experimentally measured distance of misfit dislocations was 2.72 nm. This is in good agreement with the theoretical value derived from the in-plane lattice mismatch of InN and GaN, which indicated that nearly full relaxation of the interfacial strain between the two crystal lattices was achieved.     

Radiative and nonradiative recombination processes in hydrogenated amorphous silicon as elucidated by optically detected magnetic resonance

Radiative and nonradiative recombination processes have been investigated by measurements of optically detected magnetic resonance at 2 K in hydrogenated anorphous silicon. Relevant processes are discussed on the basis of the experimental results.     

The surface structure of catalysts activated with hydrogen donors as elucidated by multinuclear solid-state NMR

¹H, ²⁷Al and ³¹P MAS, and ¹³C and ²⁹Si CP/MAS NMR spectroscopies, were used to characterize catalysts of Pd supported on various solids including SiO₂, AlPO₄ and Mg₃(PO₄)₂ that were activated with the chiral hydrogen-donor limonene. The above-mentioned techniques were used to check for the formation of an organopalladium complex between Pd²⁺ atoms and the olefin bonds in the limonene molecule on the catalyst surface. The results are compared with those obtained for catalysts activated in a hydrogen stream.     

Dynamic contrast-enhanced MRI and MR diffusion imaging to distinguish between glandular and stromal prostatic tissues

Purpose

To compare peak enhancement (PE), determined from dynamic contrast-enhanced (DCE) magnetic resonance imaging (MRI) and the magnetic resonance (MR) directionally-averaged apparent diffusion coefficient () in glandular versus stromal prostatic tissues and, with this comparison, to infer if the hypothesis that gadolinium-DTPA (Gd-DTPA) does not enter healthy glands or ducts is plausible.

Materials and Methods

MRI, MR spectroscopic imaging, DCE MRI and MR diffusion were evaluated in 17 untreated subjects with suspected or proven prostate cancer. PE and were compared in glandular-ductal tissues [normal peripheral zone and glandular benign prostatic hyperplasia (BPH)] and stromal-low ductal tissues (central gland/mixed BPH and stromal BPH).

Results

The glandular-ductal tissues had lower PE [125±6.4 (% baseline)] and higher [1.57±0.15 (s/10⁻³ mm²)] than the stromal-low ductal tissues [PE=132±5.5 (% baseline) (P<.0008), =1.18±0.20 (s/10⁻³ mm²) (P<1×10⁻⁸)]. A statistical model based upon stepwise regression was generated and completely separated the tissue types: ductal Measure = 448+669× (s/10⁻³ mm²)−10.7×PE (1/%), R²=1.0 and P<8×10⁻¹⁰.

Conclusions

The very different MR results in the glandular-ductal versus stromal-low ductal tissues suggest that these tissues have different underlying structure. These results support the hypothesis that Gd-DTPA does not enter healthy prostatic glands or ducts. This may explain the higher PE and lower that previously have been reported in prostate cancer versus healthy tissue.     

Polarized spectral features of human breast tissues through wavelet transform and principal component analysis

Fluorescence characteristics of human breast tissues are investigated through wavelet transform and principal component analysis (PCA). Wavelet transform of polarized fluorescence spectra of human breast tissues is found to localize spectral features that can reliably differentiate different tissue types. The emission range in the visible wavelength regime of 500–700 nm is analysed, with the excitation wavelength at 488 nm using laser as an excitation source, where flavin and porphyrin are some of the active fluorophores. A number of global and local parameters from principal component analysis of both high- and low-pass coefficients extracted in the wavelet domain, capturing spectral variations and subtle changes in the diseased tissues are clearly identifiable.     

Occult posttraumatic avascular necrosis of hip revealed by MRI.

Three hips with avascular necrosis in two patients had associated traumatic lesions noted on magnetic resonance imaging. Location of stress fracture in two femoral necks and acetabular contusion development prior to avascular necrosis favors these processes as being etiologic. These cases are particularly noteworthy in that nondisplaced fractures or other musculoskeletal trauma not visible on plain films has rarely been associated with subsequent development of avascular necrosis.     

The value of specific MRI features in the evaluation of suspected placental invasion

OBJECTIVE: The objective of this study was to determine imaging features that may help predict the presence of placenta accreta, placenta increta or placenta percreta on prenatal MRI scanning. SUBJECTS AND METHODS: A retrospective review of the prenatal MR scans of 10 patients with a diagnosis of placenta accreta, placenta increta or placenta percreta made by pathologic and clinical reports and of 10 patients without placental invasion was performed. Two expert MRI readers were blinded to the patients' true diagnosis and were asked to score a total of 17 MRI features of the placenta and adjacent structures. The interrater reliability was assessed using kappa statistics. The features with a moderate kappa statistic or better (kappa > .40) were then compared with the true diagnosis for each observer. RESULTS: Seven of the scored features had an interobserver reliability of kappa > .40: placenta previa (kappa = .83); abnormal uterine bulging (kappa = .48); intraplacental hemorrhage (kappa = .51); heterogeneity of signal intensity on T2-weighted (T2W) imaging (kappa = .61); the presence of dark intraplacental bands on T2W imaging (kappa = .53); increased placental thickness (kappa = .69); and visualization of the myometrium beneath the placenta on T2W imaging (kappa = .44). Using Fisher's two-sided exact test, there was a statistically significant difference between the proportion of patients with placental invasion and those without placental invasion for three of the features: abnormal uterine bulging (Rater 1, P = .005; Rater 2, P = .011); heterogeneity of T2W imaging signal intensity (Rater 1, P = .006; Rater 2, P = .010); and presence of dark intraplacental bands on T2W imaging (Rater 1, P = .003; Rater 2, P = .033). CONCLUSIONS: MRI can be a useful adjunct to ultrasound in diagnosing placenta accreta prenatally. Three features that are seen on MRI in patients with placental invasion appear to be useful for diagnosis: uterine bulging; heterogeneous signal intensity within the placenta; and the presence of dark intraplacental bands on T2W imaging.     

Primary retroperitoneal teratoma in the adult: correlation of MRI features with CT and pathology

Primary benign teratoma of the retroperitoneum is a rare tumor in the adult population. Only one case with an MRI examination has been reported in the English literature. This paper describes the CT and MRI features of a retroperitoneal teratoma in a 24-year-old male and discusses the value of MRI in the diagnosis and the preoperative imaging of such a tumor.     

The rate memory of a polymer close toT gas elucidated by reduced 4-D NMR echo experiments

We analyze the reorientational dynamics of a polymer close to the glass transition. Via comparison of two-time and four-time correlation functions, obtained from multidimensional nuclear magnetic resonance¹³C echo experiments, we are particularly sensitive to the fluctuations within the heterogeneous distribution of mobilities. From variation of the experimental parameters we obtain a detailed picture of the interplay between jump and diffusion dynamics, including the different types of motional heterogeneities and the geometry of the system.     

Enhancing the discrimination accuracy between metastases, gliomas and meningiomas on brain MRI by volumetric textural features and ensemble pattern recognition methods

Three-dimensional (3D) texture analysis of volumetric brain magnetic resonance (MR) images has been identified as an important indicator for discriminating among different brain pathologies. The purpose of this study was to evaluate the efficiency of 3D textural features using a pattern recognition system in the task of discriminating benign, malignant and metastatic brain tissues on T1 postcontrast MR imaging (MRI) series. The dataset consisted of 67 brain MRI series obtained from patients with verified and untreated intracranial tumors. The pattern recognition system was designed as an ensemble classification scheme employing a support vector machine classifier, specially modified in order to integrate the least squares features transformation logic in its kernel function. The latter, in conjunction with using 3D textural features, enabled boosting up the performance of the system in discriminating metastatic, malignant and benign brain tumors with 77.14%, 89.19% and 93.33% accuracy, respectively. The method was evaluated using an external cross-validation process; thus, results might be considered indicative of the generalization performance of the system to "unseen" cases. The proposed system might be used as an assisting tool for brain tumor characterization on volumetric MRI series.     

Microstructural evolution of ZnS during sintering monitored by optical and positron annihilation techniques

Positron lifetime and optical absorption techniques were employed to track the microstructural evolution of polycrystalline ZnS grown by Chemical Vapor Deposition (CVD). As grown material and material treated with Hot Isostatic Pressure (HIP) was sintered at temperatures ranging from 400 to 1000°C for 2–18 h. A 290 ps defect lifetime could be resolved in all samples, while an additional longer lifetime (=430 ps) was found only in samples annealed at low temperatures. This component gradually disappeared during annealing at 800°C. Associated with the disappearance of the long-lived component, the apparent bulk lifetime of the material changed from 235 to 215 ps. A 215±2 ps bulk parameter was also found for HIP-treated material annealed at temperatures greater than 400°C and hence is taken to represent the delocalized state of the positrons in ZnS. Optical absorption measurements showed that annealing at 800°C also caused the absorption profiles of the CVD and HIP samples to converge. The rate of the bulk lifetime transition correlates with the absorption changes. The observed sharpening of the absorption profile is attributed to a decrease in scattering from grain boundaries and voids, and a decrease in absorption from point defects. The 430 ps lifetime is believed to be due to trapping at voids and grain boundaries, while the 290 ps lifetime likely is due to a monovacancy stabilized as a small complex.Paper presented at the 132nd WE-Heraeus-Seminr on Positron Studies of Semiconductor Defects, Halle, Germany, 29 August to 2 September 1994     

Microstructural analysis of foam by use of NMR R2 dispersion

The spin–spin relaxation rate R₂ (=1/T₂) in hydrogel foams measured by use of a multiple spin echo sequence is found to be dependent on the echo time spacing. This property, referred to as R₂-dispersion, originates to a large extent from molecular self-diffusion of water within internal field gradients that result from magnetic susceptibility differences between the gel and air phase. Another contribution to the R₂ relaxation rate is surface relaxation. Numerical simulations are performed to investigate the relation between the foam microstructure (the mean air bubble radius and standard deviation of the air bubble radius) and foam composition properties (such as magnetic susceptibilities, diffusion coefficient and surface relaxivity) at one hand and the R₂-dispersion at the other hand. The simulated R₂-dispersions of gel foam are in agreement with the measured R₂-dispersions. By correlating the R₂-dispersion parameters and simulated microstructure properties a semi-empirical relationship is obtained that enables the mean air bubble size to be derived from measured R₂-dispersion curves. The R₂-derived mean air bubble size of a hydrogel foam is in agreement with the bubble size measured with X-ray micro-CT. This illustrates the feasibility of using ¹H R₂-dispersion measurements to determine the size of air bubbles in hydrogel foams and of alveoli in lung tissue.     

Microstructural and thermal characterization of neolithic ceramics

The aim of this research was to characterize some Neolithic ceramic finds collected during an excavation carried out at Lugo di Grezzana (Verona, Italy). Pottery shards with different paste and tempers were analyzed to better understand the manufacturing and firing technologies used for their production. Another task of the study was to determine whether highly refined artefacts, found in the site and resembling figulina-type ceramics, were of local production or imported from other places in the north of Italy, where the production of this sort of product has already been unambiguously assessed. Several results emerged from this investigation, providing indications on the finds from this Neolithic settlement and, therefore, on the technological expertise achieved by the primitive community. Moreover, a comparative study carried out on refined ceramic products found in Lugo and genuine figulina items from other northern Italian sites suggests that no trading exchange and commercial routes existed among those primitive communities.     

Microstructural and magnetic properties of CoPt nanowires

In this work, the magnetic and microstructural properties of CoPt nanowires are presented as a function of the electrolyte pH and current density during electrodeposition into anodized alumina templates. CoPt nanowires of high aspect ratio have been prepared using electrolyte pH values in the range from 2 to 6. The as-made samples exhibit a face centered cubic (fcc) structure with soft magnetic properties which transform into the face centered tetragonal (fct) L10 phase after thermal treatment. Different pH values of the electrolyte during electrodeposition lead to significantly different microstructures and, therefore, different magnetic properties. The CoPt nanowires prepared at high pH value are composed of fcc nanorods of about 25 nm in length. Thermal annealing of these samples leads to a preferred (0 0 1) orientation (along the direction perpendicular to the direction of nanowires) which increases with annealing time. On the other hand, the CoPt nanowires prepared at lower pH value are composed of uniform fcc nanograins with the size ∼2−3 nm. Magnetization curves for the later sample are virtually identical in both directions indicating an isotropic behavior.