Quantitative T1-relaxation corrected metabolite mapping of 12 metabolites in the human brain at 9.4 T

Magnetic resonance spectroscopic imaging (MRSI) is a non-invasive imaging modality that enables observation of metabolites. Applications of MRSI for neuroimaging applications has shown promise for monitoring and detecting various diseases. This study builds off previously developed techniques of short TR, 1H FID MRSI by correcting for T1-weighting of the metabolites and utilizing an internal water reference to produce quantitative (mmol kg-1) metabolite maps. This work reports and shows quantitative metabolite maps for 12 metabolites for a single slice. Voxel-specific T1-corrections for water are common in MRSI studies; however, most studies use either averaged T1-relaxation times to correct for T1-weighting of metabolites or omit this correction step entirely. This work employs the use of voxel-specific T1-corrections for metabolites in addition to water. Utilizing averaged T1-relaxation times for metabolites can bias metabolite maps for metabolites that have strong differences between T1-relaxation for GM and WM (i.e. Glu). This work systematically compares quantitative metabolite maps to single voxel quantitative results and qualitatively compares metabolite maps to previous works.

Computed tomography-based radiomics analysis to predict lymphovascular invasion in esophageal squamous cell carcinoma

Objective The present study explored the value of preoperative CT radiomics in predicting lymphovascular invasion (LVI) in esophageal squamous cell carcinoma (ESCC). Methods A retrospective analysis of 294 pathologically confirmed ESCC patients undergoing surgical resection and their preoperative chest-enhanced CT arterial images were used to delineate the target area of the lesion. All patients were randomly divided into a training cohort and a validation cohort at a ratio of 7:3. Radiomics features were extracted from single-slice, three-slice, and full-volume regions of interest (ROIs). The least absolute shrinkage and selection operator (LASSO) regression method was applied to select valuable radiomics features. Radiomics models were constructed using logistic regression method and were validated using leave group out cross-validation (LGOCV) method. The performance of the three models was evaluated using the receiver characteristic curve (ROC) and decision curve analysis (DCA). Results A total of 1218 radiomics features were separately extracted from single-slice ROIs, three-slice ROIs, and full-volume ROIs, and 16, 13 and 18 features, respectively, were retained after optimization and screening to construct a radiomics prediction model. The results showed that the AUC of the full-volume model was higher than that of the single-slice and three-slice models. According to LGOCV, the full-volume model showed the highest mean AUC for the training cohort and the validation cohort. Conclusion The full-volume radiomics model has the best predictive performance and thus can be used as an auxiliary method for clinical treatment decision making. Advances in knowledge: LVI is considered to be an important initial step for tumor dissemination. CT radiomics features correlate with LVI in ESCC and can be used as potential biomarkers for predicting LVI in ESCC.

Study of large axial space rebinning algorithm in γ-photon industrial tomography image reconstruction

The accuracy of the existing single slice and Fourier rebinning algorithms depends on the projection angle of the line of response. The increase of such projection angle with the detector size, typical in the large axial space of γ-photon industrial detection, and the loss of some projection data after rebinning, result in the degradation of the image quality. In addition, those algorithms consider the probability of positron annihilation equally distributed along the line of response, which prevents to estimate accurately the positions of the annihilation point, and can originate artifacts and noise in the reconstructed image. In this work, we propose an alternative large axial space rebinning algorithm. In that algorithm, initially the line of response is divided into transverse and axial components. Then, each line of response is uniformly rebinned into all the 2D sinogram data intersecting with it. To improve the accuracy of the estimate of the annihilation point location and suppress the noise effectively, we assign a Gaussian weight coefficient to the projection data, and optimise the rebinning algorithm with it. Finally, we reconstruct the image on the basis of the 2D sinograms with the optimised weights. On the computational side, the algorithm is also accelerated by making use of parallel computing. Both simulation and experimental results show that the proposed method improves the contrast and spatial resolution of 2D reconstructed images. Furthermore, the reconstruction time is not affected by the new method, which is therefore expected to meet the demand of γ-photon industrial inspection imaging.

Image Analysis to Optimize Calculations of Grinding Modes

Modeling helps to investigate and analyze the interrelationships of grinding parameters as a single system. The article describes a computer model of the surface layer of a grinding wheel, taking into account the cutting modes and characteristics of the circle. The subsystem of the surface layer of the grinding wheel is the most important subsystem of grinding, since it connects the main characteristics of the circle with the cutting modes and with the parameters of the workpiece. The output characteristics of the model are the parameters of the working layer with a height of several micrometers, in which micro-cutting occurs. It is impossible (or very difficult) to obtain them by conducting full-scale experiments, since the processes are instantaneous, and the cutting elements have micro-dimensions. This problem was solved by creating a simulation stochastic model based on the geometric representation of the surface layer, which clearly displays the result. The analysis of the image of this model allowed us to numerically describe the output parameters of cutting. The article offers a faster algorithm for analyzing the image of the simulation model of the surface layer. It is carried out over a matrix containing numerical information about the projection of the surface layer, the main parameters of each single slice are calculated, and only after that the result is displayed on the graphic screen. To simulate a single grinding mode, it is necessary to repeat the process of "image creation - image analysis– output of results" hundreds of times until a stable state is reached. The use of the algorithm in an automated system will allow you to create a system for automatically searching for optimal grinding modes, as well as to derive analytical dependencies of cutting modes on input parameters, for example, on the parameters of the circle and the workpiece.

Sources of variation in multicenter rectal MRI data and their effect on radiomics feature reproducibility

Objectives To investigate sources of variation in a multicenter rectal cancer MRI dataset focusing on hardware and image acquisition, segmentation methodology, and radiomics feature extraction software. Methods T2W and DWI/ADC MRIs from 649 rectal cancer patients were retrospectively acquired in 9 centers. Fifty-two imaging features (14 first-order/6 shape/32 higher-order) were extracted from each scan using whole-volume (expert/non-expert) and single-slice segmentations using two different software packages (PyRadiomics/CapTk). Influence of hardware, acquisition, and patient-intrinsic factors (age/gender/cTN-stage) on ADC was assessed using linear regression. Feature reproducibility was assessed between segmentation methods and software packages using the intraclass correlation coefficient. Results Image features differed significantly (p < 0.001) between centers with more substantial variations in ADC compared to T2W-MRI. In total, 64.3% of the variation in mean ADC was explained by differences in hardware and acquisition, compared to 0.4% by patient-intrinsic factors. Feature reproducibility between expert and non-expert segmentations was good to excellent (median ICC 0.89–0.90). Reproducibility for single-slice versus whole-volume segmentations was substantially poorer (median ICC 0.40–0.58). Between software packages, reproducibility was good to excellent (median ICC 0.99) for most features (first-order/shape/GLCM/GLRLM) but poor for higher-order (GLSZM/NGTDM) features (median ICC 0.00–0.41). Conclusions Significant variations are present in multicenter MRI data, particularly related to differences in hardware and acquisition, which will likely negatively influence subsequent analysis if not corrected for. Segmentation variations had a minor impact when using whole volume segmentations. Between software packages, higher-order features were less reproducible and caution is warranted when implementing these in prediction models. Key Points • Features derived from T2W-MRI and in particular ADC differ significantly between centers when performing multicenter data analysis. • Variations in ADC are mainly (> 60%) caused by hardware and image acquisition differences and less so (< 1%) by patient- or tumor-intrinsic variations. • Features derived using different image segmentations (expert/non-expert) were reproducible, provided that whole-volume segmentations were used. When using different feature extraction software packages with similar settings, higher-order features were less reproducible.

Automated differentiation of Alzheimer’s condition using Kernel Density Estimation based texture analysis of single slice brain MR images

Computer-assisted tools can aid in the detection of Alzheimer disease (AD) which is a progressive neurodegenerative disorder that can lead to cognitive impairments and eventually death. The accumulated effects due to AD can cause changes in the appearance of grey matter, white matter and cerebrospinal fluid in brain Magnetic Resonance (MR) images. This study aims to use Kernel Density Estimation (KDE) technique to analyse the textural changes from single slice brain MR images for the detection of AD. The preprocessed, skull stripped T1-weighted MR brain images are obtained from the publicly available OASIS database. A single axial slice per subject is chosen from a volumetric image for further processing to reduce the computational load. Multivariate KDE technique is applied to each pixel, by considering the changes in the neighbourhood based on selected bandwidth to obtain corresponding density estimates. Statistical features quantifying the distribution of density estimates are extracted to characterise textural variations in images. Linear discriminant analysis (LDA) classifier is implemented with ten-fold cross-validation for detecting AD. An optimum bandwidth of 18 for the KDE technique is selected based on the classification performance. Out of seven extracted texture features, three are found to be statistically significant in distinguishing AD. The classification with LDA yields an accuracy of 72.3% with a sensitivity of 80.6% for identifying AD from healthy subjects. The proposed method is efficient in detecting AD by revealing the textural changes within the brain slice without the involvement of any segmentation technique. Thus, the novel KDE-based texture analysis proves to be an effective tool for the automated diagnosis of AD from single slice brain MR images.

A Novel Ultra-Compact FPGA PUF: The DD-PUF

In this paper, we present a novel ultra-compact Physical Unclonable Function (PUF) architecture and its FPGA implementation. The proposed Delay Difference PUF (DD-PUF) is the most dense FPGA-compatible PUF ever reported in the literature, allowing the implementation of two PUF bits in a single slice and provides very good values for all the most important figures of merit. The architecture of the proposed PUF exploits the delay difference between two nominally identical signal paths and the metastability features of D-Latches with an asynchronous reset input. The DD-PUF has been implemented on both Xilinx Spartan-6 and Artix-7 devices and the resulting design flows which allow to accurately balance the nominal delay of the different signal paths is outlined. The circuits have been extensively tested under temperature and supply voltage variations and the results of our evaluations on both FPGA families have shown that the proposed architecture and implementation are able to fit in just 32 Configurable Logic Blocks (CLBs) without sacrificing steadiness, uniqueness and uniformity, thus outperforming most of the previously published FPGA-compatible PUFs.

Diffusion-weighted MR imaging in chronic non-bacterial osteitis: Proof-of-concept of the apparent diffusion coefficient as an outcome measure

Background The apparent diffusion coefficient (ADC), as determined by whole-body diffusion-weighted MRI, may be useful as an outcome measure for monitoring response to treatment in chronic non-bacterial osteitis. Purpose To test and demonstrate the feasibility of ADC-measurement methods for use as outcome measure in chronic non-bacterial osteitis. Materials and Methods Using data from a randomized pilot study, feasibility of change-score ADC between baseline and second MRI (ΔADC12) and third MRI (ΔADC13) as outcome measure was assessed in three settings: “whole-lesion,” “single-slice per lesion,” and “index-lesion per patient”. Bone marrow edema lesions were depicted on short tau inversion recovery sequence at baseline and copied to ADC maps at the three time-points. Correlations between the three settings were measured as were analysis of variances. Discriminant validity was assessed as inter- and intra-observer reproducibility and smallest detectable change. Results 12 subjects were enrolled, and MRI was performed at baseline and weeks 12 and 36. Pearson correlation was high ( r > 0.86; p ≤ 0.01) for ΔADC between single-slice—whole-lesion and whole-lesion—index-lesion and tended to be significant for single-slice—index-lesion settings ( p = 0.06). For ΔADC12 and ΔADC13, Bland–Altman plots showed small differences (0.02, 0.03) and narrow 95% limits-of-agreement (−0.13–0.09, −0.07–0.05 μm2/s) between whole-lesion and single-slice ROI settings. Inter-observer reproducibility measured by intra-class correlation coefficient was poor-to-fair (range: 0.09–0.31), whereas intra-observer reproducibility was good-to-excellent (range: 0.67–0.90). Smallest detectable changes were between 0.21–0.28 μm2/s. Conclusion ADC change-score as outcome measure was feasible, and the single-slice per lesion ROI setting performed almost equally to whole-lesion setting resulting in reduced assessment time.

Whole-Lesion Apparent Diffusion Coefficient Histogram Analysis: Significance for Discriminating Lung Cancer from Pulmonary Abscess and Mycobacterial Infection

Diffusion-weighted magnetic resonance imaging (DWI) can differentiate malignant from benign pulmonary nodules. However, it is difficult to differentiate pulmonary abscesses and mycobacterial infections (PAMIs) from lung cancers because PAMIs show restricted diffusion in DWI. The study purpose is to establish the role of ADC histogram for differentiating lung cancer from PAMI. There were 41 lung cancers (25 adenocarcinomas, 16 squamous cell carcinomas), and 19 PAMIs (9 pulmonary abscesses, 10 mycobacterial infections). Parameters more than 60% of the area under the ROC curve (AUC) were ADC, maximal ADC, mean ADC, median ADC, most frequency ADC, kurtosis of ADC, and volume of lesion. There were significant differences between lung cancer and PAMI in ADC, mean ADC, median ADC, and most frequency ADC. The ADC (1.19 ± 0.29 × 10−3 mm2/s) of lung cancer obtained from a single slice was significantly lower than that (1.44 ± 0.54) of PAMI (p = 0.0262). In contrast, mean, median, or most frequency ADC of lung cancer which was obtained in the ADC histogram was significantly higher than the value of each parameter of PAMI. ADC histogram could discriminate PAMIs from lung cancers by showing that AUCs of several parameters were more than 60%, and that several parameters of ADC of PAMI were significantly lower than those of lung cancer. ADC histogram has the potential to be a valuable tool to differentiate PAMI from lung cancer.