Magnetic Resonance Materials in Physics, Biology and Medicine最新文献

Objective: In the kidney, the medulla is most susceptible to damage in case of hampered perfusion or oxygenation. Due to separate regulation of cortical and medullary perfusion, measurement of both is crucial to improve the understanding of renal pathophysiology. We aim to develop and evaluate a physiologically accurate model to measure renal inner medullary (F_med) and cortical perfusion (F_cor) separately.

Materials and methods: We developed a 7-compartment model of renal perfusion and used an iterated approach to fit 10 free parameters. Model stability and accuracy were tested on both patient data and simulations. Cortical perfusion and F_T (tubular flow or glomerular filtration rate per unit of tissue volume) were compared to a conventional 2-compartment filtration model.

Results: Average (standard deviation) F_med was 37(23)mL/100 mL/min. Fitting stability as expressed by the median (interquartile range) coefficient of variation between fits was 0.0(0.0-5.8)%, with outliers up to 81%. In simulations, F_med was underestimated by around 8%. Intra-class correlation coefficients for F_cor and F_T as measured with the 2- and 7- compartment model were 0.87 and 0.63, respectively.

Discussion: We developed a pharmacokinetic model closely following renal physiology. Although the results were vulnerable for overfitting, relatively stable results could be obtained even for F_med.

Objective: To differentiate cribriform (GP4Crib+) from non-cribriform growth and Gleason 3 patterns (GP4Crib-/GP3) using MRI.

Methods: Two hundred and ninety-one operated prostate cancer men with pre-treatment MRI and whole-mount prostate histology were retrospectively included. T2-weighted, apparent diffusion coefficient (ADC) and fractional blood volume maps from 1.5/3T MRI systems were used. 592 histological GP3, GP4Crib- and GP4Crib+ regions were segmented on whole-mount specimens and manually co-registered to MRI sequences/maps. Radiomics features were extracted, and an erosion process was applied to minimize the impact of delineation uncertainties. A logistic regression model was developed to differentiate GP4Crib+ from GP3/GP4Crib- in the 465 remaining regions. The differences in balanced accuracy between the model and baseline (where all regions are labeled as GP3/GP4Crib-) and 95% confidence intervals (CI) for all metrics were assessed using bootstrapping.

Results: The logistic regression model, using the 90th percentile ADC feature with a negative coefficient, showed a balanced accuracy of 0.65 (95% CI: 0.48-0.79), receiver operating characteristic area under the curve (AUC) of 0.75 (95% CI: 0.54-0.92), a precision-recall AUC of 0.35 (95% CI: 0.14-0.68).

Conclusion: The radiomics MRI-based model, trained on Gleason sub-patterns segmented on whole-mount specimen, was able to differentiate GP4Crib+ from GP3/GP4Crib- patterns with moderate accuracy. The most dominant feature was the 90th percentile ADC. This exploratory study highlights 90th percentile ADC as a potential biomarker for cribriform growth differentiation, providing insights into future MRI-based risk assessment strategies.

Objective: To investigate whether synthetic DWI (SyDWI) calculated with short TR and zero TE can improve diffusion contrast in prostate compared to conventional DWI acquired with standard TR and TE.

Materials and methods: Thirty-two patients who underwent multiparametric MRI (mp-MRI) on a 3.0 T scanner were enrolled. For SyDWI, DWIs at b0 were acquired with two different TRs and TEs in addition to b1000 and b2000 images acquired with single conventional TR and TE. Contrast ratio (CR) was compared between SyDWI calculated with TR of 1000 ms and TE of 0 ms and conventional DWI acquired with TR of 6000 ms and TE of 70 ms.

Results: The mean CR between prostate cancer (PCa) and normal prostate, and between PCa and benign prostatic hyperplasia (BPH), is significantly higher in SyDWI compared to conventional DWI for both b-values of 1000 and 2000 s/mm². In addition, contrast within some lesions is now visualized, suggesting that tumour heterogeneity can be observed that is not seen with conventional DWI.

Conclusion: SyDWI calculated with TR of 1000 ms and TE of 0 ms significantly improves diffusion contrast between PCa and normal prostate or BPH, and within the lesion, compared to conventional DWI as a result of T1 shine-through.

Objective: Current voxel-based morphometry (VBM) studies of chemoradiotherapy effects on healthy tissues of the glioblastoma multiforme (GBM) brain faces challenges with neuroanatomical distortions (tumour, tumour oedema and resection cavities). Our aim is to compare current semi-automated segmentation methods and evaluate their reliability in investigating the effects of chemoradiotherapy on the tumoural and non-tumoural brain.

Materials and methods: VBM pipelines of CAT12 and FSL were performed and compared on an open-sourced imaging brain tumour dataset (Burdenko's Glioblastoma Progression dataset). The pre- and post-chemoradiotherapy grey (GM) and white matter (WM), and cerebrospinal fluid (CSF) volumes of whole brain, tumour-containing and nontumour containing hemispheres were derived and compared between both pipelines. Brain volumetric agreement and consistency between FSL and CAT12 were assessed using Bland-Altman plots and Intraclass correlation coefficient (ICC).

Results: Post-chemoradiotherapy GM volumes were significantly reduced in whole brain, tumour-containing and nontumour-containing hemispheres with a compensatory significant increase in CSF volumes, while WM volumes had no significant changes. Visual inspection revealed misclassification of tissue classes in the presence of neuroanatomical distortion for both pipelines, but FSL demonstrated superiority in tissue classification in the presence of a haematoma. Consequentially, Bland-Altman plots and ICC showed better agreement and consistency between FSL and CAT12 in GM_ICC (0.70 (0.53-0.82); p < 0.001), WM_ICC (0.75 (0.60-0.85); p < 0.001) and CSF_ICC (0.55 (0.32-0.71); p < 0.001) volumes of nontumour-containing hemispheres than whole brain and tumour-containing hemispheres.

Conclusions: VBM studies of chemoradiotherapy effects on the brain post-tumour resection remain challenging due to neuroanatomical distortions. A reliable alternative is to use nontumour-containing hemispheres with no anatomical distortion.

Objective: Diffusion-tensor imaging (DTI) and diffusion-weighted imaging (DWI) along perivascular spaces (ALPS) index have been proposed as noninvasive techniques to indirectly evaluate the glymphatic system function. However, these techniques are sensitive to examination parameters, limiting inter-study comparability. The definition of the region of interest (ROI) has been identified as the primary weakness of the ALPS method. Therefore, we aimed to determine which ROI characteristics would best promote consistent ALPS index analysis.

Methods: We examined 13 healthy volunteers using DTI and DWI to calculate the ALPS index, and compared and determined correlations among 11 different ROI configurations, and tested inter-method reliability.

Results: We found significant differences between different ROI configurations in the ALPS index calculation. Considering ROI characteristics and inter-method reliability, a squared ROI is the most suitable. The ICC between ROI configurations showed good-to-excellent inter-method agreement (mean ICC = 0.83). We did not find significant inter-method differences.

Conclusion: It is important to standardize the ROI characteristics for consistent ALPS index calculation.

Objective: This study presents the 3D MP-b-nSSFP sequence for multi-parametric mapping.

Methods: We evaluate several aspects of the 3D implementation, like the type of RF pulse (selective/non-selective), the readout duration, the undersampling pattern, and the acceleration factor. We use undersampled scans with subspace-constrained reconstruction and extended spiral readouts to achieve clinically acceptable scan times. The repeatability and accuracy of the $T_1$ and $T_2$ maps are compared with a reference technique in phantom and three volunteer scans.

Results: Compared with selective refocusing pulses, we observe lower bias with non-selective pulses, despite modeling the spatially varying effect of the pulses in the fitting process. $T_1$ and $T_2$ maps obtained from phantom scans were comparable to the nominal values and those from reference scans. $T_1$ values in vivo were underestimated compared to the reference scan. The maps with an acquisition matrix of $256\times 256\times 44$ and resolution $1\times 1\times 3$ ${\text{mm}}^3$ were acquired in 11 min.

Conclusion: We show that 3D MP-b-nSSFP can be used for multi-parameter mapping within clinically acceptable scan time. Phantom scans show results in good agreement with reference scan results. However, the in vivo scan underestimated $T_1$ .

Objective: The lung magnetic resonance longitudinal relaxation rate $R_1$ varies with disease but reported values in healthy subjects (HS) also differ markedly between studies.

Objective: To evaluate the reported values and variances of observed $R_1$ accounting for field strength $B_0$ , acquisition method, and disease.

Materials and methods: A systematic literature search was performed to identify studies quantifying $R_1$ or its reciprocal $T_1$ . The relationship between field strength and observed R₁ was analysed in all healthy subject data. Data were fit to the heuristic equation $R_1=B\ast {(B}_0^A)$ . The variances between-study $\mathrm{bs}$ , within-study-between-subject $\mathrm{wsbs}$ , and within-study-within-subject $\mathrm{wsws}$ , in HS were assessed. Linear correlation between observed R₁ and TE was also investigated.

Results: Fifty-nine papers were selected, 50 quantifying R₁ in HS and 20 in disease, representing 1559 human and non-human subjects, with $B_0$ ranging from 0.2 T to 9.4 T. In HS, the result of the fit on all healthy subjects was A = -0.227 ± 0.020 s^-1 T^-1, B = 0.923 ± 0.014 s^-1 with variance components $bs>wsbs>wsws$ . The reference values for R₁ at $B_0$ between 0.55 T and 7 T were derived. For inversion recovery with gradient echo readout at 1.5 T, a previously observed negative relationship between $R_1$ and TE was confirmed.

Discussion: We provide reference values for lung $R_1$ across all commonly used field strengths. The variation in HS lung observed $R_1$ reported in different studies in different centres likely reflects methodological differences. Investigators wishing to compare lung $R_1$ values with previous reports should take account of this irreproducibility, and multicentre projects should standardise to minimise between-centre variance.

Objective: To propose and validate an unsupervised MRI reconstruction method that does not require fully sampled k-space data.

Materials and methods: The proposed method, deep image prior with structured sparsity (DISCUS), extends the deep image prior (DIP) by introducing group sparsity to frame-specific code vectors, enabling the discovery of a low-dimensional manifold for capturing temporal variations. DISCUS was validated using four studies: (I) simulation of a dynamic Shepp-Logan phantom to demonstrate its manifold discovery capabilities, (II) comparison with compressed sensing and DIP-based methods using simulated single-shot late gadolinium enhancement (LGE) image series from six distinct digital cardiac phantoms in terms of normalized mean square error (NMSE) and structural similarity index measure (SSIM), (III) evaluation on retrospectively undersampled single-shot LGE data from eight patients, and (IV) evaluation on prospectively undersampled single-shot LGE data from eight patients, assessed via blind scoring from two expert readers.

Results: DISCUS outperformed competing methods, demonstrating superior reconstruction quality in terms of NMSE and SSIM (Studies I-III) and expert reader scoring (Study IV).

Discussion: An unsupervised image reconstruction method is presented and validated on simulated and measured data. These developments can benefit applications where acquiring fully sampled data is challenging.

Objective: This study aimed to develop an explainable radiomics model for stratifying prostate cancer (PCa) patients with high-risk disease via investigation of the association between cell density (CD) in the PCa region on histopathological images and multiparametric MR (mpMR) radiomics features.

Materials and methods: 137,970 radiomic features were calculated from mpMR images (101 PCa regions of 44 patients), and joint histograms (JHs) were derived from dynamic contrast-enhanced (DCE) images for each PCa region. The association between CD on histopathological images and its corresponding mpMR radiomic features in PCa regions for various grade groups and the three risk groups was evaluated using Spearman's correlation coefficient. To validate the potential of the radiomic-feature-CD association, we developed the radiomics model for stratifying patients into low/intermediate-risk and high-risk groups.

Results: There were moderate correlations of the CD with a DCE-based texture feature (WV_HH_1st_GLSZM_ZP) (ρ = 0.609, p = 0.024) and DCE-JH feature (JH_WV_HL_1st versus 5th-1st_Hist_STD) (ρ = 0.609, p = 0.024) in the high-risk group. The radiomics model had an accuracy of 0.920 for stratifying the patients of a test dataset into the low/intermediate-risk and high-risk groups.

Conclusion: The association between CD and mpMR features can be leveraged to develop the explainable radiomics for the high-risk stratification of patients with PCa.

Objective: This study evaluated the applicability and performance of the SDR4MR method at 1.5 T and 3 T across different acquisition scenarios in a clinical environment.

Materials and methods: The SDR4MR hardware consists of a broadband receiver coil connected to a software-defined radio (SDR) via optional RF attenuators. The SDR stick is plugged into the computer's USB port, which runs the SDR software and a Mathematica script to decode the RF pulse sequence. Several MRI pulse sequences were recorded: (i) a multi-echo multi-slice spin echo sequence to check the SDR4MR configuration on a well-known simple sequence; (ii) 2D and 3D sequences for which detailed information is not available in the user interface.

Results: The measured RF pulse sequences have been drawn in the style of illustrations found in MRI textbooks. Sequence times and amplitudes were estimated, and sequence details not described in the MRI user interface were retrieved.

Conclusion: The present study demonstrated the implementation of SDR4MR on clinical scanners. This easy-to-use configuration enables precise monitoring of RF pulse sequences. This method could be further improved by taking advantage of advances in SDR hardware and software.