Magnetic resonance imaging最新文献

{"title":"MRASM: A multiscale residual attention spatiotemporal model for breast tumor prediction","authors":"Liangliang Zhang,&nbsp;Hang Zheng,&nbsp;Jiawei Liu,&nbsp;Zhenzhen Wang,&nbsp;Xin Zheng,&nbsp;Qingfeng Tang,&nbsp;Xianyang Wang,&nbsp;Hui Liu","doi":"10.1016/j.mri.2025.110457","DOIUrl":"10.1016/j.mri.2025.110457","url":null,"abstract":"<div><h3>Background</h3><div>Spatial features and temporal features derived from dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) are both useful for the prediction of tumor information. However, it remains unclear whether deep spatiotemporal features can improve the diagnostic performance of the model.</div></div><div><h3>Purpose</h3><div>To improve the prediction performance of benign and malignant breast tumors by efficiently integrating deep spatial and temporal features from DCE-MRI.</div></div><div><h3>Method</h3><div>A multiscale residual attention spatiotemporal model (MRASM) based on ResNet (2 + 1)D was proposed for the prediction of benign and malignant breast tumors. Compared to ResNet (2 + 1)D, the MRASM has three improvements, including multiscale, attention, and residual modules. Model performance was evaluated using the area under the receiver operating characteristic (ROC) curve (AUC) and accuracy. Statistical comparisons between the MRASM and traditional methods were performed using the DeLong test.</div></div><div><h3>Results</h3><div>The MRASM method obtained a higher AUC of 0.9962 and accuracy of 0.9660 compared to original ResNet (2 + 1)D method (AUC = 0.9182 and accuracy = 0.8340) and traditional 3D CNN methods (best AUC = 0.8752 and accuracy = 0.8128).</div></div><div><h3>Conclusion</h3><div>The proposed MRASM method is an efficient approach for the accurate prediction of benign and malignant breast tumors.</div></div>","PeriodicalId":18165,"journal":{"name":"Magnetic resonance imaging","volume":"122 ","pages":"Article 110457"},"PeriodicalIF":2.1,"publicationDate":"2025-07-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144588019","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Segmental evaluation of myocardial ischemia in stable coronary artery disease using native T1 mapping","authors":"Runlan Luo ,&nbsp;Qi'’an Sun ,&nbsp;Yi Zhao ,&nbsp;Aihua Li ,&nbsp;Hang Qu ,&nbsp;Xiaoping Yu ,&nbsp;Wei Wang","doi":"10.1016/j.mri.2025.110456","DOIUrl":"10.1016/j.mri.2025.110456","url":null,"abstract":"<div><h3>Objective</h3><div>To evaluate the myocardial ischemic segments and related factors in stable coronary artery disease (SCAD) patients by native T1 mapping.</div></div><div><h3>Methods</h3><div>316 SCAD patients and 30 healthy controls (all right coronary dominant) underwent CMR native T1 mapping within 90 days of CCTA. Segmental native T1 values were measured using AHA 16-segment model. Patients were grouped by number of diseased coronary arteries (DCA [1–3]) with the largest diameter stenosis (DS [&lt; or ≥ 50 %]) and culprit coronary artery (CCA [LAD, LCX, RCA]), or number of coronary artery stenosis ≥50 % (CAS [0–3]), respectively. Ischemic segments were defined as native T1 values significantly elevated versus controls. Multivariable generalized estimating equations (GEE) model was used to identify independent factors.</div></div><div><h3>Results</h3><div>Single-vessel disease showed localized native T1 increases in corresponding perfusion territories, while multi-vessel disease exhibited complex ischemia patterns. Anterior and anteroseptal segments had significantly higher native T1 values in groups CAS 2 and 3 than CAS 0 and 1 (Bonferroni-adjusted <em>P</em> &lt; 0.05). GEE model identified DCA (two-vessel disease: β = 13.6 ms, <em>P</em> = 0.010), CAS (1–3: β = 10.5, 34.4 and 57.2 ms, <em>P</em> &lt; 0.05, respectively), and coronary artery calcium (CAC) score 3 and 4 (β = 12.2 and 14.5 ms, P &lt; 0.05), LAD-fractional flow reserves (LAD-FFR) (β = −47.3 ms, <em>P</em> = 0.010) and CCA (LCX: β = −9.5 ms, <em>P</em> = 0.019) as independent factors.</div></div><div><h3>Conclusion</h3><div>Native T1 mapping reveals spatially heterogenous ischemia in SCAD and is independently associated with both anatomical and functional parameters, supporting its value in personalized evaluation and management.</div></div>","PeriodicalId":18165,"journal":{"name":"Magnetic resonance imaging","volume":"122 ","pages":"Article 110456"},"PeriodicalIF":2.1,"publicationDate":"2025-07-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144600854","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Spatially resolved diffusion pore imaging using k-space readout","authors":"Lucas Oswald ,&nbsp;Julian Rauch ,&nbsp;Frederik B. Laun ,&nbsp;Mark E. Ladd ,&nbsp;Tristan A. Kuder","doi":"10.1016/j.mri.2025.110455","DOIUrl":"10.1016/j.mri.2025.110455","url":null,"abstract":"<div><div>Nuclear magnetic resonance diffusion methods are powerful tools for investigating the underlying structure of materials or tissues. Diffusion pore imaging (DPI) provides access to information about the geometric shape of pores containing diffusible substances. This technique yields an averaged image of the pores present in the imaging volume and enables measurements at a scale much smaller than that of conventional MR imaging. For applications in non-homogeneous materials such as biological tissues or heterogeneous porous media, the integration of a second spatial encoding step is essential to distinguish pore shapes in different regions of the measurement volume. Here, we present a combination of two-dimensional <em>q</em>-space and two-dimensional <em>k</em>-space acquisition on a Bruker 9.4 T small animal scanner. A 2D pore space function is reconstructed in each image voxel obtained from <em>k</em>-space. The long-narrow sequence scheme necessary for DPI was extended with a conventional <em>k</em>-space imaging readout to fill both <em>k</em>- and <em>q</em>-space. A conventional spin-echo approach with a single refocusing pulse was employed. From two different regions of interest, the sizes of capillaries with inner diameters of 15 μm and 20 μm, respectively, present in a phantom could be estimated from one- and two-dimensional projections of the pore space function. Simulations using the multiple correlation function approach exhibit good agreement with the measured one-dimensional pore space functions. Existing residual phases in the measurement data were corrected using phase reference measurements in a structureless oil phantom. In summary, spatially resolved pore imaging allows for the reconstruction of pore shapes in specific regions of interest, reinforcing the potential of DPI to non-invasively explore cellular structure. This study demonstrates the ability to reveal the voxel-averaged shape of pore distributions within a single DPI measurement on a preclinical MR scanner.</div></div>","PeriodicalId":18165,"journal":{"name":"Magnetic resonance imaging","volume":"122 ","pages":"Article 110455"},"PeriodicalIF":2.1,"publicationDate":"2025-07-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144600855","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Exploiting four-way phase-encoding benefits for robust detection and correction of EPI artifacts: Application to residual ghosts in -diffusion MRI","authors":"Anh Thai ,&nbsp;Lin Ching Chang ,&nbsp;Carlo Pierpaoli ,&nbsp;M. Okan Irfanoglu","doi":"10.1016/j.mri.2025.110454","DOIUrl":"10.1016/j.mri.2025.110454","url":null,"abstract":"<div><h3>Purpose</h3><div>To propose and develop an image processing-based methodology for detecting and correcting residual Nyquist ghost artifacts in echo planar imaging (EPI), specifically using non-diffusion-weighted (<span><math><mi>b</mi><mo>=</mo><mn>0</mn><mspace></mspace><mi>s</mi><mo>/</mo><mi>m</mi><msup><mi>m</mi><mn>2</mn></msup><mo>)</mo></math></span> images acquired with four distinct phase-encoding directions (PEDs) in diffusion MRI.</div></div><div><h3>Approach</h3><div>Previous studies have demonstrated that acquiring images with four different PEDs can improve the reproducibility of diffusion derived quantitative maps. This improvement is achieved by averaging across PEDs to reduce the impact of residual EPI Nyquist ghost artifacts. These residual artifacts originate from imperfect ghost correction applied during image reconstruction and often persist into the preprocessing pipeline, potentially biasing downstream analyses. Building on these observations, the proposed method further improves this 4-way encoding approach by leveraging the properties of signal distributions to detect artifactual regions, specifically in non-diffusion weighted (<span><math><mi>b</mi><mo>=</mo><mn>0</mn><mspace></mspace><mi>s</mi><mo>/</mo><mi>m</mi><msup><mi>m</mi><mn>2</mn></msup></math></span>). While the corrections are applied only to the <span><math><mi>b</mi><mo>=</mo><mn>0</mn></math></span> images, its effectiveness is evaluated through downstream diffusion tensor estimation, by assessing the improvements in diffusion-derived metrics such as FA and MD. Additionally, the method can be tailored for specific artifact manifestations by considering their localization due to underlying acquisition parameters.</div></div><div><h3>Results</h3><div>Simulations with known ground-truth images demonstrated high artifact detection accuracy, achieving a Dice score of 0.91 for reconstructions without parallel imaging. In the in-vivo dataset, the method also improved longitudinal reproducibility, reducing variability by 30 % in ghost-affected regions.</div></div><div><h3>Conclusion</h3><div>The proposed correction method effectively detected and corrected residual ghost artifacts without the need of any additional k-space data, specifically in non-diffusion weighted images. This retrospective approach can be directly integrated into existing processing pipelines to further improve the quality of EPI images and enhancing image quality in studies that utilize 4-way PEDs acquisition.</div></div>","PeriodicalId":18165,"journal":{"name":"Magnetic resonance imaging","volume":"122 ","pages":"Article 110454"},"PeriodicalIF":2.1,"publicationDate":"2025-07-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144600936","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Evaluating placental microstructure and microcirculation in predicting the progression of gestational hypertension to preeclampsia: a systematic comparison between virtual MR elastography, IVIM, ultrasound and lab indexes","authors":"Jing Deng ,&nbsp;Jiacheng Song ,&nbsp;Aining Zhang ,&nbsp;Feifei Qu ,&nbsp;Yanglei Wu ,&nbsp;Ting Chen","doi":"10.1016/j.mri.2025.110453","DOIUrl":"10.1016/j.mri.2025.110453","url":null,"abstract":"<div><h3>Objective</h3><div>We aimed to utilize functional magnetic resonance imaging (fMRI), including placental virtual magnetic resonance elastography (vMRE), intravoxel incoherent motion (IVIM) parameters, MRI morphological parameters, ultrasound and laboratory indexes to predict the progression of GH to PE and differentiate patients with GH and controls.</div></div><div><h3>Methods</h3><div>Between January 2018 and March 2023, we retrospectively collected 382 pregnant women and ultimately included 68 subjects: 24 with PE, 14 with GH, and 30 healthy controls. The stiffness value (<em>μ</em>_diff) from vMRE, true diffusion coefficient, pseudo-diffusion coefficient, perfusion fraction (<em>f</em>) from IVIM-based analysis, and apparent diffusion coefficient were calculated. MRI morphological parameters were evaluated using T2-weighted imaging. Doppler parameters of the umbilical artery (UA) and middle cerebral artery, cerebroplacental ratio (CPR), and laboratory tests were retrospectively collected. ANOVA analysis compared parameters across three groups. Predictive performance was evaluated using receiver operating characteristic curve analysis and area under the curve (AUC).</div></div><div><h3>Results</h3><div>Doppler parameters (pulsatility index, resistance index, peak systolic velocity/end-diastolic velocity in UA, and CPR) and fMRI parameters (<em>μ</em>_diff and <em>f</em>) effectively predicted PE in patients with GH. The fMRI combined model exhibited superior predictive efficacy compared to the Doppler model (AUC, 0.845 vs. 0.792). Their combined model further improved predictive efficacy (AUC, 0.920). The prothrombin time level, <em>μ</em>_diff and <em>f</em> could identify patients with GH and controls, with no substantial differences in all Doppler parameters.</div></div><div><h3>Conclusions</h3><div>vMRE and IVIM parameters might be superior to ultrasound and laboratory indexs in predicting the progression of GH to PE and differentiating patients with GH and controls.</div></div>","PeriodicalId":18165,"journal":{"name":"Magnetic resonance imaging","volume":"122 ","pages":"Article 110453"},"PeriodicalIF":2.1,"publicationDate":"2025-07-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144591614","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Noise decorrelation coil combination optimizes SNR of edited 1H MRS data","authors":"Amy E. Bouchard,&nbsp;Mark Mikkelsen","doi":"10.1016/j.mri.2025.110452","DOIUrl":"10.1016/j.mri.2025.110452","url":null,"abstract":"<div><h3>Introduction</h3><div>Determining the optimal radiofrequency (RF) coil combination method for magnetic resonance spectroscopy (MRS) is crucial for maximizing the signal-to-noise ratio (SNR) and reliably detecting low-concentration metabolites, such as γ-aminobutyric acid (GABA). We compared the performances of several previously proposed algorithms using GABA-edited ¹H MRS data. Given that phased-array coils often exhibit noise correlations that reduce SNR, we hypothesized that noise decorrelation algorithms would be most effective.</div></div><div><h3>Methods</h3><div>We examined six coil combination methods, with the second half accounting for noise correlations: 1) equal weighting; 2) signal weighting; 3) S/N² weighting; 4) noise-decorrelated combination (nd-comb); 5) whitened singular value decomposition (WSVD); and 6) generalized least squares (GLS). Each method was applied to 119 GABA-edited MEGA-PRESS datasets acquired on 3 T GE and Siemens MRI scanners across 11 research sites. We estimated the SNR of GABA+ and <em>N</em>-acetylaspartate (NAA) and tested for statistical differences between the six approaches. We also calculated the intersubject coefficients of variation (CVs) of GABA+/creatine (Cr) ratios.</div></div><div><h3>Results</h3><div>There were significant differences in the SNR of GABA+ and NAA between the methods. Noise decorrelation methods produced higher SNR compared to the other approaches, with nd-comb, WSVD, and GLS yielding, on average, approximately 37 % more GABA+ and 34 % more NAA SNR than equal weighting. GLS yielded the highest SNR for both GABA+ and NAA. The CVs for GABA+/Cr were generally somewhat smaller when using noise decorrelation.</div></div><div><h3>Conclusion</h3><div>As predicted, noise decorrelation coil combination, particularly GLS, produced optimal SNR for GABA-edited MRS data.</div></div>","PeriodicalId":18165,"journal":{"name":"Magnetic resonance imaging","volume":"122 ","pages":"Article 110452"},"PeriodicalIF":2.1,"publicationDate":"2025-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144557385","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Diffusion basis spectrum imaging detects axonal injury in the optic nerve following traumatic brain injury","authors":"Raj Swaroop Lavadi ,&nbsp;Kainen L. Utt ,&nbsp;Stephen N. Housley ,&nbsp;Sirisha Nouduri ,&nbsp;Tsen-Hsuan Lin ,&nbsp;Jacob Blum ,&nbsp;Brenton H. Pennicooke ,&nbsp;Sheng-Kwei Song ,&nbsp;Nitin Agarwal","doi":"10.1016/j.mri.2025.110451","DOIUrl":"10.1016/j.mri.2025.110451","url":null,"abstract":"<div><h3>Purpose</h3><div>To utilize diffusion basis spectrum imaging (DBSI), an advanced imaging modality that has been shown to distinguish between co-occurring white matter pathologies, to discern changes in the optic nerves among patients with traumatic brain injury (TBI).</div></div><div><h3>Methods</h3><div>Seven patients with TBI were prospectively recruited to undergo a 3 T magnetic resonance imaging brain scan within two months of injury, and follow-up scans at 6- and 12-months. The optic nerve was considered the region of interest. Manual alignment of the optic nerve slices was performed, followed by the deployment of an in-house script to obtain post-processed data. Diffusion tensor imaging and DBSI-derived axial (AD) and radial diffusivity (RD), reflecting axon and myelin integrity, respectively, were compared. DBSI-derived signal intensities were also compared. The Kruskal-Wallis test was performed to determine significance (<em>p</em> &lt; 0.05).</div></div><div><h3>Results</h3><div>Fourteen optic nerves from patients were compared with 18 optic nerves from control participants. The values of DTI RD were significantly greater among patients than controls (p &lt; 0.05) across all timepoints, with no corresponding differences in the AD. However, a decrease in DBSI AD (<em>p</em> &lt; 0.01) observed only at the initial scan, coupled with a stable RD, was observed among patients. Comparisons between DBSI signal intensities among patients showed no appreciable differences; however, the fiber fraction was significantly lower (<em>p</em> &lt; 0.05) at all timepoints.</div></div><div><h3>Conclusion</h3><div>Acute reductions in DBSI-derived AD and sustained reductions in fiber fraction can serve as a potential biomarker for axonal injury in the optic nerves of patients with TBI.</div></div>","PeriodicalId":18165,"journal":{"name":"Magnetic resonance imaging","volume":"122 ","pages":"Article 110451"},"PeriodicalIF":2.1,"publicationDate":"2025-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144560512","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Qualitative and quantitative analysis of functional cardiac MRI using a novel compressed SENSE sequence with artificial intelligence image reconstruction","authors":"Konstantin Klein ,&nbsp;Marcel Christian Langenbach ,&nbsp;Lenhard Pennig ,&nbsp;Thomas Schömig ,&nbsp;Robert Terzis ,&nbsp;Isabel Luisa Langenbach ,&nbsp;David Maintz ,&nbsp;Matej Gajzler ,&nbsp;Kristina Sonnabend ,&nbsp;Claas Philip Nähle","doi":"10.1016/j.mri.2025.110448","DOIUrl":"10.1016/j.mri.2025.110448","url":null,"abstract":"<div><h3>Background</h3><div>To evaluate the feasibility of combining Compressed SENSE (CS) with a newly developed deep learning-based algorithm (CS-AI) using a Convolutional Neural Network to accelerate balanced steady-state free precession (bSSFP)-sequences for cardiac magnetic resonance imaging (MRI).</div></div><div><h3>Methods</h3><div>30 healthy volunteers were examined prospectively with a 3 T MRI scanner. We acquired CINE bSSFP sequences for short axis (SA, multi-breath-hold) and four-chamber (4CH)-view of the heart. For each sequence, four different CS accelerations and CS-AI reconstructions with three different denoising parameters, CS-AI medium, CS-AI strong, and CS-AI complete, were used. Cardiac left ventricular (LV) function (i.e., ejection fraction, end-diastolic volume, end-systolic volume, and LV mass) was analyzed using the SA sequences in every CS factor and each AI level. Two readers, blinded to the acceleration and denoising levels, evaluated all sequences regarding image quality and artifacts using a 5-point Likert scale. Friedman and Dunn's multiple comparison tests were used for qualitative evaluation, ANOVA and Tukey Kramer test for quantitative metrics.</div></div><div><h3>Results</h3><div>Scan time could be decreased up to 57 % for the SA-Sequences and up to 56 % for the 4CH-Sequences compared to the clinically established sequences consisting of SA-CS3 and 4CH-CS2,5 (SA-CS3: 112 s vs. SA-CS6: 48 s; 4CH-CS2,5: 9 s vs. 4CH-CS5: 4 s, <em>p</em> &lt; 0.001). LV-functional analysis was not compromised by using accelerated MRI sequences combined with CS-AI reconstructions (all <em>p</em> &gt; 0.05). The image quality loss and artifact increase accompanying increasing acceleration levels could be entirely compensated by CS-AI post-processing, with the best results for image quality using the combination of the highest CS factor with strong AI (SA-CINE: Coef.:1.31, 95 %CI:1.05–1.58; 4CH-CINE: Coef.:1.18, 95 %CI:1.05–1.58; both <em>p</em> &lt; 0.001), and with complete AI regarding the artifact score (SA-CINE: Coef.:1.33, 95 %CI:1.06–1.60; 4CH-CINE: Coef.:1.31, 95 %CI:0.86–1.77; both p &lt; 0.001).</div></div><div><h3>Conclusion</h3><div>Combining CS sequences with AI-based image reconstruction for denoising significantly decreases scan time in cardiac imaging while upholding LV functional analysis accuracy and delivering stable outcomes for image quality and artifact reduction. This integration presents a promising advancement in cardiac MRI, promising improved efficiency without compromising diagnostic quality.</div></div>","PeriodicalId":18165,"journal":{"name":"Magnetic resonance imaging","volume":"122 ","pages":"Article 110448"},"PeriodicalIF":2.1,"publicationDate":"2025-06-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144340185","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Hierarchical refinement with adaptive deformation cascaded for multi-scale medical image registration","authors":"Naeem Hussain ,&nbsp;Zhiyue Yan ,&nbsp;Wenming Cao ,&nbsp;Muhammad Anwar","doi":"10.1016/j.mri.2025.110449","DOIUrl":"10.1016/j.mri.2025.110449","url":null,"abstract":"<div><div>Deformable image registration is a fundamental task in medical image analysis, which is crucial in enabling early detection and accurate disease diagnosis. Although transformer-based architectures have demonstrated strong potential through attention mechanisms, challenges remain in ineffective feature extraction and spatial alignment, particularly within hierarchical attention frameworks. To address these limitations, we propose a novel registration framework that integrates hierarchical feature encoding in the encoder and an adaptive cascaded refinement strategy in the decoder. The model employs hierarchical cross-attention between fixed and moving images at multiple scales, enabling more precise alignment and improved registration accuracy. The decoder incorporates the Adaptive Cascaded Module (ACM), facilitating progressive deformation field refinement across multiple resolution levels. This approach captures coarse global transformations and acceptable local variations, resulting in smooth and anatomically consistent alignment. However, rather than relying solely on the final decoder output, our framework leverages intermediate representations at each stage of the network, enhancing the robustness and precision of the registration process. Our method achieves superior accuracy and adaptability by integrating deformations across all scales.</div><div>Comprehensive experiments on two widely used 3D brain MRI datasets, OASIS and LPBA40, demonstrate that the proposed framework consistently outperforms existing state-of-the-art approaches across multiple evaluation metrics regarding accuracy, robustness, and generalizability.</div></div>","PeriodicalId":18165,"journal":{"name":"Magnetic resonance imaging","volume":"122 ","pages":"Article 110449"},"PeriodicalIF":2.1,"publicationDate":"2025-06-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144336634","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A human-compatible gradient coil for visualizing ultrasound propagation","authors":"Paul-Emile Passe-Carlus ,&nbsp;Davi Cavinatto ,&nbsp;Nathan Thyberg ,&nbsp;Carson Reed ,&nbsp;Taylor D. Webb ,&nbsp;Steven P. Allen","doi":"10.1016/j.mri.2025.110447","DOIUrl":"10.1016/j.mri.2025.110447","url":null,"abstract":"<div><div>In pursuit of a magnetic resonance imaging (MRI) based means to tailor transcranial focused ultrasound neuromodulation to a patient's unique skull morphology, this study presents a specialized gradient coil that sensitizes MRI images to ultrasonic vibrations at depths equivalent to the human cortex in a technique called the magnetic resonance hydrophone. The coil comprises a 60 mm diameter, pancake-style design that encodes acoustic displacements into MR images at the cost of an inhomogeneous encoding field. The coil was coupled with a 500 kHz, custom built, ultrasonic transducer. Both the magnetic field gradient of the coil and the acoustic field of the transducer were characterized in benchtop experiments. Acoustic standing waves were estimated in silico. Resulting MR images displayed a sinusoidal phase pattern modulated by both the transducer's acoustic field and the coil's magnetic field gradient. Acoustic pressures were estimated from the resulting images and compared to hydrophone measurements. The pancake-style coil produced a pressure measurement uncertainty pattern due to electronic noise that increased exponentially with depth. Uncertainty at locations between 0 and 30 mm of depth within a region approximately 10 mm wide scaled between approximately 20 kPa and 100 kPa. On average, the MRH underestimated the hydrophone by 12 kPa with the difference between the two following a standard deviation of 21 kPa.</div></div>","PeriodicalId":18165,"journal":{"name":"Magnetic resonance imaging","volume":"122 ","pages":"Article 110447"},"PeriodicalIF":2.1,"publicationDate":"2025-06-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144326119","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}