MR Imaging Research Articles

Classification of Muscular Dystrophies from MR Images Improves Using the Swin Transformer Deep Learning Model.

Muscular dystrophies present diagnostic challenges, requiring accurate classification for effective diagnosis and treatment. This study investigates the efficacy of deep learning methodologies in classifying these disorders using skeletal muscle MRI scans. Specifically, we assess the performance of the Swin Transformer (SwinT) architecture against traditional convolutional neural networks (CNNs) in distinguishing between healthy individuals, Becker muscular dystrophy (BMD), and limb-girdle muscular Dystrophy type 2 (LGMD2) patients. Moreover, 3T MRI scans from a retrospective dataset of 75 scans (from 54 subjects) were utilized, with multiparametric protocols capturing various MRI contrasts, including T1-weighted and Dixon sequences. The dataset included 17 scans from healthy volunteers, 27 from BMD patients, and 31 from LGMD2 patients. SwinT and CNNs were trained and validated using a subset of the dataset, with the performance evaluated based on accuracy and F-score. Results indicate the superior accuracy of SwinT (0.96), particularly when employing fat fraction (FF) images as input; it served as a valuable parameter for enhancing classification accuracy. Despite limitations, including a modest cohort size, this study provides valuable insights into the application of AI-driven approaches for precise neuromuscular disorder classification, with potential implications for improving patient care.

Multiparameter MRI-based radiomics nomogram for preoperative prediction of brain invasion in atypical meningioma:a multicentre study

ObjectiveTo develop a nomogram based on tumor and peritumoral edema (PE) radiomics features extracted from preoperative multiparameter MRI for predicting brain invasion (BI) in atypical meningioma (AM).MethodsIn this retrospective study, according to the 2021 WHO classification criteria, a total of 469 patients with pathologically confirmed AM from three medical centres were enrolled and divided into training (n = 273), internal validation (n = 117) and external validation (n = 79) cohorts. BI was diagnosed based on the histopathological examination. Preoperative contrast-enhanced T1-weighted MR images (T1C) and T2-weighted MR images (T2) for extracting meningioma features and T2-fluid attenuated inversion recovery (FLAIR) sequences for extracting meningioma and PE features were obtained. The multiple logistic regression was applied to develop separate multiparameter radiomics models for comparison. A nomogram was developed by combining radiomics features and clinical risk factors, and the clinical usefulness of the nomogram was verified using decision curve analysis.ResultsAmong the clinical factors, PE volume and PE/tumor volume ratio are the risk of BI in AM. The combined nomogram based on multiparameter MRI radiomics features of meningioma and PE and clinical indicators achieved the best performance in predicting BI in AM, with area under the curve values of 0.862 (95% CI, 0.819–0.905) in the training cohort, 0.834 (95% CI, 0.780–0.908) in the internal validation cohort and 0.867 (95% CI, 0.785–0.950) in the external validation cohort, respectively.ConclusionsThe nomogram based on tumor and PE radiomics features extracted from preoperative multiparameter MRI and clinical factors can predict the risk of BI in patients with AM.

Non-Alcoholic Fatty Liver Disease: Evaluation of Hepatic Steatosis with In-Phase and Out-Of-Phase MR Imaging

Non-Alcoholic Fatty Liver Disease: Evaluation of Hepatic Steatosis with In-Phase and Out-Of-Phase MR Imaging

Noninvasive platelet membrane‐coated Fe3O4 nanoparticles identify vulnerable atherosclerotic plaques

AbstractVulnerable atherosclerotic plaques serve as the primary pathological basis for fatal cardiovascular and cerebrovascular diseases. The precise identification and treatment of these vulnerable plaques hold paramount clinical importance in mitigating the incidence of myocardial infarction and stroke. Nevertheless, the identification of vulnerable plaques within the diffuse atherosclerotic plaques dispersed throughout the systemic circulation continues to pose a substantial challenge in clinical practice. Double emulsion solvent evaporation method, specifically the water‐in‐oil‐in‐water (W/O/W) technique, was employed to fabricate Fe3O4‐based poly (lactic‐co‐glycolic acid) (PLGA) nanoparticles (Fe3O4@PLGA). Platelet membranes (PM) were extracted through hypotonic lysis, followed by ultrasound‐assisted encapsulation onto the surface of Fe3O4@PLGA, resulting in the formation of PM‐coated Fe3O4 nanoparticles (PM/Fe3O4@PLGA). Characterization of PM/Fe3O4@PLGA involved the use of dynamic light scattering, transmission electron microscopy, western blotting, and magnetic resonance imaging (MRI). A model of atherosclerotic vulnerable plaques was constructed by carotid artery coarctation and a high‐fat diet fed to ApoE−/− (Apolipoprotein E knockout) mice. Immunofluorescence and MRI techniques were employed to verify the functionality of PM/Fe3O4@PLGA. In this study, we initially synthesized Fe3O4@PLGA as the core material. Subsequently, a platelet membrane was employed as a coating for the Fe3O4@PLGA, aiming to enable the detection of vulnerable atherosclerotic plaques through MRI. In vitro, PM/Fe3O4@PLGA not only exhibited excellent biosafety but also showed targeted collagen characteristics and MR imaging performance. In vivo, the adhesion of PM/Fe3O4@PLGA to atherosclerotic lesions was confirmed in a mouse model of vulnerable atherosclerotic plaques. Simultaneously, PM/Fe3O4@PLGA as a novel contrast agent for MRI has shown effective identification of vulnerable atherosclerotic plaques. In terms of safety profile in vivo, PM/Fe3O4@PLGA has not demonstrated significant organ toxicity or inflammatory response in the bloodstream. In this study, we successfully developed a platelet‐membrane‐coated nanoparticle system for the targeted delivery of Fe3O4@PLGA to vulnerable atherosclerotic plaques. This innovative system allows for the visualization of vulnerable plaques using MRI, thereby demonstrating its potential for enhancing the clinical diagnosis of vulnerable atherosclerotic plaques.

MR imaging features of uterine necrosis following modified B-Lynch sutures: a case report

BackgroundB-Lynch braided compression sutures are applied as a lifesaving procedure to control atonic post-partum hemorrhage if the patient is not responding to uterotonics and other conservative methods. These sutures are applied to avoid hysterectomy so as to preserve fertility. However, if not applied properly, these can lead to a rare but serious complication of uterine necrosis secondary to uterine ischemia and sepsis. The patient then ends up in hysterectomy defeating the sole purpose of maintaining fertility for which the sutures were applied at the first place.Case presentationHere we describe a case of primipara who had caesarian section complicated by post-partum hemorrhage which was managed with modified B-Lynch compression sutures for achieving hemostasis. The patient presented three weeks post-procedure with signs and symptoms of inflammation and sepsis. The MR imaging revealed features of uterine necrosis. The patient underwent hysterectomy unwillingly as no other option was left to save her life. A detailed MR imaging of this case has been described given the scarcity of MR features of uterine necrosis following compressive sutures in medical literature.ConclusionsThe case is reported to make the clinicians aware of the possibility of such a dreaded complication following compressive sutures for post-partum hemorrhage and to emphasize on early and timely follow-up of such patients so that uterus could be salvaged.

MADR-Net: multi-level attention dilated residual neural network for segmentation of medical images

Medical image segmentation has made a significant contribution towards delivering affordable healthcare by facilitating the automatic identification of anatomical structures and other regions of interest. Although convolution neural networks have become prominent in the field of medical image segmentation, they suffer from certain limitations. In this study, we present a reliable framework for producing performant outcomes for the segmentation of pathological structures of 2D medical images. Our framework consists of a novel deep learning architecture, called deep multi-level attention dilated residual neural network (MADR-Net), designed to improve the performance of medical image segmentation. MADR-Net uses a U-Net encoder/decoder backbone in combination with multi-level residual blocks and atrous pyramid scene parsing pooling. To improve the segmentation results, channel-spatial attention blocks were added in the skip connection to capture both the global and local features and superseded the bottleneck layer with an ASPP block. Furthermore, we introduce a hybrid loss function that has an excellent convergence property and enhances the performance of the medical image segmentation task. We extensively validated the proposed MADR-Net on four typical yet challenging medical image segmentation tasks: (1) Left ventricle, left atrium, and myocardial wall segmentation from Echocardiogram images in the CAMUS dataset, (2) Skin cancer segmentation from dermoscopy images in ISIC 2017 dataset, (3) Electron microscopy in FIB-SEM dataset, and (4) Fluid attenuated inversion recovery abnormality from MR images in LGG segmentation dataset. The proposed algorithm yielded significant results when compared to state-of-the-art architectures such as U-Net, Residual U-Net, and Attention U-Net. The proposed MADR-Net consistently outperformed the classical U-Net by 5.43%, 3.43%, and 3.92% relative improvement in terms of dice coefficient, respectively, for electron microscopy, dermoscopy, and MRI. The experimental results demonstrate superior performance on single and multi-class datasets and that the proposed MADR-Net can be utilized as a baseline for the assessment of cross-dataset and segmentation tasks.

Accelerated Screening of Protein-Ligand Interactions via Parallel T2-Weighted 19F-MRI.

In drug discovery, ligands are sought that modulate the (mal-)function of medicinally relevant target proteins. In order to develop new drugs, typically a multitude of potential ligands are initially screened for binding and subsequently characterized for their affinity. Nuclear magnetic resonance (NMR) is a well-established and highly sensitive technology for characterizing such interactions. However, it has limited throughput, because only one sample can be measured at a time. In contrast, magnetic resonance imaging (MRI) is inherently parallel and MR parameters can conveniently be encoded in its images, potentially offering increased sample throughput. We explore this application using a custom-built 9-fold sample holder and a 19F-MRI coil. With this setup, we show that ligand binding can be detected by T2-weighted 19F-MRI using 4-(trifluoromethyl)benzamidine (TFBA) and trypsin as the reporter ligand and target protein, respectively. Furthermore, we demonstrate that the affinity of nonfluorinated ligands can be determined in a competition format by monitoring the dose-dependent displacement of TFBA. By comparing 19F-T2-weighted MR images of TFBA in the presence of different benzamidine (BA) concentrations-all recorded in parallel-the affinity of BA could be derived. Therefore, this approach promises parallel characterization of protein-ligand interactions and increased throughput of biochemical assays, with potential for increased sensitivity when combined with hyperpolarization techniques.

Positional contrastive learning for improved thigh muscle segmentation in MR images.

The accurate segmentation of individual muscles is essential for quantitative MRI analysis of thigh images. Deep learning methods have achieved state-of-the-art results in segmentation, but they require large numbers of labeled data to perform well. However, labeling individual thigh muscles slice by slice for numerous volumes is a laborious and time-consuming task, which limits the availability of annotated datasets. To address this challenge, self-supervised learning (SSL) emerges as a promising technique to enhance model performance by pretraining the model on unlabeled data. A recent approach, called positional contrastive learning, exploits the information given by the axial position of the slices to learn features transferable on the segmentation task. The aim of this work was to propose positional contrastive SSL for the segmentation of individual thigh muscles from MRI acquisitions in a population of elderly healthy subjects and to evaluate it on different levels of limited annotated data. An unlabeled dataset of 72 T1w MRI thigh acquisitions was available for SSL pretraining, while a labeled dataset of 52 volumes was employed for the final segmentation task, split into training and test sets. The effectiveness of SSL pretraining to fine-tune a U-Net architecture for thigh muscle segmentation was compared with that of a randomly initialized model (RND), considering an increasing number of annotated volumes (S= 1, 2, 5, 10, 20, 30, 40). Our results demonstrated that SSL yields substantial improvements in Dice similarity coefficient (DSC) when using a very limited number of labeled volumes (e.g., for = 1, DSC 0.631 versus 0.530 for SSL and RND, respectively). Moreover, enhancements are achievable even when utilizing the full number of labeled subjects, with DSC = 0.927 for SSL and 0.924 for RND. In conclusion, positional contrastive SSL was effective in obtaining more accurate thigh muscle segmentation, even with a very low number of labeled data, with a potential impact of speeding up the annotation process in clinics.

Precise diagnosis of cardiac‐cerebral vascular diseases with magnetic resonance imaging‐based nanoprobes

AbstractCardiac‐cerebral vascular diseases (CCVDs) are acknowledged as a major threat to public health, leading to more than one‐third of all deaths worldwide. The complex anatomical structure and immune features of blood vessels significantly affect the development of CCVDs, and magnetic resonance angiography (MRA) is one of the main diagnostic approaches for the accurate diagnosis and prognosis of CCVDs. However, MRA suffers from intrinsic problems derived from its blood flow‐dependency, and the clinical Gd‐chelating contrast agents are limited by their rapid vascular extravasation. Over the past decade, spurred on by nanoscience and nanotechnology, numerous contrast agents based on magnetic nanomaterials have been developed to enhance the contrast of MRA, with these including iron oxide nanoparticles, rare earth‐doped nanoparticles, and metal‐organic coordination polymers. The molecular MR imaging of vasculopathy using specific nanoprobes has been explored to obtain a better understanding of the molecular aspects of CCVDs. In this review, the state of the art in MRA nanoprobes is introduced, and recent achievements in the diagnosis of CCVDs using MR imaging are summarized. Additionally, the future prospects and limitations of MRA based on nanoprobes are discussed. The current review provides methodological designs and ideas for subsequent MRA nanoprobes.

Normal cardiac dimensions by magnetic resonance imaging and topographic anatomy of the adult arabian one-humped camel (Camelus dromedarius)

BackgroundDromedaries’ normal heart architecture and size have not been adequately examined utilizing magnetic resonance imaging (MRI) and topographic anatomy.Resultwe aimed to investigate the regular appearance of the heart and its dimensions, using MRI and cross-sectional anatomy, in mature Arabian one-humped camels (Camelus dromedarius). We also analyzed hematological and cardiac biochemical markers. MRI scans were conducted on twelve camel heart cadavers using a closed 1.5-Tesla magnet with fast spin echo (FSE) weighted sequences. Subsequently, the hearts were cross-sectionally sliced. Additionally, hematobiochemical studies were conducted on ten mature live camels. The study analyzed standard cardiac dimensions including HL, BW, RA, LA, RV, LV, IVS, LAD, RAD, RVD, AoD, TCVD, and MVD. The results showed a strong positive correlation between the cardiac dimensions obtained from both gross analysis and MR images, with no significant difference between them. On both gross and MRI images, the usual structures of the heart were identified and labeled. Along with the cardiac markers (creatine kinase and troponin), the average hematological values and standard biochemical parameters were also described.ConclusionAccording to what we know, this investigation demonstrates, for the first time the typical heart structures and dimensions of the heart in dromedaries, and it could serve as a basis for diagnosing cardiac disorders in these animals.

Right Ventricular Strain Derived from Cardiac MRI Feature Tracking for the Diagnosis and Prognosis of Arrhythmogenic Right Ventricular Cardiomyopathy.

Purpose To demonstrate the myocardial strain characteristics of patients with arrhythmogenic right ventricular cardiomyopathy (ARVC), based on revised Task Force Criteria (rTFC), and to explore the prognostic value of strain analysis in ARVC. Materials and Methods This retrospective study included 247 patients (median age, 38 years [IQR, 28-48 years]; 167 male, 80 female) diagnosed with ARVC, based on rTFC, between 2014 and 2018. Patients were divided into "possible" (n =25), "borderline" (n = 40), and "definite" (n = 182) ARVC groups following rTFC. Biventricular global strain parameters were calculated using cardiac MRI feature tracking (FT). The primary outcome was defined as a composite of cardiovascular events, including cardiovascular death, heart transplantation, and appropriate implantable cardioverter defibrillator discharge. Univariable and multivariable cumulative logistic regression and Cox proportional hazards regression analysis were used to evaluate the diagnostic and prognostic value of right ventricle (RV) strain parameters. Results Patients with definite ARVC had significantly reduced RV global strain in all three directions compared with possible or borderline groups (all P < .001). RV global longitudinal strain (GLS) was an independent predictor for disease (odds ratio, 1.09 [95% CI: 1.02, 1.16]; P = .009). During a median follow-up of 3.4 years (IQR, 2.0-4.9 years), 55 patients developed primary end point events. Multivariable analysis showed that RV GLS was independently associated with the occurrence of cardiovascular events (hazard ratio, 1.15 [95% CI: 1.07, 1.24]; P < .001). Kaplan-Meier analysis showed that patients with RV GLS worse than median had a higher risk of combined cardiovascular events (log-rank P < .001). Conclusion RV GLS derived from cardiac MRI FT demonstrated good diagnostic and prognostic value in ARVC. Keywords: MR Imaging, Image Postprocessing, Cardiac, Right Ventricle, Cardiomyopathies, Arrhythmogenic Right Ventricular Cardiomyopathy, Revised Task Force Criteria, Cardiovascular MR, Feature Tracking, Cardiovascular Events Supplemental material is available for this article. © RSNA, 2024.

Manganese carbonate superparticles as DNA- and pH-responsive magnetic resonance imaging contrast agents

Manganese carbonate nanoparticles (NPs) have been developed as pH-triggered responsive magnetic resonance imaging (MRI) contrast agents for tumor diagnosis, but the signal amplification of MRI is still limited due to the lack of response to other pathological parameters of tumors. Here, random single-stranded deoxyribonucleic acid (ssDNA) was explored as a biomimetic template to prepare manganese carbonate superparticles (MnCO3 SPs) assembled from ultrasmall particles. We focused on the degradation ability of DNA-MnCO3 SPs triggered by DNA and its trigger mechanism. The results show that different sequences of DNA could trigger the disassembly of these DNA-MnCO3 SPs and result in MR signal amplification. Further investigations show the DNA-MnCO3 SPs possess good physiological stability, but increased degradation sensitivity under dual triggering of DNA and low pH. In vivo MR and FL dual-modal imaging for tumor region display T1-signal rapid amplification, suggesting rapid responsiveness of DNA-MnCO3 SPs to DNA. Therefore, the design of DNA- and pH-triggered DNA-MnCO3 SPs may provide a new idea for the construction of next-generation activatable contrast agents with high specificity and high sensitivity.

Deep learning-based quantification of osteonecrosis using magnetic resonance images in Gaucher disease

Gaucher disease is one of the most common lysosomal storage disorders. Osteonecrosis is a principal clinical manifestation of Gaucher disease and often leads to joint collapse and fractures. T1-weighted (T1w) modality in MRI is widely used to monitor bone involvement in Gaucher disease and to diagnose osteonecrosis. However, objective and quantitative methods for characterizing osteonecrosis are still limited. In this work, we present a deep learning-based quantification approach for the segmentation of osteonecrosis and the extraction of characteristic parameters. We first constructed two independent U-net models to segment the osteonecrosis and bone marrow unaffected by osteonecrosis (UBM) in spine and femur respectively, based on T1w images from patients in the UK national Gaucherite study database. We manually delineated parcellation maps including osteonecrosis and UBM from 364 T1w images (176 for spine, 188 for femur) as the training datasets, and the trained models were subsequently applied to all the 917 T1w images in the database. To quantify the segmentation, we calculated morphological parameters including the volume of osteonecrosis, the volume of UBM, and the fraction of total marrow occupied by osteonecrosis. Then, we examined the correlation between calculated features and the bone marrow burden score for marrow infiltration of the corresponding image, and no strong correlation was found. In addition, we analyzed the influence of splenectomy and the interval between the age at first symptom and the age of onset of treatment on the quantitative measurements of osteonecrosis. The results are consistent with previous studies, showing that prior splenectomy is closely associated with the fractional volume of osteonecrosis, and there is a positive relationship between the duration of untreated disease and the quantifications of osteonecrosis. We propose this technique as an efficient and reliable tool for assessing the extent of osteonecrosis in MR images of patients and improving prediction of clinically important adverse events.

Impact of gadopiclenol on decision making in patients with brain metastases: A post-hoc study.

e14003 Background: Gadopiclenol (Elucirem, Guerbet, France) is a new high relaxivity macrocyclic gadolinium-based contrast agent (GBCA), recently approved by the FDA and EMA at the dose of 0.05 mmol/kg. The aim of this post-hoc analysis was to evaluate the impact of contrast-enhanced MRI with gadopiclenol at 0.1 mmol/kg on decision making and radiotherapy (RT) treatment planning of brain metastases (BM). Methods: This is a post-hoc analysis of data from a phase IIb study where patients underwent two separate MRI examinations, one with gadopiclenol at 0.025, 0.05, 0.1 or 0.2 mmol/kg and one with gadobenate dimeglumine at 0.1 mmol/kg. MR images of patients who received both GBCAs at 0.1 mmol/kg, with ≥1 BM detected in either of the scans, were subjected to a blinded reader analysis and contouring by two radiation oncology experts. For each patient, treatment plans (stereotactic radiosurgery [SRS] or whole-brain radiotherapy [WBRT]) were determined for both MRIs, with the gross target volume (GTV) indicating the contrast-enhancing aspects of the tumor. Mean GTVs and normal tissue volumes receiving 12 Gy (V12), as well as the Dice similarity coefficient (DSC) were obtained for the paired contours. The Spearman´s rank (ρ) correlation was additionally calculated. Furthermore, three experts blindly evaluated the contrast enhancement of each lesion for contouring purposes and subjectively qualified them as “better”, in detriment of the counterpart, or “equal”. Results: In total, images from 13 adult patients were analyzed. MRI with gadopiclenol depicted additional BM as compared with gadobenate dimeglumine in 7 patients (54%). The treatment plan was changed in 2 patients (15%), from no treatment to SRS and from SRS to WBRT. Gadopiclenol depicted additional BM in these 2 patients (from 0 and 10 BM with gadobenate dimeglumine to 1 and 15 BM with gadopiclenol). The mean GTVs and V12 were comparable between gadopiclenol and gadobenate dimeglumine (p=0.694, p=1.974). The mean DSC was 0.70 (SD: 0.14, ρ0.82). From a total of 36 answers, contrast enhancement was qualified as better with gadopiclenol in 21 (58.3%) evaluations, better with gadobenate dimeglumine in 8 (22.2%) evaluations, while no difference was observed in 7 (19.4%) evaluations. Conclusions: Gadopiclenol at 0.1 mmol/kg improved BM detection and contrast enhancement with potential impact on RT treatment decisions.

Reduced functional connectivity induced by longitudinal alterations of structure and perfusion may be associated with cognitive impairment in patients on maintenance hemodialysis.

Hemodialysis (HD) leads to cognitive impairment; however, the pathophysiology of maintenance HD remains unclear. This study aimed to investigate the longitudinal alterations in gray matter volume (GMV) and cerebral blood flow (CBF) in patients on HD at follow-up compared with baseline, examine the alterations in functional connectivity (FC) by defining co-changed brain regions as seed points, and investigate the correlation between the co-changed brain regions and neuropsychological test scores. Twenty-seven patients with HD and 30 healthy controls were enrolled in this study. All participants underwent high-resolution T1-weighted imaging, arterial spin labeling, and functional MR imaging to measure GMV, CBF, and FC. The patients on HD were assessed at baseline and 3years subsequently. The right and left medial superior frontal gyrus (SFGmed.L) exhibited significantly lower GMV and CBF in patients on HD at follow-up compared with baseline and lower FC between the SFGmed.L and left middle temporal gyrus (MTG.L). FC between the SFGmed.L and MTG.L was positively correlated with neuropsychological test scores in the HD group at follow-up. Reduced GMV and CBF may result in decreased FC between the SFGmed.L and MTG.L, which may be associated with cognitive impairment in patients on maintenance HD. Our findings provide unique insights into the pathological mechanisms of patients on maintenance HD with cognitive impairment.

Delta-radiomics Analysis Based on Multi-phase Contrast-enhanced MRI to Predict Early Recurrence in Hepatocellular Carcinoma After Percutaneous Thermal Ablation

Rationale and ObjectivesIt is critical to predict early recurrence (ER) after percutaneous thermal ablation (PTA) for hepatocellular carcinoma (HCC). We aimed to develop and validate a delta-radiomics nomogram based on multi-phase contrast-enhanced magnetic resonance imaging (MRI) to preoperatively predict ER of HCC after PTA. Materials and MethodsWe retrospectively enrolled 164 patients with HCC and divided them into training, temporal validation, and other-scanner validation cohorts (n=110, 29, and 25, respectively). The volumes of interest of the intratumoral and/or peritumoral regions were delineated on preoperative multi-phase MR images. Original radiomics features were extracted from each phase, and delta-radiomics features were calculated. Logistic regression was used to train the corresponding radiomics models. The clinical and radiological characteristics were evaluated and combined to establish a clinical-radiological model. A fusion model comprising the best radiomics scores and clinical-radiological risk factors was constructed and presented as a nomogram. The performance of each model was evaluated and recurrence-free survival (RFS) was assessed. ResultsChild-Pugh grade B, high-risk tumor location, and an incomplete/absent tumor capsule were independent predictors of ER. The optimal radiomics model comprised 12 delta-radiomics features with areas under the curve (AUCs) of 0.834, 0.795, and 0.769 in the training, temporal validation, and other-scanner validation cohorts, respectively. The nomogram showed the best predictive performance with AUCs as 0.893, 0.854, and 0.827 in the three datasets. There was a statistically significant difference in RFS between the risk groups calculated using the delta-radiomics model and nomogram. ConclusionsThe nomogram combined with the delta-radiomic score and clinical-radiological risk factors could non-invasively predict ER of HCC after PTA.

High-Dimensional MR Reconstruction Integrating Subspace and Adaptive Generative Models.

To develop a new method that integrates subspace and generative image models for high-dimensional MR image reconstruction. We proposed a formulation that synergizes a low-dimensional subspace model of high-dimensional images, an adaptive generative image prior serving as spatial constraints on the sequence of "contrast-weighted" images or spatial coefficients of the subspace model, and a conventional sparsity regularization. A special pretraining plus subject-specific network adaptation strategy was proposed to construct an accurate generative-network-based representation for images with varying contrasts. An iterative algorithm was introduced to jointly update the subspace coefficients and the multi-resolution latent space of the generative image model that leveraged an recently proposed intermediate layer optimization technique for network inversion. We evaluated the utility of the proposed method for two high-dimensional imaging applications: accelerated MR parameter mapping and high-resolution MR spectroscopic imaging. Improved performance over state-of-the-art subspace-based methods was demonstrated in both cases. The proposed method provided a new way to address high-dimensional MR image reconstruction problems by incorporating an adaptive generative model as a data-driven spatial prior for constraining subspace reconstruction. Our work demonstrated the potential of integrating data-driven and adaptive generative priors with canonical low-dimensional modeling for high-dimensional imaging problems.

Signal intensity and volume of carotid intraplaque hemorrhage on magnetic resonance imaging and the risk of ipsilateral cerebrovascular events: The Plaque At RISK (PARISK) study

BackgroundThe Plaque at RISK (PARISK) study demonstrated that patients with a carotid plaque with intraplaque hemorrhage (IPH) have an increased risk of recurrent ipsilateral ischemic cerebrovascular events. It was previously reported that symptomatic carotid plaques with IPH showed higher IPH signal intensity ratios (SIR) and larger IPH volumes than asymptomatic plaques. We explored whether IPH SIR and IPH volume are associated with future ipsilateral ischemic cerebrovascular events beyond the presence of IPH. MethodsTIA and ischemic stroke patients with mild-to-moderate carotid stenosis and an ipsilateral IPH-positive carotid plaque (n=89) from the PARISK study were included. The clinical endpoint was a new ipsilateral ischemic cerebrovascular event during 5 years of follow-up, while the imaging-based endpoint was a new ipsilateral brain infarct on brain MRI after 2 years (n=69). Trained observers delineated IPH, a hyperintense region compared to surrounding muscle tissue on hyper T1-weighted MR images. The IPH SIR was the maximal signal intensity in the IPH region divided by the mean signal intensity of adjacent muscle tissue. The associations between IPH SIR or volume and the clinical and imaging-based endpoint were investigated using Cox-proportional hazard models and logistic regression, respectively. ResultsDuring 5.1 (interquartile range (IQR): 3.1-5.6) years of follow-up, 21 ipsilateral cerebrovascular ischemic events were identified. Twelve new ipsilateral brain infarcts were identified on the 2-year neuro MRI. There was no association for IPH SIR or IPH volume with the clinical endpoint (HR: 0.89 [95% CI: 0.67-1.10] and HR: 0.91 [0.69-1.19] per 100µl increase, respectively) nor with the imaging-based endpoint (OR: 1.04 [0.75-1.45] and OR: 1.21 [0.87-1.68] per 100µl increase, respectively). ConclusionsIPH SIR and IPH volume were not associated with future ipsilateral ischemic cerebrovascular events. Therefore, quantitative assessment of IPH does not seem to provide additional value beyond the presence of IPH for stroke risk assessment.Trial registrationThe PARISK study was registered on ClinicalTrials.gov with ID NCT01208025 on 21 September 2010 (https://clinicaltrials.gov/study/NCT01208025).

Evaluation of treatment response by multiparametric MR imaging in locally advanced rectal tumors following neoadjuvant chemotherapy.

To investigate the effectiveness of multiparametric MRI examination in determining tumor response after neoadjuvant chemoradiotherapy (CRT) in locally advanced rectal tumors. 46 patients with locally advanced rectal adenocarcinoma were included and were divided into two groups as complete responders and nonresponders based on Mandard score. On MRI, relative T2w signal intensity and ADC values obtained before and after treatment and tumour volumes in dynamic contrast enhanced images (DCI) were used to determine complete response to treatment. There were no significant differences between mean ADC values obtained by single slice ADC and three circular ROI methods. There were significant differences between two groups in terms of Post-CRT ADC value, ΔADC and %ΔADC obtained by whole tumour volume ADC method (p < 0.05). There were significant differences between Pre-CRT and Post-CRT volume values. ΔV DCI and %ΔV DCI, ΔV ADC and T2w volume values were significantly lower in complete responders (p < 0.05). In multivariate analysis, sensitivity and specificity were calculated as 88.9% and 91.9% (AUC = 0.943) when Post-CRT mean ADC value and Post-CRT DCI volume values were used together, and sensitivity and specificity were calculated as 88.9% and 94.6% (AUC = 0.949) when ΔADC and Post-CRT DCI volume values were used together. Whole tumour volume mean ADC value is the most useful method to determine treatment response. Post-CRT DCI volume measurement stands out as the most useful method in assessing complete response alone. The highest diagnostic values are achieved when the post-CRT DCI volume is combined with the ADC change value of the whole tumor volume.

Hypoperfusion of Amygdala in Chronic Migraine: An Exploratory Quantitative Perfusion Imaging Using 3D Pseudo-Continuous Arterial Spin Labeling.

Although the amygdala has structural and functional abnormalities in Chronic Migraine (CM), less is known about the altered perfusion of the amygdala in CM. The current study aimed to assess amygdala perfusion in CM using a contrast agent-free and quantitative approach. 15 Normal Controls (NC) and 13 patients with CM during the migraine interval were assessed for brain structure and subjected to 3D Pseudo- Continuous Arterial Spin Labeling (3D-PCASL) MR imaging. The Cerebral Blood Flow (CBF) value of the amygdala was automatically extracted based on the individual amygdala mask for all participants. The independent sample t-test, Receiver Operating Characteristic (ROC) curve, and correlation analysis were used to evaluate the perfusion changes in CM. Bilateral amygdala cerebral perfusion was lower in CM (left amygdala, 42.21±4.49 ml/100mg/min; right amygdala, 42.38±4.41 ml/100mg/min) than in NC (left amygdala, 48.31±6.92 ml/100mg/min; right amygdala, 47.88±6.53 ml/100mg/min) (left, p = 0.01; right, p = 0.02). There was no significant correlation between the perfusion of bilateral amygdalas and the clinical variables. Also, there was no significant difference in the volume of bilateral amygdalas between the two groups. The Area Under the Curve (AUC) of the CBF values of the left and right amygdala was 0.78 (95%CI: 0.58-0.91) and 0.75 (95%CI: 0.55-0.89), respectively. The cut-off value was 44.24 ml/100mg/min (left amygdala, with sensitivity 76.90% and specificity 78.70%) and 46.75 ml/100mg/min (right amygdala, with sensitivity 92.3% and specificity 58.80%), respectively. CM presented bilateral hypoperfusion in the amygdala, offering potential diagnostic value in distinguishing CM from NC. The 3D-PCASL could be regarded as a simple and efficient neuroimaging tool to assess the perfusion status in CM patients.