Automated Detection of Cervical Spinal Stenosis and Cord Compression via Vision Transformer and Rules-Based Classification.

BACKGROUND AND PURPOSE: Cervical spinal cord compression, defined as spinal cord deformity and severe narrowing of the spinal canal in the cervical region, can lead to severe clinical consequences, including intractable pain, sensory disturbance, paralysis, and even death, and may require emergent intervention to prevent negative outcomes. Despite the critical nature of cord compression, no automated tool is available to alert clinical radiologists to the presence of such findings. This study aims to demonstrate the ability of a vision transformer (ViT) model for the accurate detection of cervical cord compression. MATERIALS AND METHODS: A clinically diverse cohort of 142 cervical spine MRIs was identified, 34% of which were normal or had mild stenosis, 31% with moderate stenosis, and 35% with cord compression. Utilizing gradient-echo images, slices were labeled as no cord compression/mild stenosis, moderate stenosis, or severe stenosis/cord compression. Segmentation of the spinal canal was performed and confirmed by neuroradiology faculty. A pretrained ViT model was fine-tuned to predict section-level severity by using a train:validation:test split of 60:20:20. Each examination was assigned an overall severity based on the highest level of section severity, with an examination labeled as positive for cord compression if ≥1 section was predicted in the severe category. Additionally, 2 convolutional neural network (CNN) models (ResNet50, DenseNet121) were tested in the same manner. RESULTS: The ViT model outperformed both CNN models at the section level, achieving section-level accuracy of 82%, compared with 72% and 78% for ResNet and DenseNet121, respectively. ViT patient-level classification achieved accuracy of 93%, sensitivity of 0.90, positive predictive value of 0.90, specificity of 0.95, and negative predictive value of 0.95. Receiver operating characteristic area under the curve was greater for ViT than either CNN. CONCLUSIONS: This classification approach using a ViT model and rules-based classification accurately detects the presence of cervical spinal cord compression at the patient level. In this study, the ViT model outperformed both conventional CNN approaches at the section and patient levels. If implemented into the clinical setting, such a tool may streamline neuroradiology workflow, improving efficiency and consistency.

A kinematic model-based analysis framework for 3D Cine-DENSE-validation with an axially compressed gel phantom and application in sheep before and after antero-apical myocardial infarction.

PURPOSE: Myocardial strain is increasingly used to assess left ventricular (LV) function. Incorporation of LV deformation into finite element (FE) modeling environment with subsequent strain calculation will allow analysis to reach its full potential. We describe a new kinematic model-based analysis framework (KMAF) to calculate strain from 3D cine-DENSE (displacement encoding with stimulated echoes) MRI. METHODS: Cine-DENSE allows measurement of 3D myocardial displacement with high spatial accuracy. The KMAF framework uses cine cardiovascular magnetic resonance (CMR) to facilitate cine-DENSE segmentation, interpolates cine-DENSE displacement, and kinematically deforms an FE model to calculate strain. This framework was validated in an axially compressed gel phantom and applied in 10 healthy sheep and 5 sheep after myocardial infarction (MI). RESULTS: Excellent Bland-Altman agreement of peak circumferential (Ecc ) and longitudinal (Ell ) strain (mean difference = 0.021 ± 0.04 and -0.006 ± 0.03, respectively), was found between KMAF estimates and idealized FE simulation. Err had a mean difference of -0.014 but larger variation (±0.12). Cine-DENSE estimated end-systolic (ES) Ecc , Ell and Err exhibited significant spatial variation for healthy sheep. Displacement magnitude was reduced on average by 27%, 42%, and 56% after MI in the remote, adjacent and MI regions, respectively. CONCLUSIONS: The KMAF framework allows accurate calculation of 3D LV Ecc and Ell from cine-DENSE.

Electric scooter craniofacial trauma.

OBJECTIVE: The use of standing electronic scooters associated with micromobility applications (e-scooters) has risen nationally. The aim of this study was to obtain a detailed view of soft tissue and bony craniofacial injury associated with e-scooter-related trauma. METHODS: Single-institution retrospective case series of patients presenting to a level 1 trauma center emergency department or trauma unit with documented e-scooter-related craniofacial injury. RESULTS: Of 203 included patients, 188 (92.6%) patients sustained craniofacial injury. One hundred thirty-one (64.5%) had exclusively soft tissue injury, 3 (1.5%) exclusively bony injury, 51 (25.1%) both soft and bony injuries, and twenty-five (12.3%) patients sustained dental injury. Aesthetic units most frequently sustaining acute soft tissue injury were the forehead (n = 106, 34.6%), scalp (n = 36, 11.8%), chin (n = 34, 11.1%), upper lip (n = 32, 10.5%), and cheek (n = 31, 10.1%). Aesthetic subunits most often sustaining acute soft tissue injury included the brow (42, 13.7%), central forehead (39, 12.7%), lateral forehead (n = 25, 8.2%), and upper lip vermillion (n = 23, 7.5%). Craniofacial osseous fracture most often occurred in the orbit (n = 42, 24.6%) and maxilla (n = 40, 23.4%). Individual osseous segments most frequently sustaining acute fracture included the anterior maxillary sinus wall (n = 22, 12.9%), nasal bone (n = 20, 11.7%), lateral orbital wall (n = 16, 9.4%), orbital floor (n = 15, 8.8%), and zygomatic bone (13, 7.6%). CONCLUSIONS: Our analysis demonstrates that most patients presenting to our center with craniofacial trauma sustained acute bony fracture, most often to the midface. Our data of common injuries associated with e-scooter trauma could inform implementation in the form of facial safety equipment or safety skills training for e-scooter riders. LEVEL OF EVIDENCE: 4.

Radiologic Modalities and Response Assessment Schemes for Clinical and Preclinical Oncology Imaging.

Clinical drug trials for oncology have resulted in universal protocols for medical imaging in order to standardize protocols for image procurement, radiologic interpretation, and therapeutic response assessment. In recent years, there has been increasing interest in using large animal models to study oncologic disease, though few standards currently exist for imaging of large animal models. This article briefly reviews medical imaging modalities, the current state-of-the-art in radiologic diagnostic criteria and response assessment schemes for evaluating therapeutic response and disease progression, and translation of radiologic imaging protocols and standards to large animal models of malignant disease.

Shape-appearance constrained segmentation and separation of vein and artery in pulsatile tinnitus patients based on MR angiography and flow MRI.

This study reports on the development and evaluation of a novel segmentation method for extracting the internal jugular vein and the adjacent carotid artery from magnetic resonance (MR) images of patients with pulsatile tinnitus. A narrow band level set method with combined shape and appearance constraints was developed and applied to high-resolution MR images from 17 pulsatile tinnitus patients (age 52 ±â€¯23 years, 10 females). The proposed method was validated by comparing with the manual segmentation as well as by identifying the jugular vein and carotid artery based on 4D flow MRI in which the two types of vessels have opposing flow. Our study showed that the vein and artery are in contact with each other on 30.2% of all the slices. Dice value, Peak signal-to-noise ratio (PSNR), Hausdorff distance and mean sum of square distance (MSSD) between automatic and manual segmentation were 89.13 ±â€¯2.84%, 27.36 ±â€¯2.39%, 17.2 ±â€¯6.9 mm, 7.4 ±â€¯5.5 mm, demonstrating good segmentation accuracy. The average Dice similarity coefficient and the coefficient of variation compared with 4D flow MRI was 91.42 ±â€¯1.63% and 89.28 ±â€¯4.54% for the internal jugular vein and the carotid artery. The present pipeline for automatic internal jugular vein quantification holds promise for efficient image interpretation in large-scale cohort studies.

Visualizing wall enhancement over time in unruptured intracranial aneurysms using 3D vessel wall imaging.

BACKGROUND: Few studies directed at assessing the visualization of the walls of unruptured aneurysms have used higher-resolution 3D MRI vessel wall imaging. Prospective longitudinal studies are also needed to screen vessel wall changes in unruptured aneurysms. PURPOSE: To compare the aneurysm wall visualization on pre- and post-3D isotropic T1 -weighted Sampling Perfection with Application-optimized Contrasts by using different flip angle Evolutions (SPACE) images and to explore whether there is a change in wall enhancement at follow up. STUDY TYPE: Prospective. POPULATION: Twenty-nine patients with a total of 35 unruptured intracranial aneurysms. SEQUENCE: 3D T1 -weighted pre- and postcontrast SPACE (0.5 mm isotropic) at 3.0T. ASSESSMENT: The aneurysm wall visibility (0-5 scale) between pre- and postcontrast images as well as the wall enhancement (0-5 scale) between follow-up and baseline studies (6-30 months, average 12.7 months) were compared. Differences in wall visibility and enhancement were also investigated as a function of aneurysm diameter and location. STATISTICAL TEST: The Wilcoxon signed rank paired test was used to compare the wall visibility score between pre- and postcontrast SPACE images, as well as wall enhancement between follow-up and baseline. The Mann-Whitney and Kruskal-Wallis tests were used to investigate the enhancement difference between different diameters and locations. RESULTS: Postcontrast images had significantly higher wall visibility (P = 0.01). A wall enhancement score ≥2 was found in 71% of the aneurysms. Changes in levels of wall enhancement were found in 17% of the aneurysms at follow-up studies, but those changes were small. Wall visibility and enhancement scores of large aneurysms were significantly higher than small ones (P < 0.001). DATA CONCLUSION: 3D T1 -weighted higher resolution SPACE can be used to assess changes in enhancement at follow-up studies. Contrast SPACE image provides better aneurysm wall visibility and improves visualization of the aneurysm wall. LEVEL OF EVIDENCE: 2 Technical Efficacy: Stage 3 J. Magn. Reson. Imaging 2019;50:193-200.

Short term doxycycline treatment induces sustained improvement in myocardial infarction border zone contractility.

Decreased contractility in the non-ischemic border zone surrounding a MI is in part due to degradation of cardiomyocyte sarcomeric components by intracellular matrix metalloproteinase-2 (MMP-2). We recently reported that MMP-2 levels were increased in the border zone after a MI and that treatment with doxycycline for two weeks after MI was associated with normalization of MMP-2 levels and improvement in ex-vivo contractile protein developed force in the myocardial border zone. The purpose of the current study was to determine if there is a sustained effect of short term treatment with doxycycline (Dox) on border zone function in a large animal model of antero-apical myocardial infarction (MI). Antero-apical MI was created in 14 sheep. Seven sheep received doxycycline 0.8 mg/kg/hr IV for two weeks. Cardiac MRI was performed two weeks before, and then two and six weeks after MI. Two sheep died prior to MRI at six weeks from surgical/anesthesia-related causes. The remaining 12 sheep completed the protocol. Doxycycline induced a sustained reduction in intracellular MMP-2 by Western blot (3649±643 MI+Dox vs 9236±114 MI relative intensity; p = 0.0009), an improvement in ex-vivo contractility (65.3±2.0 MI+Dox vs 39.7±0.8 MI mN/mm2; p<0.0001) and an increase in ventricular wall thickness at end-systole 1.0 cm from the infarct edge (12.4±0.6 MI+Dox vs 10.0±0.5 MI mm; p = 0.0095). Administration of doxycycline for a limited two week period is associated with a sustained improvement in ex-vivo contractility and an increase in wall thickness at end-systole in the border zone six weeks after MI. These findings were associated with a reduction in intracellular MMP-2 activity.

Highly accelerated intracranial 4D flow MRI: evaluation of healthy volunteers and patients with intracranial aneurysms.

OBJECTIVES: To evaluate an accelerated 4D flow MRI method that provides high temporal resolution in a clinically feasible acquisition time for intracranial velocity imaging. MATERIALS AND METHODS: Accelerated 4D flow MRI was developed by using a pseudo-random variable-density Cartesian undersampling strategy (CIRCUS) with the combination of k-t, parallel imaging and compressed sensing image reconstruction techniques (k-t SPARSE-SENSE). Four-dimensional flow data were acquired on five healthy volunteers and eight patients with intracranial aneurysms using CIRCUS (acceleration factor of R = 4, termed CIRCUS4) and GRAPPA (R = 2, termed GRAPPA2) as the reference method. Images with three times higher temporal resolution (R = 12, CIRCUS12) were also reconstructed from the same acquisition as CIRCUS4. Qualitative and quantitative image assessment was performed on the images acquired with different methods, and complex flow patterns in the aneurysms were identified and compared. RESULTS: Four-dimensional flow MRI with CIRCUS was achieved in 5 min and allowed further improved temporal resolution of <30 ms. Volunteer studies showed similar qualitative and quantitative evaluation obtained with the proposed approach compared to the reference (overall image scores: GRAPPA2 3.2 ± 0.6; CIRCUS4 3.1 ± 0.7; CIRCUS12 3.3 ± 0.4; difference of the peak velocities: -3.83 ± 7.72 cm/s between CIRCUS4 and GRAPPA2, -1.72 ± 8.41 cm/s between CIRCUS12 and GRAPPA2). In patients with intracranial aneurysms, the higher temporal resolution improved capturing of the flow features in intracranial aneurysms (pathline visualization scores: GRAPPA2 2.2 ± 0.2; CIRCUS4 2.5 ± 0.5; CIRCUS12 2.7 ± 0.6). CONCLUSION: The proposed rapid 4D flow MRI with a high temporal resolution is a promising tool for evaluating intracranial aneurysms in a clinically feasible acquisition time.

Segmentation of lumen and outer wall of abdominal aortic aneurysms from 3D black-blood MRI with a registration based geodesic active contour model.

Segmentation of the geometric morphology of abdominal aortic aneurysm is important for interventional planning. However, the segmentation of both the lumen and the outer wall of aneurysm in magnetic resonance (MR) image remains challenging. This study proposes a registration based segmentation methodology for efficiently segmenting MR images of abdominal aortic aneurysms. The proposed methodology first registers the contrast enhanced MR angiography (CE-MRA) and black-blood MR images, and then uses the Hough transform and geometric active contours to extract the vessel lumen by delineating the inner vessel wall directly from the CE-MRA. The proposed registration based geometric active contour is applied to black-blood MR images to generate the outer wall contour. The inner and outer vessel wall are then fused presenting the complete vessel lumen and wall segmentation. The results obtained from 19 cases showed that the proposed registration based geometric active contour model was efficient and comparable to manual segmentation and provided a high segmentation accuracy with an average Dice value reaching 89.79%.

Flow patterns in the jugular veins of pulsatile tinnitus patients.

Pulsatile Tinnitus (PT) is a pulse-synchronous sound heard in the absence of an external source. PT is often related to abnormal flow in vascular structures near the cochlea. One vascular territory implicated in PT is the internal jugular vein (IJV). Using computational fluid dynamics (CFD) based on patient-specific Magnetic Resonance Imaging (MRI), we investigated the flow within the IJV of seven subjects, four symptomatic and three asymptomatic of PT. We found that there were two extreme anatomic types classified by the shape and position of the jugular bulbs: elevated and rounded. PT patients had elevated jugular bulbs that led to a distinctive helical flow pattern within the proximal internal jugular vein. Asymptomatic subjects generally had rounded jugular bulbs that neatly redirected flow from the sigmoid sinus directly into the jugular vein. These two flow patterns were quantified by calculating the length-averaged streamline curvature of the flow within the proximal jugular vein: 130.3±8.1m-1 for geometries with rounded bulbs, 260.7±29.4m-1 for those with elevated bulbs (P<0.005). Our results suggest that variations in the jugular bulb geometry lead to distinct flow patterns that are linked to PT, but further investigation is needed to determine if the vortex pattern is causal to sound generation.

The manifestation of vortical and secondary flow in the cerebral venous outflow tract: An in vivo MR velocimetry study.

Aberrations in flow in the cerebral venous outflow tract (CVOT) have been implicated as the cause of several pathologic conditions including idiopathic intracranial hypertension (IIH), multiple sclerosis (MS), and pulsatile tinnitus (PT). The advent of 4D flow magnetic resonance imaging (4D-flow MRI) has recently allowed researchers to evaluate blood flow patterns in the arterial structures with great success. We utilized similar imaging techniques and found several distinct flow characteristics in the CVOT of subjects with and without lumenal irregularities. We present the flow patterns of 8 out of 38 subjects who have varying heights of the internal jugular bulb and varying lumenal irregularities including stenosis and diverticulum. In the internal jugular vein (IJV) with an elevated jugular bulb (JB), 4Dflow MRI revealed a characteristic spiral flow that was dependent on the level of JB elevation. Vortical flow was also observed in the diverticula of the venous sinuses and IJV. The diversity of flow complexity in the CVOT illustrates the potential importance of hemodynamic investigations in elucidating venous pathologies.

Isotropic 3D black blood MRI of abdominal aortic aneurysm wall and intraluminal thrombus.

INTRODUCTION: The aortic wall and intraluminal thrombus (ILT) have been increasingly studied as potential markers of progressive disease with abdominal aortic aneurysms (AAAs). Our goal was to develop a high resolution, 3D black blood MR technique for AAA wall and ILT imaging within a clinically acceptable scan time. METHODS: Twenty two patients with AAAs (maximal diameter 4.3±1.0cm), along with five healthy volunteers, were imaged at 3T with a 3D T1-weighted fast-spin-echo sequence using variable flip angle trains (SPACE) with a preparation pulse (DANTE) for suppressing blood signal. Volunteers and ten patients were also scanned with SPACE alone for comparison purposes. The signal to noise ratio (SNR) and the aortic wall/ILT to lumen contrast to noise ratio (CNR) were measured. Qualitative image scores (1-4 scale) assessing the inner lumen and outer wall boundaries of AAA were performed by two blinded reviewers. In patients with ILT, the ratio of ILT signal intensity (ILTSI) over psoas muscle SI (MuscleSI) was calculated, and the signal heterogeneity of ILT was quantified as standard deviation (SD) over the mean. RESULTS: All subjects were imaged successfully with an average scan time of 7.8±0.7minutes. The DANTE preparation pulse for blood suppression substantially reduced flow artifacts in SPACE with lower lumen SNR (8.8 vs. 21.4, p<0.001) and improved the wall/ILT to lumen CNR (9.9 vs. 6.3, p<0.001) in patients. Qualitative assessment showed improved visualization of lumen boundaries (73% higher scores on average, p=0.01) and comparable visualization of outer wall boundary (p>0.05). ILT was present in ten patients, with relatively high signal and a wide SD (average ILTSI/MuscleSI 1.42±0.48 (range 0.75-2.11)) and with SD/mean of 27.7%±6.6% (range 19.6%-39.4%). CONCLUSION: High resolution, 3D black blood MRI of AAAs can be achieved in a clinical accepted scan time with reduction of flow artifacts using the DANTE preparation pulse. Signal characteristics of ILT can be quantified and may be used for improved patient-specific risk stratification.

Vascular Imaging With Ferumoxytol as a Contrast Agent.

OBJECTIVE: Ferumoxytol is increasingly reported as an alternative to gadolinium-based contrast agents for MR angiography (MRA), particularly for patients with renal failure. This article summarizes more than 3 years of clinical experience with ferumoxytol-enhanced MRA for a range of indications and anatomic regions. CONCLUSION: Ferumoxytol-enhanced MRA has many advantages including that it is safe for patients with renal failure and provides a lengthy plateau of vascular signal as a blood pool agent that allows longer navigated MRA sequences.