Ultrafast lumbar spine MRI protocol using deep learning-based reconstruction: diagnostic equivalence to a conventional protocol.

OBJECTIVE: To evaluate the diagnostic equivalency between an ultrafast (1 min 53 s) lumbar MRI protocol using deep learning-based reconstruction and a conventional lumbar MRI protocol (12 min 31 s). MATERIALS AND METHODS: This study included 58 patients who underwent lumbar MRI using both conventional and ultrafast protocols, including sagittal T1-weighted, T2-weighted, short-TI inversion recovery, and axial T2-weighted sequences. Compared with the conventional protocol, the ultrafast protocol shortened the acquisition time to approximately one-sixth. To compensate for the decreased signal-to-noise ratio caused by the acceleration, deep learning-based reconstruction was applied. Three neuroradiologists graded degenerative changes and analyzed for presence of other pathologies. For the grading of degenerative changes, interprotocol intrareader agreement was assessed using kappa statics. Interchangeability between the two protocols was also tested by calculating the individual equivalence index between the intraprotocol interreader agreement and interprotocol interreader agreement. For the detection of other pathologies, interprotocol intrareader agreement was assessed. RESULTS: For the grading of degenerative changes, the kappa values for interprotocol intrareader agreement of all three readers ranged from 0.707 to 0.804, indicating substantial to almost perfect agreement. Except for foraminal stenosis and disc contour on axial images, the 95% confidence interval of the individual equivalence index was < 5%, indicating the two protocols were interchangeable. For the detection of other pathologies, the interprotocol intrareader agreement rates were > 98% for each individual pathology. CONCLUSIONS: Our proposed ultrafast lumbar spine MRI protocol provided almost equivalent diagnostic results to that of the conventional protocol, except for some degenerative changes.

Spontaneous closure of non-cavernous sinus dural arteriovenous fistulas: A case series and systematic review of the literature.

BACKGROUND AND PURPOSE: To report 9 new cases of non-cavernous sinus dural arteriovenous fistulas (NCS-DAVFs) that closed spontaneously and systematically review reports of other cases in the literature. MATERIAL AND METHODS: We performed a retrospective analysis of 9 cases from 2 institutions of NCS-DAVFs that closed spontaneously. Using PubMed and Scopus in accordance with the PRISMA guidelines, we systematically reviewed English language articles about NCS-DAVFs showing spontaneous closure. RESULTS: Review of the cases from 2 institutions identified 9 cases of NCS-DAVFs showing spontaneous closure in follow-up magnetic resonance angiography (MRA), and the systematic review of the literature yielded an additional 38 cases, which had been diagnosed by repeated arteriography. Collectively, the patients included 23 men and 24 women with a mean age of 54 years. The shunts were located in the transverse-sigmoid sinus in 24 cases (51%), anterior condylar confluence in 11, and other locations in 12. Based on the venous drainage pattern on arteriography, 27 cases (57%) were classified as low-risk NCS-DAVF (without cortical venous reflux) and 17 were classified as high-risk NCS-DAVF (with cortical venous reflux). Shunt closure was observed within 3 months in 17 cases (36%). Extrinsic predisposing factors for shunt closure were detected in 14 cases (30%). These included angiography in 7 cases, sinus recanalization in 4, development of sinus occlusion in 2, and sinus compression by a newly developed hematoma in 1. CONCLUSION: Spontaneous closures of NCS-DAVFs can occur for both high- and low-risk types. One-third of these closures occur within 3 months.

Improvement of the diagnostic accuracy for intracranial haemorrhage using deep learning-based computer-assisted detection.

PURPOSE: To elucidate the effect of deep learning-based computer-assisted detection (CAD) on the performance of different-level physicians in detecting intracranial haemorrhage using CT. METHODS: A total of 40 head CT datasets (normal, 16; haemorrhagic, 24) were evaluated by 15 physicians (5 board-certificated radiologists, 5 radiology residents, and 5 medical interns). The physicians attended 2 reading sessions without and with CAD. All physicians annotated the haemorrhagic regions with a degree of confidence, and the reading time was recorded in each case. Our CAD system was developed using 433 patients' head CT images (normal, 203; haemorrhagic, 230), and haemorrhage rates were displayed as corresponding probability heat maps using U-Net and a machine learning-based false-positive removal method. Sensitivity, specificity, accuracy, and figure of merit (FOM) were calculated based on the annotations and confidence levels. RESULTS: In patient-based evaluation, the mean accuracy of all physicians significantly increased from 83.7 to 89.7% (p < 0.001) after using CAD. Additionally, accuracies of board-certificated radiologists, radiology residents, and interns were 92.5, 82.5, and 76.0% without CAD and 97.5, 90.5, and 81.0% with CAD, respectively. The mean FOM of all physicians increased from 0.78 to 0.82 (p = 0.004) after using CAD. The reading time was significantly lower when CAD (43 s) was used than when it was not (68 s, p < 0.001) for all physicians. CONCLUSION: The CAD system developed using deep learning significantly improved the diagnostic performance and reduced the reading time among all physicians in detecting intracranial haemorrhage.

Silent susceptibility-weighted angiography to detect hemorrhagic lesions in the brain: a clinical and phantom study.

PURPOSE: To compare the effectiveness of silent susceptibility-weighted angiography (sSWAN), a new imaging technique with lower acoustic noise, with conventional susceptibility-weighted angiography (cSWAN) in the detection of intracranial hemorrhagic lesions. METHODS: We measured the acoustic and background noise during sSWAN and cSWAN imaging and calculated the contrast-to-noise ratio (CNR) of the phantom consisting of eight chambers with different concentrations of superparamagnetic iron oxide. In the clinical study, we calculated the CNRs of hemorrhagic lesions in 15 patients and evaluated the images for conspicuity and artifact on each sequence and scored them on a 4-point scale. We also evaluated whether hypointense areas observed on sSWAN or cSWAN increased in size from those on T2*-weighted imaging (T2*-WI). RESULTS: Acoustic noise for sSWAN (57.9 ± 0.32 dB [background noise 51.3 dB]) was significantly less than that for cSWAN (89.0 ± 0.22 dB [background noise 50.9 dB]). The CNRs of phantoms for sSWAN were slightly but not significantly lower than those for cSWAN (P = 0.18). The CNRs of hemorrhagic lesions did not show significant differences between sSWAN and cSWAN (P = 0.17). There were no significant differences between sSWAN and cSWAN with respect to the scores for conspicuity, artifact, and change in size of hypointense areas from T2*-WI. CONCLUSION: sSWAN is equivalent to cSWAN with respect to the image quality for the detection of hemorrhagic lesions but has lower acoustic noise.

Comparative study of pulsed-continuous arterial spin labeling and dynamic susceptibility contrast imaging by histogram analysis in evaluation of glial tumors.

PURPOSE: Arterial spin labeling (ASL) is a non-invasive perfusion technique that may be an alternative to dynamic susceptibility contrast magnetic resonance imaging (DSC-MRI) for assessment of brain tumors. To our knowledge, there have been no reports on histogram analysis of ASL. The purpose of this study was to determine whether ASL is comparable with DSC-MRI in terms of differentiating high-grade and low-grade gliomas by evaluating the histogram analysis of cerebral blood flow (CBF) in the entire tumor. METHODS: Thirty-four patients with pathologically proven glioma underwent ASL and DSC-MRI. High-signal areas on contrast-enhanced T1-weighted images or high-intensity areas on fluid-attenuated inversion recovery images were designated as the volumes of interest (VOIs). ASL-CBF, DSC-CBF, and DSC-cerebral blood volume maps were constructed and co-registered to the VOI. Perfusion histogram analyses of the whole VOI and statistical analyses were performed to compare the ASL and DSC images. RESULTS: There was no significant difference in the mean values for any of the histogram metrics in both of the low-grade gliomas (n = 15) and the high-grade gliomas (n = 19). Strong correlations were seen in the 75th percentile, mean, median, and standard deviation values between the ASL and DSC images. The area under the curve values tended to be greater for the DSC images than for the ASL images. CONCLUSIONS: DSC-MRI is superior to ASL for distinguishing high-grade from low-grade glioma. ASL could be an alternative evaluation method when DSC-MRI cannot be used, e.g., in patients with renal failure, those in whom repeated examination is required, and in children.

[Successful treatment with bortezomib in a patient with systemic primary AL amyloidosis manifesting severe gastrointestinal bleeding].

We present a systemic primary AL amyloidosis patient with severe gastrointestinal bleeding for whom treatment with bortezomib and dexamethasone was very effective. An 83-year-old woman was admitted to our hospital suffering from multiple gastric ulcers with bleeding. She had monoclonal protein (IgG, λ) in her serum. Systemic primary AL amyloidosis was diagnosed. She was also suspected of having amyloid deposition in cardiac muscle, as well as her bladder and intestinal mucosa. However, we were unable to carry out other invasive examinations due to her bleeding tendency. She was treated with bortezomib and dexamethasone. Her bleeding tendency gradually diminished. After one year, the bleeding tendency had completely resolved. However, she still had amyloid deposition in her gastric mucosa and skin.

Local Area Tactile Stimulation Using Interference of Multi-Frequency Airborne Ultrasound.

In spatiotemporal modulation (STM) and lateral modulation (LM) used in conventional mid-air ultrasound tactile stimulation, single or multiple focuses are moved by switching the ultrasound transducer phases. A problem with the phase switching method is the limitation of the focus motion speed due to rapid phase switching that causes sound pressure fluctuations. This paper proposes an LM method using multiple-frequency ultrasound to shift the ultrasound focal point without switching the phase. This method can demonstrate a continuous and stable moving stimulus with high-frequency components, without producing unnecessary audible noise. Using the proposed broadband LM covering up to 400 Hz, we found that a high-frequency 400 Hz LM applied at a finger pad can display a stimulation area with the diameters comparable to or less than the half wavelength of 40 kHz ultrasound, where the perceptual size was evaluated as 4. 2 mm for the long axis diameter and 3. 4 mm for the short axis diameter.

UltLever: Ultrasound-Driven Passive Haptic Actuator Based on Amplifying Radiation Force Using a Simple Lever Mechanism.

A lightweight haptic display that does not interfere with the user's natural movement is required for an immersive haptic experience. This study proposes a lightweight, powerful, and responsive passive haptic actuator driven by airborne focused ultrasound. This 6.2 g completely plastic passive device amplifies an applied ultrasound radiation force by a factor of 35 using a simple lever mechanism, presenting an amplified force of 0.7 N to the user's finger pad. 2-30 Hz vibration can also be presented. Since the radiation force is presented at the speed of sound, the amplified force is presented at high speed even with the high amplification rate of a lever, achieving such strong force and vibration presentation. Physical measurements showed that the amplified force was 0.7 N for the 20.48 mN input radiation force, and the amplitude of the presented vibration was over 0.1 N at 2-30 Hz. A psychophysical experiment showed that the vibration and force were perceivable with a device output level of -7.7 dB. In the future, we will explore methodologies around device design to present desired tactile sensations.

Comparison of natural language processing algorithms in assessing the importance of head computed tomography reports written in Japanese.

PURPOSE: To propose a five-point scale for radiology report importance called Report Importance Category (RIC) and to compare the performance of natural language processing (NLP) algorithms in assessing RIC using head computed tomography (CT) reports written in Japanese. MATERIALS AND METHODS: 3728 Japanese head CT reports performed at Osaka University Hospital in 2020 were included. RIC (category 0: no findings, category 1: minor findings, category 2: routine follow-up, category 3: careful follow-up, and category 4: examination or therapy) was established based not only on patient severity but also on the novelty of the information. The manual assessment of RIC for the reports was performed under the consensus of two out of four neuroradiologists. The performance of four NLP models for classifying RIC was compared using fivefold cross-validation: logistic regression, bidirectional long-short-term memory (BiLSTM), general bidirectional encoder representations of transformers (general BERT), and domain-specific BERT (BERT for medical domain). RESULTS: The proportion of each RIC in the whole data set was 15.0%, 26.7%, 44.2%, 7.7%, and 6.4%, respectively. Domain-specific BERT showed the highest accuracy (0.8434 ± 0.0063) in assessing RIC and significantly higher AUC in categories 1 (0.9813 ± 0.0011), 2 (0.9492 ± 0.0045), 3 (0.9637 ± 0.0050), and 4 (0.9548 ± 0.0074) than the other models (p < .05). Analysis using layer-integrated gradients showed that the domain-specific BERT model could detect important words, such as disease names in reports. CONCLUSIONS: Domain-specific BERT has superiority over the other models in assessing our newly proposed criteria called RIC of head CT radiology reports. The accumulation of similar and further studies of has a potential to contribute to medical safety by preventing missed important findings by clinicians.

Noncontact Haptic Rendering of Static Contact With Convex Surface Using Circular Movement of Ultrasound Focus on a Finger Pad.

A noncontact tactile stimulus can be presented by focusing airborne ultrasound on the human skin. Focused ultrasound has recently been reported to produce not only vibration but also static pressure sensation on the palm by modulating the sound pressure distribution at a low frequency. This finding expands the potential for tactile rendering in ultrasound haptics as static pressure sensation is perceived with a high spatial resolution. In this study, we verified that focused ultrasound can render a static pressure sensation associated with contact with a small convex surface on a finger pad. This static contact rendering enables noncontact tactile reproduction of a fine uneven surface using ultrasound. In the experiments, four ultrasound foci were simultaneously and circularly rotated on a finger pad at 5 Hz. When the orbit radius was 3 mm, vibration and focal movements were barely perceptible, and the stimulus was perceived as static pressure. Moreover, under the condition, the pressure sensation rendered a contact with a small convex surface with a radius of 2 mm. The perceived intensity of the static contact sensation was equivalent to a physical contact force of 0.24 N on average, 10.9 times the radiation force physically applied to the skin.

Sufficient Time-Frequency Resolution for Reproducing Vibrotactile Sensation.

In vibrotactile stimuli, it is essential to reproduce realistic tactile sensations to enhance the immersiveness of applications. To reproduce more realistic tactile experiences, various tools have been proposed to fine-tune and design vibrotactile sensations. Considering the situation where users adjust parameters manually, providing tactile sensations with fewer parameters is desirable. This study examines the coarsest resolution in the time and frequency dimensions necessary to present tactile sensations as realistic as vibrations recorded by the sensor. Time and frequency are fundamental parameters to express vibrations as a spectrogram, and we considered it important to investigate how much coarser the resolution could be without changing perception. We focus on the textural vibrations and the preliminary experiment compared actual texture vibrations with the reconstructed vibration as coarse as possible in the frequency dimension. The result showed that the frequency resolution above 172 Hz makes it difficult to distinguish between the vibrations. The main experiment, a similar discrimination experiment, verified the time resolution using the averaging filter of vibration intensity over time. The results indicate that with the update interval set to 30 ms, the discrimination rate compared to the original vibration is approximately 60%. This percentage is below the chance level of 75%, indicating that distinguishing between the two is difficult. Based on our experiments, it is necessary to have a frequency resolution of at least 172 Hz and a time resolution that updates intensity at a rate of 30 fps or higher to recreate tactile sensations comparable to actual vibrations.

Sound Pressure Field Reconstruction for Airborne Ultrasound Tactile Display Encountering Obstacles.

Airborne ultrasound tactile display (AUTD) is used to provide non-contact tactile sensations with specific foci sound fields through the optimization of transducer phases. However, most existing optimization approaches are not directly applicable in case of an inhomogeneous medium, such as in the presence of obstacles between the AUTD and objective sound field. Certain methods can perform optimizations by considering the sound-scattering surfaces of the obstacles to compute the transmission matrix, which requires several complex measurements. This study proposed two methods to reconstruct the sound field under an inhomogeneous medium, wherein the need to calculate the impact of the obstacles was eliminated. The two methods are Bayesian optimization and greedy algorithm with brute-force search. Further, the process of the foci field generation was assumed as a black box. The proposed methods require only the pressure intensity at the control point generated by the input phases, discarding the need for transmission matrix in the presence of obstacles. Moreover, these methods offer the advantage of optimization of the phases in the presence of obstacles. This study explains the working of proposed methods in different forms of foci fields encountering obstacles.

Non-Vibratory Pressure Sensation Produced by Ultrasound Focus Moving Laterally and Repetitively With Fine Spatial Step Width.

Focused airborne ultrasound provides various noncontact spatiotemporal pressure patterns on the skin. However, the presentation of static force remains an untouched issue because the static radiation force by ultrasound is too weak for the human hand to perceive. Hence, creatable sensations have been limited to vibrations or some dynamically changing feelings. This study demonstrates that a non-vibratory pressure sensation is presented by low-frequency Lateral Modulation (LM) with a fine spatial step width. LM is a pressure modulation method that moves a single ultrasound focus laterally and repetitively along the skin surface. The produced sensation in this study was not perfectly static, but the vibratory perception contained in the stimulus was significantly suppressed under a condition while maintaining its intense perception. We found the condition was 5 to 15 Hz in the LM frequency with a motion step width of less than 1 mm. In a comparison test in the most vibration-suppressed case, the participants reported 0.21 N as an equivalent force to the LM stimulus, significantly higher than the 0.027 N force physically applied by the ultrasound. The statistical analysis also showed that the step width of the LM had a significant effect on its vibratory sensation but not on the intensity of the evoked pressure sensation.

Remote Friction Reduction on Polystyrene Foam Surface by Focused Airborne Ultrasound.

In recent years, various tactile displays having the ability to change their surface friction have been proposed. These displays can express many types of textures and shapes that the materials used for them do not possess. In our study, we found that the ultrasound converged on the surface of polystyrene foam reduces the surface friction. This method has potential applications in disposable and three-dimensional tactile displays. In this study, physical and psychophysical experiments were conducted to verify the effectiveness of the proposed method and to examine the basic conditions under which it is perceived. As a result, we confirmed that the surface friction was reduced on the polystyrene foam, which may be due to the squeeze film effect caused by the external ultrasound excitation of the surface.

Ultrafast cervcial spine MRI protocol using deep learning-based reconstruction: Diagnostic equivalence to a conventional protocol.

PURPOSE: A major drawback of magnetic resonance imaging (MRI) is its limited imaging speed. This study proposed an ultrafast cervical spine MRI protocol (2 min 57 s) using deep learning-based reconstruction (DLR) and compared the diagnostic results to those of conventional MRI protocols (12 min 54 s). METHODS: Fifty patients who underwent cervical spine MRI using both conventional and ultrafast protocols, including sagittal T1-weighted, T2-weighted, short-TI inversion recovery, and axial T2*-weighted imaging were included in this study. The ultrafast protocol shortened the acquisition time to approximately-one-fourth of that of the conventional protocol by reducing the phase matrix, oversampling rate, and number of excitations, and by applying compressed sensing. To compensate for the decreased signal-to-noise ratio caused by acceleration, noise reduction using DLR was performed. For image interpretation, three neuroradiologists graded or classified degenerative changes, including central canal stenosis, foraminal stenosis, endplate degeneration, disc degeneration, and disc hernia. The presence of other pathologies was also recorded. Given the absence of a reference standard, we tested the interchangeability of the two protocols by calculating the 95% confidence interval (CI) of the individual equivalence index. We also assessed the inter-protocol intra-reader agreement using kappa statistics. RESULTS: Except for endplate degeneration, the 95 % CI of the individual equivalence index for all variables did not exceed 5 %, indicating interchangeability between the two protocols. The kappa values ranged from 0.600 to 0.977, indicating substantial to almost perfect agreement. CONCLUSIONS: The proposed ultrafast MRI protocol yielded almost equivalent diagnostic results compared as the conventional protocol.

Formation of peelable rough gold patterns on an ionic liquid template.

The ability to control metal patterns at the micro- and nanoscales, along with the development of a simple fabrication method, is important to many applications in the fields of materials science, biological sensing, electronics, and photonics. Herein, a simple approach to fabricating gold micropatterns with controlled roughness is reported. In this approach, gold is evaporated onto a striped liquid micropattern formed on self-organized microwrinkles. Gold nanoribbons with higher roughness form on the liquid part of the substrate because the deposited gold atoms can diffuse, grow, and aggregate at the liquid-air interface, whereas flat gold films form on the solid part. The rough gold nanoribbons formed on the liquid can then be peeled off through contact with water. The extinction spectrum of the rough gold nanoribbons suggests characteristic surface-plasmon absorption. This shows the possibility of using rough gold nanoribbons with controlled shape in plasmonic technology.

Radiation Pressure Field Reconstruction for Ultrasound Midair Haptics by Greedy Algorithm With Brute-Force Search.

Non-contact tactile presentation using ultrasound phased arrays is becoming a powerful method for providing haptic feedback on bare skin without restricting the user's movement. In such ultrasonic mid-air haptics, it is often necessary to generate multiple ultrasonic foci simultaneously, which requires solving the inverse problem of amplitudes and phases of the transducers in a phased array. Conventionally, matrix calculation methods have been used to solve this inverse problem. However, a matrix calculation requires a non-negligible amount of time when the number of control points and the number of transducers in the array are large. In this article, we propose a simple method based on a greedy algorithm and brute-force search to solve the field reconstruction problem. The proposed method directly optimizes the desired field without matrix calculation or target field phase optimization. The empirical results indicate that the proposed method can reproduce the target sound with an accuracy of more than 80%.

Remote Friction Reduction on Resonant Film Surface by Airborne Ultrasound.

We propose a film device that can be attached to flat surfaces, including touch panels, to remotely reduce surface friction by irradiating airborne ultrasound. In this article, we present a film-air resonance structure that produces large-amplitude surface vibrations excited by airborne ultrasound. We confirmed via simulation that the surface amplitude increases to a level sufficient to reduce friction at the designed frequency. It was also observed in an experiment using a prototype that the friction between a finger and film surface is sharply reduced, and the surface vibrates with sufficient amplitude when touched with a finger.

AUTD3: Scalable Airborne Ultrasound Tactile Display.

Through nonlinear effects, airborne ultrasound phased arrays enable mid-air tactile presentations, as well as auditory presentation and acoustic levitation. To create workplaces flexibly, we have developed a scalable phased array system in which multiple modules can be connected via Ethernet cables and controlled from a PC or other host device. Each module has 249 transducers and the software used can individually specify the phase and amplitude of each of the connected transducers. Using EtherCAT for communication, the system achieves high accuracy synchronization among the connected modules. In this article, we describe the details of the hardware and software architecture of the developed system and evaluate it. We experimentally confirmed the synchronization of 20 modules within an accuracy of 0.1 µs and the phase and amplitude can be specified at 8 bits resolution. In addition, using nine modules, we confirmed that we could make a focal point of the size consistent with the theory at 500 mm above the array surface.

Improved visualization of the chorda tympani nerve using ultra-high-resolution computed tomography.

BACKGROUND: Recognition of the anatomical course of the chorda tympani nerve (CTN) is important for preventing iatrogenic injuries during middle-ear surgery. PURPOSE: This study aims to compare visualization of the CTN using two computed tomography (CT) methods: conventional high-resolution CT (C-HRCT) and ultra-high-resolution CT (U-HRCT). MATERIALS AND METHODS: We performed a retrospective visual assessment of 59 CTNs in normal temporal bones of 54 consecutive patients who underwent both C-HRCT and U-HRCT. After dividing CTN into three anatomical segments (posterior canaliculus, tympanic segment, and anterior canaliculus), two neuroradiologists scored the visualizations on a four-point scale. RESULTS: On C-HRCT, the visual scores of the posterior canaliculus, tympanic segment, and anterior canaliculus were 3.5 ± 0.7, 1.6 ± 0.6, and 3.1 ± 0.7, respectively. The respective values were significantly higher in all segments on U-HRCT: 3.9 ± 0.2, 2.4 ± 0.6, 3.5 ± 0.6 (p < 0.01). Although the difference in scores between methods was greatest for the tympanic segment, the visual score on U-HRCT was lower for the tympanic segment than for the anterior and posterior segments (p < 0.01). CONCLUSION: Ultra-high-resolution CT provides superior visualization of the CTN, especially the tympanic segment.