Lightweight deep neural networks for cholelithiasis and cholecystitis detection by point-of-care ultrasound.

BACKGROUND AND OBJECTIVE: Emergency physicians (EPs) frequently deal with abdominal pain, including that is caused by either gallstones or acute cholecystitis. Easy access and low cost justify point-of-care ultrasound (POCUS) use as a first-line test to detect these diseases; yet, the detection performance of POCUS by EPs is unreliable, causing misdiagnoses with serious impacts. This study aimed to develop a machine learning system to detect and localize gallstones and to detect acute cholecystitis by ultrasound (US) still images taken by physicians or technicians for preliminary diagnoses. METHODS: Abdominal US images (> 89,000) were collected from 2386 patients in a hospital database. We constructed training sets for gallstones with or without cholecystitis (N = 10,971) and cholecystitis with or without gallstones (N = 7348) as positives. Validation sets were also constructed for gallstones (N = 2664) and cholecystitis (N = 1919). We applied a single-shot multibox detector (SSD) and a feature pyramid network (FPN) to classify and localize objects using image features extracted by ResNet-50 for gallstones, and MobileNet V2 to classify cholecystitis. The deep learning models were pretrained using the COCO-2017 and ILSVRC-2012 datasets. RESULTS: Using the validation sets, the SSD-FPN-ResNet-50 and MobileNet V2 achieved areas under the receiver operating characteristics curve of 0.92 and 0.94, respectively. The inference speeds were 21 (47.6 frames per second, fps) and 7 ms (142.9 fps). CONCLUSIONS: A machine learning system was developed to detect and localize gallstones, and to detect cholecystitis, with acceptable discrimination and speed. This is the first study to develop this system for either gallstone or cholecystitis detection with absence or presence of each one. After clinical trials, this system may be used to assist EPs, including those in remote areas, for detecting these diseases.

Deep learning for abdominal ultrasound: A computer-aided diagnostic system for the severity of fatty liver.

BACKGROUND: The prevalence of nonalcoholic fatty liver disease is increasing over time worldwide, with similar trends to those of diabetes and obesity. A liver biopsy, the gold standard of diagnosis, is not favored due to its invasiveness. Meanwhile, noninvasive evaluation methods of fatty liver are still either very expensive or demonstrate poor diagnostic performances, thus, limiting their applications. We developed neural network-based models to assess fatty liver and classify the severity using B-mode ultrasound (US) images. METHODS: We followed standards for reporting of diagnostic accuracy guidelines to report this study. In this retrospective study, we utilized B-mode US images from a consecutive series of patients to develop four-class, two-class, and three-class diagnostic prediction models. The images were eligible if confirmed by at least two gastroenterologists. We compared pretrained convolutional neural network models, consisting of visual geometry group (VGG)19, ResNet-50 v2, MobileNet v2, Xception, and Inception v2. For validation, we utilized 20% of the dataset resulting in >100 images for each severity category. RESULTS: There were 21,855 images from 2,070 patients classified as normal (N = 11,307), mild (N = 4,467), moderate (N = 3,155), or severe steatosis (N = 2,926). We used ResNet-50 v2 for the final model as the best ones. The areas under the receiver operating characteristic curves were 0.974 (mild steatosis vs others), 0.971 (moderate steatosis vs others), 0.981 (severe steatosis vs others), 0.985 (any severity vs normal), and 0.996 (moderate-to-severe steatosis/clinically abnormal vs normal-to-mild steatosis/clinically normal). CONCLUSION: Our deep learning models achieved comparable predictive performances to the most accurate, yet expensive, noninvasive diagnostic methods for fatty liver. Because of the discriminative ability, including for mild steatosis, significant impacts on clinical applications for fatty liver are expected. However, we need to overcome machine-dependent variation, motion artifacts, lacking of second confirmation from any other tools, and hospital-dependent regional bias.

Neutralization of five subgenotypes of Enterovirus 71 by Taiwanese human plasma and Taiwanese plasma derived intravenous immunoglobulin.

Enterovirus 71 (EV71) commonly occurs in children, causing hand, foot and mouth disease (HFMD) in about 29% of patients. Studies have suggested that patients develop meningitis and encephalopathy with a mortality rate of 4-26%. EV71 subgenotypes including B4, B5, C2, C4 and C5 have caused HFMD epidemics in Taiwan. In terms of therapeutical strategy, intravenous immunoglobulin (IVIG) has been shown to improve patient conditions. In this study, the EV71 neutralizing titer was evaluated in 75 human plasmas and 8 lots of Taiwanese plasma derived IVIG. Results showed that human plasmas and IVIG significantly neutralized B4 and C2 subgenotypes. Four percent of human plasma contained neutralizing antibody titer of 1:128 against B4 and C2. Most IVIG lots possessed a median effective dose of over 100 against B4 and C2. IVIG lots had an average neutralizing capacity of 5.60, 0.90, 4.30, 1.12 and 0.77 log10 CCID50/ml against B4, B5, C2, C4 and C5, respectively. In conclusion, effective neutralization of B4 and C2 could be due to their earlier appearance in the EV71 epidemiology timeline of Taiwan. IVIG derived from Taiwanese plasma may be desirable for treatment of patients infected with EV71 of specific subgenotypes.

Trajectory of depression symptoms and related factors in later life--a population based study.

BACKGROUND: A number of studies of elderly depression have examined the association between depression and its related factors; however, they did not observe how the development of related factors influenced elderly depression over time. This study aims to understand the rising trajectory of depressive symptoms found in elderly male and female, and whether the causation of depression due to related factors would be heterogeneous between male and female over a decade from 1993 to 2003. METHOD: This study uses the data from the "Survey of Health and Living Status of the Elderly in Taiwan". A total of 1017 elderly Taiwanese, aged 60 and over, completed four survey modules. The latent growth curve model and general growth mixture model were used to identify the trajectories of depression symptoms, and to assess their related factors in the elderly male and female in Taiwan during the 10 year survey study. RESULT: The results showed that a rising trajectory of depressive symptoms came about for both Taiwan elderly male and female over ten years. We also found that in the trajectories of depressive symptoms between elderly male and female across four waves of data, the results were heterogeneous. The trajectory of depressive symptoms in males was affected by perceived health and disability; the trajectory of depressive symptoms in females was influenced by disability and social support. CONCLUSION: These findings demonstrated that the final depressive symptoms of the elderly would differ from the elderly with developmentally different depressive related factors. The findings also highlighted the developmental trajectories of depressive symptom in terms of their related factors in the elderly.

Simple mobile single-sided NMR apparatus with a relatively homogeneous B0 distribution.

This work presents a simple design for a mobile single-sided nuclear magnetic resonance (NMR) apparatus with a relatively homogeneous static magnetic field (B(0)) distribution. In the proposed design, the B(0) magnetic field of the apparatus is synthesized using only two permanent magnet blocks, i.e., a cube (main) magnet and a small shim magnet placed above the main magnet. The magnetic flux of the shim magnet partially cancels out that of the main magnet, subsequently creating a smooth B(0) profile above the shim magnet where low-resolution NMR experiments are performed. Compared with many previously published designs, this straightforward design simplifies the construction of the apparatus and simultaneously generates a B(0) field parallel to the apparatus surface, allowing the use of a simple loop-type radiofrequency (RF) coil. Additionally, an apparatus prototype is constructed according to the proposed design. Weighing only 1.8 kg, the constructed apparatus has a compact structure and can be held in the palm of a hand. The apparatus generates a B(0) strength of about 0.0746 T. Within a B(0) field deviation of 3 mT, the region with a relatively homogeneous B(0) distribution extends to about 11 mm above the shim magnet. The proposed apparatus can detect a clear Hahn echo or Carr-Purcell-Meiboom-Gill (CPMG) echoes of a pencil eraser block or a bottle of oil placed on the apparatus in 5 s with signal averaging using an RF transmitter power of only 19 W; the detection range of the apparatus exceeds 6 mm. The strength of the residual static magnetic field gradient of the apparatus is roughly estimated at 0.58 T/m. Applying different CPMG echo spacings in this residual static gradient leads to various transverse relaxation time (T(2)) contrasts for liquids with distinct viscosities such as water and oil. Two nondestructive inspection applications of the apparatus, including correlating the concentrations of magnetic nanoparticle solutions with their measured transverse relaxation rates (R(2)) and monitoring the outgassing from an opened bottle of oxygen-supersaturated water by measuring its longitudinal relaxation rate (R(1)), are also demonstrated.

Single-sided mobile NMR apparatus using the transverse flux of a single permanent magnet.

This study presents a simple design for a mobile, single-sided nuclear magnetic resonance (NMR) apparatus which uses the magnetic flux parallel to the magnetization direction of a single, disc-shaped permanent magnet polarized in radial direction. The stray magnetic field above the magnet is approximately parallel to the magnetization direction of the magnet and is utilized as the B(0) magnetic field of the apparatus. The apparatus weighs 1.8 kg, has a compact structure and can be held in one's palm. The apparatus generates a B(0) field strength of about 0.279 T at the center of apparatus surface and can acquire a clear Hahn echo signal of a pencil eraser block lying on the RF coil in one shot. Moreover, a strong static magnetic field gradient exists in the direction perpendicular to the apparatus surface. The strength of the static magnetic field gradient near the center of the apparatus surface is about 10.2 T/m; one-dimensional imaging of thin objects and liquid self-diffusion coefficient measurements can be performed therein. The available spatial resolution of the one-dimensional imaging experiments using a 5 x 5 mm horizontal sample area is about 200 mum. Several nondestructive inspection applications of the apparatus, including distinguishing between polyethylene grains of different densities, characterizing epoxy putties of distinct set times and evaluating the fat content percentages of milk powders, are also demonstrated. Compared with many previously published designs, the proposed design bears a simple structure and generates a B(0) magnetic field parallel to the apparatus surface, simplifying apparatus construction and simultaneously rendering the selection of the radiofrequency coil relatively flexible.

Single-sided mobile NMR with a Halbach magnet.

A single-sided mobile NMR apparatus with a small Halbach magnet was constructed for the first time. It is lightweight, compact and exhibits good sensitivity. The weight of the device is only 2 kg, and the NMR signal of the pencil eraser block can be detected in one shot using the device. This study describes the characteristics of this instrument, including the profile of static magnetic flux density, B0, the sensitivity in the depth direction and its effectiveness in one-dimensional profiling. Its usefulness in differentiating soft materials and evaluating the extent of damage of a material is demonstrated based on T2 relaxation data. The moisture absorbance also can be observed from the increase of the echo amplitude of the NMR spin echo signal.

Implementation of High-Temperature Superconducting tapes RF coils for 3T MRI system.

One way to reduce receiving coil noise in MRI scans is using non-resistive high-temperature superconducting (HTS) coils [1]. They show advantages of much lower cost and easier fabrication over HTS thin iilm coils. In this work, we built a 200mm in diameter Bi2Sr2Ca2Cu3Ox(Bi-2223) tape HTS RF coil and demonstrated that the SNR of using the HTS tape coil was 2.22 folds higher than that of the traditional copper coil for a phantom MR study. Test results were in agreement with predictions, and the error of predicted SNR gains and measured SNR gains is about 0.9%. The HTS coil can be expected to generate higher SNR gain after optimization. In the future, in-vivo experiments will be conducted to farther test the capability of the HTS tape coil. Further applications functional MRI is under investigation to test the power of this HTSC system in our 3T system.

Non-invasive Fiber Tracking on Diffusion Tensor MRI Using High-Temperature Superconducting Tape RF coil.

High-temperature superconducting (HTS) coil is one of the best ways to increase the signal-to-noise ratio (SNR). Bi2Sr2Ca2Cu3Ox (Bi-2223) tapes were suitable to use because of the easier fabrications and lower cost. In this study, we built HTS Bi-2223 tape coils and demonstrated that the SNR of using the HTS tape coil was 3 or 4 folds higher than that of the traditional copper coil for a rat brain MR study. Acquisition time of MR diffusion tensor imaging (DTI) can be reduced by factor of 9 for the same signal-to-noise. Accuracy of fiber tracking using DTI is also significantly improved by a factor of 2.5 or so using HTS coil. In summary, with this HTSC system, a 3T MR system could reach the high signal-to-noise of 12 T MR system with the advantage of less T2 shortening effects at high field. Currents researches are focused on brain connectivity and fMRI studies.