Numerical Approach to Facial Palsy Using a Novel Registration Method with 3D Facial Landmark.

Treatment of facial palsy is essential because neglecting this disorder can lead to serious sequelae and further damage. For an objective evaluation and consistent rehabilitation training program of facial palsy patients, a clinician's evaluation must be simultaneously performed alongside quantitative evaluation. Recent research has evaluated facial palsy using 68 facial landmarks as features. However, facial palsy has numerous features, whereas existing studies use relatively few landmarks; moreover, they do not confirm the degree of improvement in the patient. In addition, as the face of a normal person is not perfectly symmetrical, it must be compared with previous images taken at a different time. Therefore, we introduce three methods to numerically approach measuring the degree of facial palsy after extracting 478 3D facial landmarks from 2D RGB images taken at different times. The proposed numerical approach performs registration to compare the same facial palsy patients at different times. We scale landmarks by performing scale matching before global registration. After scale matching, coarse registration is performed with global registration. Point-to-plane ICP is performed using the transformation matrix obtained from global registration as the initial matrix. After registration, the distance symmetry, angular symmetry, and amount of landmark movement are calculated for the left and right sides of the face. The degree of facial palsy at a certain point in time can be approached numerically and can be compared with the degree of palsy at other times. For the same facial expressions, the degree of facial palsy at different times can be measured through distance and angle symmetry. For different facial expressions, the simultaneous degree of facial palsy in the left and right sides can be compared through the amount of landmark movement. Through experiments, the proposed method was tested using the facial palsy patient database at different times. The experiments involved clinicians and confirmed that using the proposed numerical approach can help assess the progression of facial palsy.

A Transparent Nanopatterned Chemiresistor: Visible-Light Plasmonic Sensor for Trace-Level NO2 Detection at Room Temperature.

The highly selective detection of trace gases using transparent sensors at room temperature remains challenging. Herein, transparent nanopatterned chemiresistors composed of aligned 1D Au-SnO2 nanofibers, which can detect toxic NO2 gas at room temperature under visible light illumination is reported. Ten straight Au-SnO2 nanofibers are patterned on a glass substrate with transparent electrodes assisted by direct-write, near-field electrospinning, whose extremely low coverage of sensing materials (≈0.3%) lead to the high transparency (≈93%) of the sensor. The sensor exhibits a highly selective, sensitive, and reproducible response to sub-ppm levels of NO2 , and its detection limit is as low as 6 ppb. The unique room-temperature NO2 sensing under visible light emanates from the localized surface plasmonic resonance effect of Au nanoparticles, thereby enabling the design of new transparent oxide-based gas sensors without external heaters or light sources. The patterning of nanofibers with extremely low coverage provides a general strategy to design diverse compositions of gas sensors, which can facilitate the development of a wide range of new applications in transparent electronics and smart windows wirelessly connected to the Internet of Things.

TIMA SLAM: Tracking Independently and Mapping Altogether for an Uncalibrated Multi-Camera System.

We present a novel simultaneous localization and mapping (SLAM) system that extends the state-of-the-art ORB-SLAM2 for multi-camera usage without precalibration. In this system, each camera is tracked independently on a shared map, and the extrinsic parameters of each camera in the fixed multi-camera system are estimated online up to a scalar ambiguity (for RGB cameras). Thus, the laborious precalibration of extrinsic parameters between cameras becomes needless. By optimizing the map, the keyframe poses, and the relative poses of the multi-camera system simultaneously, observations from the multiple cameras are utilized robustly, and the accuracy of the shared map is improved. The system is not only compatible with RGB sensors but also works on RGB-D cameras. For RGB cameras, the performance of the system tested on the well-known EuRoC/ASL and KITTI datasets that are in the stereo configuration for indoor and outdoor environments, respectively, as well as our dataset that consists of three cameras with small overlapping regions. For the RGB-D tests, we created a dataset that consists of two cameras for an indoor environment. The experimental results showed that the proposed method successfully provides an accurate multi-camera SLAM system without precalibration of the multi-cameras.

Clinical efficacy of upper limb robotic therapy in people with tetraplegia: a pilot randomized controlled trial.

STUDY DESIGN: A pilot randomized controlled trial. OBJECTIVES: To evaluate the clinical efficacy of upper limb robotic therapy in people with tetraplegia. SETTING: Inpatient rehabilitation hospital in Seoul, Korea. METHODS: Participants were randomly allocated to a robotic therapy (RT) or occupational therapy (OT) group. Both groups received usual care plus 30 min of additional therapy per day for 4 weeks. The additional therapy provided to the OT group was OT, and the additional therapy provided to RT group was RT using the Armeo Power. Primary outcomes were the Medical Research Council scale of each key muscle and Upper Extremity Motor Score (UEMS) for the trained arm. Secondary outcomes were the Spinal Cord Independence Measurement version III (SCIM-III) subscale and total score. Evaluations were performed at baseline and 4 weeks. RESULTS: A total of 34 individuals with tetraplegia were included; 17 in each group. At 4 weeks, the median (IQR) change in UEMS in the RT group was 1/25 (0 to 3) points compared with 0/25 (-1 to 1) points in the OT group (p = 0.03). The median (IQR) change in total SCIM-III score in the RT group was 7/100 (1.5 to 11) points compared with 0/100 (-8 to 4) points in the OT group (p < 0.01). CONCLUSIONS: There were small improvements in motor strength and SCIM-III scores in the RT group, but there were no statistically significant differences between the groups. Further studies are required for a better understanding of the effects of RT for people with tetraplegia.

Correction: Clinical efficacy of upper limb robotic therapy in people with tetraplegia: a pilot randomized controlled trial.

In the original version of the article the authors incorrectly stated that: "One case study provided evidence of some improvements in motor performance and spasticity [19], while several other studies only provided evidence on the feasibility of UER as an assessment tool; however, the fact that the manufacturer funded these studies lessens their objectivity [15, 20-24]." This is not correct as the manufacturers did not fund the studies. The correct phrase therefore should have read: "One case study provided evidence of some improvements in motor performance and spasticity [19], while several other studies provided evidence focusing more on the feasibility of UER rather than the clinical efficacy [15, 20-24]." The authors would like to apologise for this error.This has been corrected in both the PDF and HTML versions of the Article.

Structure-From-Motion in 3D Space Using 2D Lidars.

This paper presents a novel structure-from-motion methodology using 2D lidars (Light Detection And Ranging). In 3D space, 2D lidars do not provide sufficient information for pose estimation. For this reason, additional sensors have been used along with the lidar measurement. In this paper, we use a sensor system that consists of only 2D lidars, without any additional sensors. We propose a new method of estimating both the 6D pose of the system and the surrounding 3D structures. We compute the pose of the system using line segments of scan data and their corresponding planes. After discarding the outliers, both the pose and the 3D structures are refined via nonlinear optimization. Experiments with both synthetic and real data show the accuracy and robustness of the proposed method.

A New Strategy for Humidity Independent Oxide Chemiresistors: Dynamic Self-Refreshing of In2 O3 Sensing Surface Assisted by Layer-by-Layer Coated CeO2 Nanoclusters.

The humidity dependence of the gas sensing characteristics of metal oxide semiconductors has been one of the greatest obstacles for gas sensor applications during the last five decades because ambient humidity dynamically changes with the environmental conditions. Herein, a new and novel strategy is reported to eliminate the humidity dependence of the gas sensing characteristics of oxide chemiresistors via dynamic self-refreshing of the sensing surface affected by water vapor chemisorption. The sensor resistance and gas response of pure In2 O3 hollow spheres significantly change and deteriorate in humid atmospheres. In contrast, the humidity dependence becomes negligible when an optimal concentration of CeO2 nanoclusters is uniformly loaded onto In2 O3 hollow spheres via layer-by-layer (LBL) assembly. Moreover, In2 O3 sensors LBL-coated with CeO2 nanoclusters show fast response/recovery, low detection limit (500 ppb), and high selectivity to acetone even in highly humid conditions (relative humidity 80%). The mechanism underlying the dynamic refreshing of the In2 O3 sensing surfaces regardless of humidity variation is investigated in relation to the role of CeO2 and the chemical interaction among CeO2 , In2 O3 , and water vapor. This strategy can be widely used to design high performance gas sensors including disease diagnosis via breath analysis and pollutant monitoring.

BNNT vs. CNT: influence of structural defects on damping characteristics of nanocomposites.

The viscoelastic damping of nanocomposites reinforced with BNNTs and CNTs was compared. MD simulations revealed that the interfacial damping of pristine-CNT was superior to that of pristine-BNNT. The contrasting effects of structural defects were elucidated using interfacial adhesion, interphase, and overlapping phonon density of states in the nanotubes and polymers.

Dysregulation of histone deacetylases in ocular diseases.

Ocular diseases are a growing global concern and have a significant impact on the quality of life. Cataracts, glaucoma, age-related macular degeneration, and diabetic retinopathy are the most prevalent ocular diseases. Their prevalence and the global market size are also increasing. However, the available pharmacotherapy is currently limited. These diseases share common pathophysiological features, including neovascularization, inflammation, and/or neurodegeneration. Histone deacetylases (HDACs) are a class of enzymes that catalyze the removal of acetyl groups from lysine residues of histone and nonhistone proteins. HDACs are crucial for regulating various cellular processes, such as gene expression, protein stability, localization, and function. They have also been studied in various research fields, including cancer, inflammatory diseases, neurological disorders, and vascular diseases. Our study aimed to investigate the relationship between HDACs and ocular diseases, to identify a new strategy for pharmacotherapy. This review article explores the role of HDACs in ocular diseases, specifically focusing on diabetic retinopathy, age-related macular degeneration, and retinopathy of prematurity, as well as optic nerve disorders, such as glaucoma and optic neuropathy. Additionally, we explore the interplay between HDACs and key regulators of fibrosis and angiogenesis, such as TGF-ß and VEGF, highlighting the potential of targeting HDAC as novel therapeutic strategies for ocular diseases.

Large litters have a detrimental impact on litter performance and postpartum maternal behaviour in primiparous sows.

BACKGROUND: Our previous study confirmed that large litter size adversely affects prepartum maternal hormones and behaviour, concurrently with heightened oxidative stress in primiparous sows. The purpose of this study was to examine the effect of large litter size on litter performance, postpartum maternal behaviour, salivary cortisol levels, and colostral immunoglobulin levels in sows, as well as investigate their correlations with the levels of oxidative stress parameters. RESULTS: A total of 24 primiparous sows (Landrace[Formula: see text]Large white) and their offspring were categorised into two groups based on litter size: NORMAL (n = 8) with litter size ranging from 7 to 14 (mean 11.5[Formula: see text]2.7), and LARGE (n=16) with litter size ranging from 15 to 20 (mean 15.9[Formula: see text]1.4). All sows were housed in a group housing system during gestation and transitioned to an adaptable loose housing system (2.4[Formula: see text]2.3 m) during the farrowing and lactation periods. The nursing and carefulness behaviour of the sows was monitored over a 24-h period between 72 and 96 h after parturition. Saliva samples were collected for cortisol assay on 35, 21, and 7 days before parturition (D-35, D-21, and D-7, respectively), as well as on days 1, 7, and 28 after parturition (D1, D7, and D28, respectively). On D1, higher piglet mortality rates were observed among the LARGE group compared to the NORMAL group (p<0.01). The total and successful nursing behaviours of the sows were less frequent in the LARGE group than in the NORMAL group (p<0.05, for both), and the carefulness score of the LARGE group was also lower than that of the NORMAL group (p< 0.01). On D1, cortisol levels in LARGE sows were higher than those in NORMAL sows (p< 0.05), and for other time points (D-21, D-7, D7, and D28), cortisol levels in LARGE sows tended to be higher than those in NORMAL sows (p < 0.10, for all). Successful nursing behaviour displayed negative correlations with levels of salivary cortisol and certain oxidative stress parameters measured on D1. CONCLUSIONS: These findings suggest that the strategy for alleviating physiological and oxidative stress during the peripartum periods could benefit potential postpartum maternal behaviour and litter performance in the sows with large litters.

Effects of Nesting Material Provision and High-Dose Vitamin C Supplementation during the Peripartum Period on Prepartum Nest-Building Behavior, Farrowing Process, Oxidative Stress Status, Cortisol Levels, and Preovulatory Follicle Development in Hyperprolific Sows.

Hyperprolific sows often experience increased oxidative stress during late gestation and lactation periods, which can adversely affect the farrowing process and overall lactation performance. This study examines the influence of providing a coconut coir mat (CCM; 1 × 1 m) as nesting material, supplementing high-dose vit-C (HVC; 20% vit-C, 10 g/kg feed) as an antioxidant, or both on maternal behavior, the farrowing process, oxidative status, cortisol levels, and preovulatory follicle developments in sows with large litters. In total, 35 sows (Landrace × Yorkshire; litter size 15.43 ± 0.27) were allocated to the following four treatment groups: control (n = 9, basal diet), vit-C (n = 8, basal diet + HVC), mat (n = 10, basal diet + CCM), and mat + vit-C (n = 8, basal diet + HVC + CCM). A post-hoc analysis showed that compared with sows that were not provided CCM, mat and mat + vit-C groups demonstrated increased durations of nest-building behavior during the period from 24 h to 12 h before parturition (p < 0.05 for both), reduced farrowing durations, and decreased intervals from birth to first udder contact (p < 0.01 for both). The mat group exhibited lower advanced oxidation protein product (AOPP) levels during late gestation and lactation periods than the control group (p < 0.05). Sows with HVC supplementation showed longer farrowing durations than those without HVC supplementation (p < 0.0001). The vit-C group had higher salivary cortisol levels on day 1 after farrowing than the other treatment groups (p < 0.05). Furthermore, the follicle diameters on day 3 after weaning in the vit-C group tended to be smaller than those in the control group (p = 0.077). HVC supplementation prolonged farrowing and increased the physiological stress on postpartum, and no advantageous effects on maternal behavior and developmental progression of preovulatory follicles were observed. Hence, alternative solutions beyond nutritional approaches are required to address increased oxidative stress in hyperprolific sows and secure their welfare and reproductive performance. The present results substantiated the positive impact of providing CCM as nesting material for sows with large litters on nest-building behavior and the farrowing process, which could mitigate the deleterious consequences induced by peripartum physiological and oxidative stress.

Extracellular Vesicles Derived from Adipose Stem Cells Alleviate Systemic Sclerosis by Inhibiting TGF-ß Pathway.

Systemic sclerosis is an autoimmune disease characterized by inflammatory reactions and fibrosis. Myofibroblasts are considered therapeutic targets for preventing and reversing the pathogenesis of fibrosis in systemic sclerosis. Although the mechanisms that differentiate into myofibroblasts are diverse, transforming growth factor ß (TGF-ß) is known to be a key mediator of fibrosis in systemic sclerosis. This study investigated the effects of extracellular vesicles derived from human adipose stem cells (ASC-EVs) in an in vivo systemic sclerosis model and in vitro TGF-ß1-induced dermal fibroblasts. The therapeutic effects of ASC-EVs on the in vivo systemic sclerosis model were evaluated based on dermal thickness and the number of α-smooth muscle actin (α-SMA)-expressing cells using hematoxylin and eosin staining and immunohistochemistry. Administration of ASC-EVs decreased both the dermal thickness and α-SMA expressing cell number as well as the mRNA levels of fibrotic genes, such as Acta2, Ccn2, Col1a1 and Comp. Additionally, we discovered that ASC-EVs can decrease the expression of α-SMA and CTGF and suppress the TGF-ß pathway by inhibiting the activation of SMAD2 in dermal fibroblasts induced by TGF-ß1. Finally, TGF-ß1-induced dermal fibroblasts underwent selective death through ASC-EVs treatment. These results indicate that ASC-EVs could provide a therapeutic approach for preventing and reversing systemic sclerosis.

Accuracy of posteroanterior cephalogram landmarks and measurements identification using a cascaded convolutional neural network algorithm: A multicenter study.

Objective: : To quantify the effects of midline-related landmark identification on midline deviation measurements in posteroanterior (PA) cephalograms using a cascaded convolutional neural network (CNN). Methods: : A total of 2,903 PA cephalogram images obtained from 9 university hospitals were divided into training, internal validation, and test sets (n = 2,150, 376, and 377). As the gold standard, 2 orthodontic professors marked the bilateral landmarks, including the frontozygomatic suture point and latero-orbitale (LO), and the midline landmarks, including the crista galli, anterior nasal spine (ANS), upper dental midpoint (UDM), lower dental midpoint (LDM), and menton (Me). For the test, Examiner-1 and Examiner-2 (3-year and 1-year orthodontic residents) and the Cascaded-CNN models marked the landmarks. After point-to-point errors of landmark identification, the successful detection rate (SDR) and distance and direction of the midline landmark deviation from the midsagittal line (ANS-mid, UDM-mid, LDM-mid, and Me-mid) were measured, and statistical analysis was performed. Results: : The cascaded-CNN algorithm showed a clinically acceptable level of point-to-point error (1.26 mm vs. 1.57 mm in Examiner-1 and 1.75 mm in Examiner-2). The average SDR within the 2 mm range was 83.2%, with high accuracy at the LO (right, 96.9%; left, 97.1%), and UDM (96.9%). The absolute measurement errors were less than 1 mm for ANS-mid, UDM-mid, and LDM-mid compared with the gold standard. Conclusions: : The cascaded-CNN model may be considered an effective tool for the auto-identification of midline landmarks and quantification of midline deviation in PA cephalograms of adult patients, regardless of variations in the image acquisition method.

Weakly Supervised Contrastive Learning for Unsupervised Vehicle Reidentification.

Reidentification (Re-id) of vehicles in a multicamera system is an essential process for traffic control automation. Previously, there have been efforts to reidentify vehicles based on shots of images with identity (id) labels, where the model training relies on the quality and quantity of the labels. However, labeling vehicle ids is a labor-intensive procedure. Instead of relying on expensive labels, we propose to exploit camera and tracklet ids that are automatically obtainable during a Re-id dataset construction. In this article, we present weakly supervised contrastive learning (WSCL) and domain adaptation (DA) techniques using camera and tracklet ids for unsupervised vehicle Re-id. We define each camera id as a subdomain and tracklet id as a label of a vehicle within each subdomain, i.e., weak label in the Re-id scenario. Within each subdomain, contrastive learning using tracklet ids is applied to learn a representation of vehicles. Then, DA is performed to match vehicle ids across the subdomains. We demonstrate the effectiveness of our method for unsupervised vehicle Re-id using various benchmarks. Experimental results show that the proposed method outperforms the recent state-of-the-art unsupervised Re-id methods. The source code is publicly available on https://github.com/andreYoo/WSCL_VeReid.

Clinicopathologic significance of the delta-like ligand 4, vascular endothelial growth factor, and hypoxia-inducible factor-2α in gallbladder cancer.

BACKGROUND: Gallbladder cancer (GBC) is usually detected in advanced stages with a low 5-year survival rate. Delta-like ligand 4 (DLL4), vascular endothelial growth factor (VEGF), and hypoxia-inducible factor-2alpha (HIF2α) have been studied for their role in tumorigenesis and potential for therapeutic target, and multiple clinical trials of the agents targeting them are ongoing. We investigated the expression of these markers in surgically resected GBC and tried to reveal their association with the clinicopathologic features, mutual correlation of their expression, and prognosis of the GBC patients by their expression. METHODS: We constructed the tissue microarray blocks of 99 surgically resected GBC specimens and performed immunohistochemistry of DLL4, VEGF, and HIF2α. We used the quantitative digital image analysis to evaluate DLL4 and VEGF expression, while the expression of HIF2α was scored manually. RESULTS: The expression of VEGF and HIF2α showed a significant trend with tumor differentiation (p= .028 and p= .006, respectively). We found that the high DLL4 and VEGF expression were significantly correlated with lymph node metastasis (p= .047, both). The expression of VEGF and HIF2α were significantly correlated (p < .001). The GBC patients with low HIF2α expression showed shorter recurrence-free survival than those with high HIF2α expression. CONCLUSIONS: This study suggested the possibility of the usage of DLL4 and VEGF to predict the lymph node metastasis and the possibility of VEGF and HIF2α to predict the expression level mutually. Further studies may be needed to validate our study results and eventually accelerate the introduction of the targeted therapy in GBC.

An Iterative Method for Unsupervised Robust Anomaly Detection Under Data Contamination.

Most deep anomaly detection models are based on learning normality from datasets due to the difficulty of defining abnormality by its diverse and inconsistent nature. Therefore, it has been a common practice to learn normality under the assumption that anomalous data are absent in a training dataset, which we call normality assumption. However, in practice, the normality assumption is often violated due to the nature of real data distributions that includes anomalous tails, i.e., a contaminated dataset. Thereby, the gap between the assumption and actual training data affects detrimentally in learning of an anomaly detection model. In this work, we propose a learning framework to reduce this gap and achieve better normality representation. Our key idea is to identify sample-wise normality and utilize it as an importance weight, which is updated iteratively during the training. Our framework is designed to be model-agnostic and hyperparameter insensitive so that it applies to a wide range of existing methods without careful parameter tuning. We apply our framework to three different representative approaches of deep anomaly detection that are classified into one-class classification-, probabilistic model-, and reconstruction-based approaches. In addition, we address the importance of a termination condition for iterative methods and propose a termination criterion inspired by the anomaly detection objective. We validate that our framework improves the robustness of the anomaly detection models under different levels of contamination ratios on five anomaly detection benchmark datasets and two image datasets. On various contaminated datasets, our framework improves the performance of three representative anomaly detection methods, measured by area under the ROC curve.

Artificial intelligence-based numerical analysis of the quality of facial reanimation: A comparative retrospective cohort study between one-stage dual innervation and single innervation.

This study aimed to investigate the difference in facial reanimation surgery using functional gracilis muscle transfer between the masseteric nerve alone and its combined use with cross face nerve graft (CFNG), which has not been explored before. A novel analysis method based on artificial intelligence (AI) was employed to compare the outcomes of the two approaches. Using AI, 3-dimensional facial landmarks were extracted from 2-dimensional photographs, and distance and angular symmetry scores were calculated. The patients were divided into two groups, with Group 1 undergoing one-stage CFNG and masseteric nerve dual innervation, and Group 2 receiving only masseteric nerve. The symmetry scores were obtained before and 1 year after surgery to assess the degree of change. Of the 35 patients, Group 1 included 13 patients, and Group 2 included 22 patients. The analysis revealed that, in the resting state, the change in the symmetry score of the mouth corner showed distance symmetry (2.55 ± 2.94, 0.52 ± 2.75 for Group 1 and Group 2, respectively, p = 0.048) and angle symmetry (1.21 ± 1.43, 0.02 ± 0.22 for Group 1 and Group 2, respectively, p = 0.001), which were significantly improved in Group 1, indicating a more symmetric pattern after surgery. In the smile state, only the angle symmetry was improved more symmetrically in Group 1 (3.20 ± 2.38, 1.49 ± 2.22 for Group 1 and Group 2, respectively, p = 0.041). Within the limitations of the study it seems that this new analysis method enabled a more accurate numerical symmetry score to be obtained, and while the degree of mouth corner excursion was sufficient with only the masseteric nerve, accompanying CFNG led to further improvement in symmetry in the resting state.

Trends in the incidence and prevalence of dysphagia requiring medical attention among adults in South Korea, 2006-2016: A nationwide population study.

The prevalence of dysphagia is increasing, resulting in socioeconomic burden, but previous reports have only been based on a limited populations. Therefore, we aimed to investigate the nationwide incidence and prevalence of dysphagia requiring medical attention to provide adequate information for healthcare planning and resource allocation. In this nationwide retrospective cohort study, the data of adults aged ≥20 years recorded from 2006 to 2016 were sourced from the Korean National Health Insurance Service database. Medical claim codes based on ICD-10-CM were used to define dysphagia and possible causes. The annual incidence and prevalence of dysphagia were calculated. Cox regression was used to estimate dysphagia risk in people with possible dysphagia etiology. Survival analysis was performed to estimate the mortality and hazard ratio of dysphagia. The crude annual incidence of dysphagia increased continuously from 7.14 in 2006 to 15.64 in 2016. The crude annual prevalence of dysphagia in 2006 was 0.09% and increased annually to 0.25% in 2016. Stroke (odds ratio [OR]: 7.86, 95% confidence interval [CI]: 5.76-6.68), neurodegenerative disease (OR: 6.20, 95% CI: 5.76-6.68), cancer (OR: 5.59, 95% CI: 5.17-6.06), and chronic obstructive pulmonary disease (OR: 2.94, 95% CI: 2.71-3.18) were associated with a high risk of dysphagia. The mortality in the dysphagia group was 3.12 times higher than that in the non-dysphagia group (hazard ratio: 3.12, 95% CI: 3.03-3.23). The incidence and prevalence of dysphagia requiring medical attention are increasing annually. The increasing trend was conspicuous in the geriatric population. The presence of stroke, neurodegenerative disease, cancer, and chronic obstructive pulmonary disease is associated with a high risk of dysphagia. Therefore, adequate screening, diagnosis, and management of dysphagia in the older population must be emphasized in geriatric healthcare.

MuSiC-ViT: A multi-task Siamese convolutional vision transformer for differentiating change from no-change in follow-up chest radiographs.

A major responsibility of radiologists in routine clinical practice is to read follow-up chest radiographs (CXRs) to identify changes in a patient's condition. Diagnosing meaningful changes in follow-up CXRs is challenging because radiologists must differentiate disease changes from natural or benign variations. Here, we suggest using a multi-task Siamese convolutional vision transformer (MuSiC-ViT) with an anatomy-matching module (AMM) to mimic the radiologist's cognitive process for differentiating baseline change from no-change. MuSiC-ViT uses the convolutional neural networks (CNNs) meet vision transformers model that combines CNN and transformer architecture. It has three major components: a Siamese network architecture, an AMM, and multi-task learning. Because the input is a pair of CXRs, a Siamese network was adopted for the encoder. The AMM is an attention module that focuses on related regions in the CXR pairs. To mimic a radiologist's cognitive process, MuSiC-ViT was trained using multi-task learning, normal/abnormal and change/no-change classification, and anatomy-matching. Among 406 K CXRs studied, 88 K change and 115 K no-change pairs were acquired for the training dataset. The internal validation dataset consisted of 1,620 pairs. To demonstrate the robustness of MuSiC-ViT, we verified the results with two other validation datasets. MuSiC-ViT respectively achieved accuracies and area under the receiver operating characteristic curves of 0.728 and 0.797 on the internal validation dataset, 0.614 and 0.784 on the first external validation dataset, and 0.745 and 0.858 on a second temporally separated validation dataset. All code is available at https://github.com/chokyungjin/MuSiC-ViT.

Orthognathic surgical planning using graph CNN with dual embedding module: External validations with multi-hospital datasets.

BACKGROUND AND OBJECTIVE: Despite recent development of AI, prediction of the surgical movement in the maxilla and mandible by OGS might be more difficult than that of tooth movement by orthodontic treatment. To evaluate the prediction accuracy of the surgical movement using pairs of pre-(T0) and post-surgical (T1) lateral cephalograms (lat-ceph) of orthognathic surgery (OGS) patients and dual embedding module-graph convolution neural network (DEM-GCNN) model. METHODS: 599 pairs from 3 institutions were used as training, internal validation, and internal test sets and 201 pairs from other 6 institutions were used as external test set. DEM-GCNN model (IEM, learning the lat-ceph images; LTEM, learning the landmarks) was developed to predict the amount and direction of surgical movement of ANS and PNS in the maxilla and B-point and Md1crown in the mandible. The distance between T1 landmark coordinates actually moved by OGS (ground truth) and predicted by DEM-GCNN model and pre-existed CNN-based Model-C (learning the lat-ceph images) was compared. RESULTS: In both internal and external tests, DEM-GCNN did not exhibit significant difference from ground truth in all landmarks (ANS, PNS, B-point, Md1crown, all P > 0.05). When the accumulated successful detection rate for each landmark was compared, DEM-GCNN showed higher values than Model-C in both the internal and external tests. In violin plots exhibiting the error distribution of the prediction results, both internal and external tests showed that DEM-GCNN had significant performance improvement in PNS, ANS, B-point, Md1crown than Model-C. DEM-GCNN showed significantly lower prediction error values than Model-C (one-jaw surgery, B-point, Md1crown, all P < 0.005; two-jaw surgery, PNS, ANS, all P < 0.05; B point, Md1crown, all P < 0.005). CONCLUSION: We developed a robust OGS planning model with maximized generalizability despite diverse qualities of lat-cephs from 9 institutions.