Visual Scene-Aware Hybrid and Multi-Modal Feature Aggregation for Facial Expression Recognition.

Facial expression recognition (FER) technology has made considerable progress with the rapid development of deep learning. However, conventional FER techniques are mainly designed and trained for videos that are artificially acquired in a limited environment, so they may not operate robustly on videos acquired in a wild environment suffering from varying illuminations and head poses. In order to solve this problem and improve the ultimate performance of FER, this paper proposes a new architecture that extends a state-of-the-art FER scheme and a multi-modal neural network that can effectively fuse image and landmark information. To this end, we propose three methods. To maximize the performance of the recurrent neural network (RNN) in the previous scheme, we first propose a frame substitution module that replaces the latent features of less important frames with those of important frames based on inter-frame correlation. Second, we propose a method for extracting facial landmark features based on the correlation between frames. Third, we propose a new multi-modal fusion method that effectively fuses video and facial landmark information at the feature level. By applying attention based on the characteristics of each modality to the features of the modality, novel fusion is achieved. Experimental results show that the proposed method provides remarkable performance, with 51.4% accuracy for the wild AFEW dataset, 98.5% accuracy for the CK+ dataset and 81.9% accuracy for the MMI dataset, outperforming the state-of-the-art networks.

Measurement of Coracohumeral Distance in 3 Shoulder Positions Using Dynamic Ultrasonography: Correlation With Subscapularis Tear.

PURPOSE: To detect differences between the values of dynamic coracohumeral distance (CHD) measured using ultrasonography (USG) in different shoulder rotations and to investigate its correlation with subscapularis tear. METHODS: We prospectively enrolled consecutive patients (n = 168) who were scheduled to have arthroscopic rotator cuff repair. Patients with a history of previous shoulder surgery or shoulder fracture and patients with external rotation less than 30° were excluded from the study. Dynamic CHD was measured using USG in 3 different shoulder positions: external rotation, neutral and internal rotation. We evaluated the intrarater reliability with 3 times repetition of measurement. Patients were divided into 1 of 3 groups according to arthroscopic findings: intact subscapularis, partial-thickness tear, and full-thickness tear of the subscapularis. The control group (n = 23) included patients without rotator cuff tears from the outpatient clinic. Subgroup analysis-according to the presence of dynamic subcoracoid stenosis, defined as a CHD less than 6 mm measured in internal rotation-was performed to find the clinical effect of dynamic subcoracoid stenosis. RESULTS: A partial-thickness tear of the subscapularis tendon was present in 60 patients (35.7%) and a full-thickness tear in 26 patients (15.4%) among 168 patients. The CHD was maximum in external rotation and the narrowest in internal rotation. There were no statistical differences in the CHDs between groups with different subscapularis tear status. According to the presence of dynamic subcoracoid stenosis, patients with dynamic subcoracoid stenosis had a significantly higher incidence of partial-thickness subscapularis tear than those without stenosis (P = .022). CONCLUSIONS: The CHD values were narrowest in shoulder internal rotation, which is thought to be the pathogenic position. We could not confirm the correlation between CHD and subscapularis tear. However, patients who have dynamic subcoracoid stenosis had significantly higher incidence of subscapularis tear than others without dynamic stenosis. LEVEL OF EVIDENCE: Level II, prospective prognostic cohort study.

Rate of correction after asymmetrical physeal suppression in valgus deformity: analysis using a linear mixed model application.

BACKGROUND: This study was performed to estimate the rate of angular correction after asymmetrical physeal suppression and analyze the factors that influence the rate of correction by using a linear mixed model application. METHODS: A total of 175 physes (72 distal femoral, 70 proximal tibial, and 33 distal tibial) from 78 consecutive patients with valgus angular deformity of the lower limb who underwent asymmetrical physeal suppression were included. The anatomic lateral distal femoral angle, the anatomic lateral proximal tibial angle, and the anatomic lateral distal tibial angle were measured from the teleroentgenograms of the patients' preoperative visit and periodic follow-ups. The rate of angular correction was adjusted by multiple factors by using a linear mixed model with age, sex, and surgical method as the fixed effects and each subject as the random effect. The final model included the age-specific and surgical method-specific rate and sex-specific and surgical method-specific intercept. Multivariate analysis was performed for this model. RESULTS: In younger children (boys 14 y or younger and girls 12 y or younger), the rate of correction of valgus deformity at the distal femur, proximal tibia, and distal tibia was 0.71 degrees/month (8.5 degrees/y), 0.40 degrees/month (4.8 degrees/y), and 0.48 degrees/month (5.8 degrees/y), respectively. In older children, the rate of correction of valgus deformity at the distal femur, proximal tibia, and distal tibia was 0.39 degrees/month (4.7 degrees/y), 0.29 degrees/month (3.5 degrees/y), and 0.48 degrees/month (5.8 degrees/y), respectively. The rate of correction at the distal femur was significantly lower in older children (P = 0.025). The rate of angular correction at the proximal tibia was significantly faster in the screw group than in the staple group (P = 0.046). CONCLUSIONS: Asymmetrical physeal suppression with staples, percutaneous transphyseal screws, and permanent method all are effective methods for treating valgus deformity in growing children. When we treat valgus deformity in growing children, we should take into consideration the fact that the rate of correction at the distal femur is lower in older children, and that at the proximal tibia is faster in the screw group. LEVEL OF EVIDENCE: Therapeutic level III.

Focal Bone Marrow Lesions: A Complication of Ultrasound Diathermy.

Ultrasound diathermy is widely used for the treatment of musculoskeletal disorders and other soft tissue injuries. Its use as a therapeutic modality is believed to be safe, with very few reported complications. Here, we report two patients who developed focal bone marrow abnormalities after receiving ultrasound diathermy. Both patients' magnetic resonance (MR) evaluations revealed linear subchondral bone lesions of the superolateral humeral head similar to those in osteonecrosis. The patients' symptoms subsequently improved, and available follow-up MR evaluation revealed near complete resolution of bone lesions. These findings suggest that ultrasound diathermy, and its interaction with bone tissue through thermal mechanisms, can cause focal bone marrow abnormalities. Furthermore, the bone marrow abnormalities seem to be transient, resolving upon cessation of ultrasound diathermy, therefore osteonecrosis should be differentiated from this temporal lesion.

[Acute Suppurative Appendicitis Diagnosed by Acute Lower Gastrointestinal Hemorrhage].

A 49-year-old man visited the emergency room of Korea University Ansan Hospital with hematochezia starting the day before the visit. Recently, he was on anti-platelet medication due to hypertension. The patient had no definite symptoms other than hematochezia. Digital rectal exam was positive and laboratory tests showed severe anemia. Sigmoidoscopy was initiated and almost no fecal material was observed in the intestinal tract, allowing insertion into the cecum. Active bleeding from the appendiceal opening was noted. On abdominal CT, contrast enhancement was observed at the tip of the appendix. Under suspicion of acute appendicitis, we consulted with a surgeon. The patient underwent appendectomy with partial cecal resection. Pathologic examination revealed a diagnosis of appendix bleeding due to acute suppurative appendicitis. The patient had no further bleeding after surgery and was discharged in a stable state. Careful observation by the endoscopist is necessary for accurate diagnosis of lower gastrointestinal hemorrhage. Appendiceal hemorrhage is very rarely reported, but it has various pathophysiologies. CT scan is useful when appendiceal hemorrhage is confirmed by endoscopic findings. Surgical treatment was needed in almost all cases reported worldwide. If bleeding from the appendix is confirmed, surgical treatment should be considered for both therapeutic and diagnostic purposes.

[The Effect of Helicobacter pylori Infection on Recurrence of Gastric Hyperplastic Polyp after Endoscopic Removal].

Background/Aims: Several previous studies suggest that eradication of Helicobacter pylori (H. pylori) leads to the disappearance of gastric hyperplastic polyps. However, little is known about the effect of H. pylori status and eradication on the recurrence of gastric polyps after endoscopic removal. Here, we investigated the recurrence of gastric polyps according to the final H. pylori status in patients who underwent endoscopic removal of gastric hyperplastic polyps. Methods: Between January 2011 and December 2016, patients who underwent endoscopic removal of gastric hyperplastic polyps and were followed-up for more than two months were enrolled. The success of H. pylori eradication was assessed by histology and rapid urease test or urea breath test, at least 4 weeks after the completion of eradication treatment. At follow-up, the recurrence of gastric polyp was evaluated via esophagogastroduodenoscopy. Results: Seventy-nine patients were enrolled. During the mean follow-up period of 16.4 months, the recurrence rate of gastric polyp was 25.3%. Among those who received H. pylori eradication therapy, the H. pylori persistent group showed a higher recurrence of polyp than the H. pylori eradicated group; but there was no statistical significance (42.9% vs. 21.7%, p=0.269). Regarding the final H. pylori infection status, the recurrence rate of gastric polyps was significantly higher in the H. pylori positive group than in the H. pylori negative group (42.9% vs. 18.9%, p=0.031). In multivariate analysis, the final H. pylori infection status was a significant risk factor for gastric polyp recurrence after endoscopic removal. Conclusions: The final positive H. pylori infection status is significantly associated with higher recurrence of gastric hyperplastic polyps after endoscopic removal.

[A Case of Tenofovir-associated Fanconi Syndrome in Patient with Chronic Hepatitis B].

Tenofovir disoproxil fumarate (TDF) is one of the most widely used treatment options for human immunodeficiency virus (HIV) and HBV infections. Despite its efficacy and safety, some cases of nephrotoxicity have been reported in the treatment of HIV patients. Even more recently, very few cases of Fanconi syndrome associated with tenofovir therapy in HBV monoinfection have been reported. Herein, we report a case of a 47-year-old male with an HBV monoinfection, who developed Fanconi syndrome and a secondary osteomalacia with multiple bone pain. After TDF withdrawal and supplementation of calcitriol, his renal function was reverted. Although the overall risk of TDF-associated nephrotoxicity is very low, both glomerular and tubular function should be monitored in patients undergoing TDF treatment.

Comparison between the radionuclide salivagram and videofluoroscopic swallowing study methods for evaluating patients with aspiration pneumonia.

OBJECTIVE: A videofluoroscopic swallowing study (VFSS) is generally used to assess dysphagia and aspiration, although false-negative results may be obtained. Saliva aspiration may cause false-negative VFSS findings in elderly people. A radionuclide salivagram can be useful in detecting saliva aspiration. We therefore compared these two methods for the evaluation of patients with aspiration pneumonia. METHODS: We prospectively enrolled 50 patients with aspiration pneumonia into this study. All patients underwent VFSS, and were evaluated on the Penetration-Aspiration Scale. A salivagram was performed on the day after VFSS and each patient received 1 mL Tc-99 m DTPA sublingually. Both dynamic and delayed images were obtained. RESULTS: Aspiration was observed in 34% of patients by salivagram and in 42% by VFSS. The frequency of tests that were positive for aspiration pneumonia using a combination of the two methods was 52%. Five patients with positive aspiration findings on salivagram were negative on VFSS. The association between VFSS and salivagram findings was significant but the total agreement was 72%. The frequency of aspiration pneumonia was significantly associated with an abnormal salivagram finding. CONCLUSIONS: A salivagram may be a useful method for the evaluation of aspiration in elderly people, and a combination of the VFSS and salivagram methods could enable a more effective investigation of aspiration pneumonia.

Effect of lumbar stabilization and dynamic lumbar strengthening exercises in patients with chronic low back pain.

OBJECTIVE: To compare the effects of lumbar stabilization exercises and lumbar dynamic strengthening exercises on the maximal isometric strength of the lumbar extensors, pain severity and functional disability in patients with chronic low back pain (LBP). METHODS: Patients suffering nonspecific LBP for more than 3 months were included prospectively and randomized into lumbar stabilization exercise group (n=11) or lumbar dynamic strengthening exercise group (n=10). Exercises were performed for 1 hour, twice weekly, for 8 weeks. The strength of the lumbar extensors was measured at various angles ranging from 0° to 72° at intervals of 12°, using a MedX. The visual analog scale (VAS) and the Oswestry Low Back Pain Disability Questionnaire (ODQ) were used to measure the severity of LBP and functional disability before and after the exercise. RESULTS: Compared with the baseline, lumbar extension strength at all angles improved significantly in both groups after 8 weeks. The improvements were significantly greater in the lumbar stabilization exercise group at 0° and 12° of lumbar flexion. VAS decreased significantly after treatment; however, the changes were not significantly different between the groups. ODQ scores improved significantly in the stabilization exercise group only. CONCLUSION: Both lumbar stabilization and dynamic strengthening exercise strengthened the lumbar extensors and reduced LBP. However, the lumbar stabilization exercise was more effective in lumbar extensor strengthening and functional improvement in patients with nonspecific chronic LBP.

Revisit of Broden's view for intraarticular calcaneal fracture.

BACKGROUND: This study was performed to investigate the relationship between coronal computed tomography (CT) and Broden's view in terms of location of the fracture line and fracture pattern. METHODS: Forty-five feet of 45 patients with intraarticular calcaneal fractures were evaluated. The mean age of the patients was 46.3 years (standard deviation, 18.1; range, 15 to 80 years), and there were 34 men and 11 women. The Broden's views were acquired using the ray sum projection, reviewed, and correlated with the coronal CT image to determine the location of the fracture on the posterior facet and fracture pattern described by the Sanders classification. The quantified location of the fracture line was defined as the distance between the medial margin of posterior facet and the fracture line divided by the whole length of the posterior facet, which was expressed as a percentage. RESULTS: The fracture line on the Broden's view was positioned at 22.3% (standard deviation, 29.6) laterally compared to that on coronal CT (p < 0.01). Although all cases showed posterior facet involvement on the CT scan, the fracture line was positioned lateral to the posterior facet in 6 cases (13.3%) in the Broden's view. The coronal CT and Broden's view showed a low level of agreement in the fracture pattern according to the Sanders classification, with kappa values of 0.23. CONCLUSIONS: Surgeons should consider that the fracture line on the Broden's view shows positioning laterally compared to coronal CT and they should consider that the fracture line at the lateral to posterior facet on the Broden's view might be an intraarticular fracture line. There are some limitations when applying the Sanders classification with the Broden's view.

Inter-rater Reliability of Videofluoroscopic Dysphagia Scale.

OBJECTIVE: To investigate the inter-rater agreement using the Videofluoroscopic Dysphagia Scale (VDS). METHOD: The present study was designed as a multicenter, single-blind trial. A Videofluoroscopic Swallowing Study (VFSS) was performed using the protocol described by J.A Logemann. Thick-fluid, pureed food, mechanically altered food, regularly textured food, and thin-fluid boluses were sequentially swallowed. Each participant received a 3 ml bolus followed by a 5 ml bolus of each food material, in the order mentioned above. All study procedures were video recorded. Discs containing these video recordings in random order were distributed to interpreters who were blinded to the participant information. The video recordings were evaluated using a standardized VDS sheet and the inter-rater reliability was calculated. RESULTS: In total, 100 patients participated in this study and 10 interpreters analyzed the findings. Inter-rater reliability was fair in terms of lip closure (κ: 0.325), oral transit time (0.253), delayed triggering of pharyngeal swallowing (0.300), vallecular residue (0.275), laryngeal elevation (0.345), pyriform sinus residue (0.310), coating of the pharyngeal wall (0.310), and aspiration (0.393). However, other parameters of the oral phase were lower than those of the pharyngeal phase (0.06-0.153). Moreover, the summation of VDS reliability (intraclass correlation coefficient: 0.556) showed moderate agreement. CONCLUSION: VDS shows a moderate rate of agreement for evaluating the swallowing function. However, many of the parameters demonstrated a lower rate of agreement, particularly the oral phase parameters.

The use of a ti-ni shape memory alloy ring bone fixator during the retrograde nailing of supracondylar femoral fractures.

PURPOSE: To identify the effects of using a Ti-Ni shape memory alloy ring shaped bone fixator (SMA-rBF) during the retrograde nailing of supracondylar femoral fractures. MATERIALS AND METHODS: The authors reviewed 25 patients with a supracondylar femoral fracture treated by retrograde intramedullary nailing with or without SMA-rBF (group S, 12/25; group N, 13/25). Radiological measurements of angular deformity were performed and functional assessments were made using the Sanders grading system. RESULTS: All fractures healed after an average of 12.2 weeks (range, 9-15 weeks) in group N and after 11.6 weeks (range, 10-13 weeks) in group S (p=0.351). The mean angle of coronal angular deformity was valgus 0.8° (range, varus 2.3°-valgus 4.5°) in group N and valgus 0.7° (range, varus 1.0°-valgus 2.4°) in group S (p=0.892). The mean angle of sagittal angular deformity was 1.0° in extension (range, flexion 3.2°-extension 3.1°) in group N and 0° (range, flexion 2.1°-extension 1.2°) in group S (p=0.022). However, functional grading evaluations revealed no differences between the two groups. CONCLUSIONS: When reduction of a distal femoral fracture with retrograde nailing was difficult additional mini-open reduction and fixation with a ring shaped SMA did not delay or prevent bony union and resulted in good postoperative alignment.