Prediction of the composition of urinary stones using deep learning.

PURPOSE: This study aimed to predict the composition of urolithiasis using deep learning from urinary stone images. MATERIALS AND METHODS: We classified 1,332 stones into 31 classes according to the stone composition. The top 4 classes with a frequency of 110 or more (class 1: calcium oxalate monohydrate [COM] 100%, class 2: COM 80%+struvite 20%, class 3: COM 60%+calcium oxalate dihydrate [COD] 40%, class 4: uric acid 100%) were selected. With the 965 stone images of the top 4 classes, we used the seven convolutional neural networks (CNN) to classify urinary stones and compared their classification performances. RESULTS: Among the seven models, Xception_Ir0.001 showed the highest accuracy, precision, and recall and was selected as the CNN model to predict the stone composition. The sensitivity and specificity for the 4 classes by Xception_Ir0.001 were as follows: class 1 (94.24%, 91.73%), class 2 (85.42%, 96.14%), class 3 (86.86%, 99.59%), and class 4 (94.96%, 98.82%). The sensitivity and specificity of the individual components of the stones were as follows. COM (98.82%, 94.96%), COD (86.86%, 99.64%), struvite (85.42%, 95.59%), and uric acid (94.96%, 98.82%). The area under the curves for class 1, 2, 3, and 4 were 0.98, 0.97, 1.00, and 1.00, respectively. CONCLUSIONS: This study showed the feasibility of deep learning for the diagnostic ability to assess urinary stone composition from images. It can be an alternative tool for conventional stone analysis and provide decision support to urologists, improving the effectiveness of diagnosis and treatment.

Use of combined treatment of 3rd-generation cephalosporin, azithromycin and antiviral agents on moderate SARs-CoV-2 patients in South Korea: A retrospective cohort study.

OBJECTIVES: To assess efficacy and safety of the combined treatment of antibiotics (3rd-generation cephalosporin and azithromycin) and antiviral agents (lopinavir/ritonavir or hydroxychloroquine) on moderate COVID-19 patients in South Korea. METHODS: A retrospective cohort study of the 358 laboratory-confirmed SARS-CoV-2 (COVID-19) patients was conducted. 299 patients met inclusion criteria for analysis. Propensity score matching (PSM) and Cox regression method were used to control and adjust for confounding factors. Mild to moderate COVID-19 patients were managed with either CA/LoP (cephalosporin, azithromycin, and lopinavir/ritonavir) (n = 57), CA/HQ (cephalosporin, azithromycin, and hydroxychloroquine) (n = 25) or standard supportive care (n = 217). We analyzed the association between treatment group and standard supportive group in terms of three endpoints: time to symptom resolution, time to viral clearance, and hospital stay duration. Using propensity-score matching analysis, three rounds of propensity-matching analysis were performed to balance baseline characteristics among three cohorts. RESULTS: Kaplan-Meier curves fitted using propensity score-matched data revealed no significant differences on time to symptom resolution, time to viral clearance, hospital stay duration among the three treatment arms (CA/LoP vs Standard, log-rank p-value = 0.2, 0.58, and 0.74 respectively for the three endpoints) (CA/HQ vs Standard, log-rank p-value = 0.46, 0.99, and 0.75 respectively). Similarly, Cox regression analysis on matched cohorts of CA/LoP and standard supportive group showed that hazard ratios of time to symptom resolution (HR: 1.447 [95%-CI: 0.813-2.577]), time to viral clearance(HR: 0.861, [95%-CI: 0.485-1.527]), and hospital stay duration (HR: 0.902, [95%-CI: 0.510-1.595]) were not significant. For CA/HQ and standard supportive group, hazard ratios of the three endpoints all showed no statistical significance (HR: 1.331 [95%-CI:0.631-2.809], 1.005 [95%-CI:0.480-2.105], and 0.887, [95%-CI:0.422-1.862] respectively). No severe adverse event or death was observed in all groups. CONCLUSIONS: Combined treatment of 3rd cephalosporin, azithromycin and either low-dose lopinavir/ritonavir or hydroxychloroquine was not associated with better clinical outcomes in terms of time to symptom resolution, time to viral clearance, and hospital stay duration compared to standard supportive treatment alone. Microbiological evidence should be closely monitored when treating SARS-CoV-2 patients with antibiotics to prevent indiscreet administration of empirical antimicrobial treatments.

Association between clinical risk factors and severity of dysphagia after extubation based on a videofluoroscopic swallowing study.

BACKGROUND/AIMS: This study aimed to evaluate the correlation between clinical risk factors of post-extubation dysphagia (PED) and the severity of impaired pharyngeal swallowing function assessed via videofluoroscopic swallowing studies (VFSSs). METHODS: This study was a retrospective review of medical records. Of 116 patients who were admitted to the intensive care unit and underwent VFSS, 32 who had non-neurologic disorders and experienced prolonged intubation (for more than 48 hours) were diagnosed with PED. The severity of PED was evaluated by using a functional dysphagia scale (FDS) and a penetration aspiration scale (PAS), on the basis of VFSS. RESULTS: The Simplified Acute Physiology Score 3 and total FDS score were positively correlated (r = 0.40, p = 0.02). Intubation duration was positively correlated with total PAS and FDS scores (r = 0.62, p < 0.001; r = 0.65, p < 0.001, respectively). The amounts of residue in the valleculae (RV) and pyriform sinuses (RP) were associated with intubation duration (r = 0.58, p < 0.001; r = 0.57, p < 0.001, respectively). Multivariate regression analysis revealed that intubation duration was significantly associated with the total FDS score, RV and RP subscales of the FDS, and total PAS score. CONCLUSION: The severity of impaired swallowing function, particularly the amount of residue in the pharyngeal recesses assessed via VFSS, was strongly associated with both severity of medical illness and intubation duration. Intubation duration could be a prognostic factor for assessing impaired swallowing function on the basis of VFSS.

Optimal placement of electrodes for treatment of post-stroke dysphagia by neuromuscular electrical stimulation combined with effortful swallowing.

INTRODUCTION: We aimed to determine the optimal placement of electrodes for neuromuscular electrical stimulation (NMES) for post-stroke dysphagia therapy. METHODS: 31 patients with post-stroke dysphagia were randomised to three groups according to NMES electrode placement. In Group A (n = 10), two pairs of electrodes were attached horizontally on the suprahyoid and infrahyoid muscles. In Group B (n = 11), one pair of electrodes was attached horizontally on the suprahyoid muscles while the other was attached vertically on the infrahyoid muscles. In Group C (n = 10), the electrodes were attached vertically, with one pair above the hyoid bone and the other above the cricoid cartilage. All patients received rehabilitation treatment via NMES combined with effortful swallowing training five times weekly for four weeks. The effect of NMES electrode placement was assessed in terms of the Functional Dysphagia Scale (FDS) and Dysphagia Outcome and Severity Scale (DOSS) scores. RESULTS: Group A showed significantly greater improvement than Group B in overall FDS (p = 0.009) and pharyngeal-phase FDS (FDS-P; p = 0.005) scores. Group A also showed significant improvement when compared with Group C in overall FDS (p = 0.001) and FDS-P (p = 0.001) scores. CONCLUSION: Horizontal placement of the NMES electrodes on the suprahyoid and infrahyoid muscles for the treatment of post-stroke dysphagia by NMES combined with effortful swallowing was more effective than the horizontal and vertical placement of electrodes on the suprahyoid and infrahyoid muscles, respectively, and their vertical placement above the hyoid bone and cricoid cartilage.

Validity of the Buttoning Test in Hand Disability Evaluation of Patients With Stroke.

OBJECTIVE: To investigate the relationship between the buttoning test and Jebsen-Taylor Hand Function Test (JTHFT), and to determine the validity of using the buttoning test as a tool to evaluate hand disability in patients with stroke. METHODS: This was a retrospective study of the medical records of 151 ischemic stroke patients affecting the dominant hand. Patients underwent the buttoning test and JTHFT for their affected hand. All patients were divided into three groups depending on how quickly they fastened a button (group A, not completed; group B, slowly completed over 18 seconds; and group C, completed within 18 seconds). RESULTS: The button fastening time was negatively correlated with the total score and subtest scores of the JTHFT. Patients who experienced difficulty during the buttoning test had lower mean scores in the JTHFT (group A, 28.0±23.9; group B, 62.9±21.7; group C, 75.4±13.3; p<0.0001, Jonckheere-Terpstra test). We observed significant differences in JTHFT scores among the three groups (p<0.017, Mann-Whitney U-test), although there were considerable overlaps in JTHFT scores between the groups. Significant differences were also found in the subtest scores of the JTHFT, which include fine hand motor function (writing letters, p=0.009; moving small objects, p=0.003; stacking checkers, p=0.001 between groups B and C), among the three groups. CONCLUSION: Considering its relationship with the JTHFT and validity, the buttoning test can be considered appropriate for evaluation of hand disability in patients with stroke.

Effect of Family Caregiving on Depression in the First 3 Months After Spinal Cord Injury.

OBJECTIVE: To investigate the effect of family caregiving on depression in the first 3 months after spinal cord injury (SCI). METHODS: A retrospective study was carried out on 76 patients diagnosed with an SCI from January 2013 to December 2016 at the Department of Physical Medicine and Rehabilitation of Kyungpook National University Hospital, Korea. Clinical characteristics including age, gender, level of injury, completeness of the injury, time since injury, caregiver information, etiology, and functional data were collected through a retrospective review of medical records. Depression was assessed using the Beck Depression Inventory (BDI). Patients with 14 or more points were classified as depressed and those with scores of 13 or less as non-depressed group. RESULTS: Of the 76 patients, 33 were in the depressed group with an average BDI of 21.27±6.17 and 43 patients included in the non-depressed group with an average BDI of 4.56±4.20. The BDI score of patients cared by unlicensed assistive personnel (UAP) was significantly higher than that of patients cared by their families (p=0.020). Univariate regression analysis showed that motor complete injury (p=0.027), UAP caregiving (p=0.022), and Ambulatory Motor Index (p=0.019) were associated with depression after SCI. Multivariate binary logistic regression analysis showed that motor completeness (p=0.002) and UAP caregiving (p=0.002) were independent risk factors. CONCLUSION: Compared with UAP, family caregivers lowered the prevalence of depression in the first 3 months after SCI.