Using Transfer Learning of Convolutional Neural Network on Neck Radiographs to Identify Acute Epiglottitis.

Acute epiglottitis (AE) is a life-threatening condition and needs to be recognized timely. Diagnosis of AE with a lateral neck radiograph yields poor reliability and sensitivity. Convolutional neural networks (CNN) are powerful tools to assist the analysis of medical images. This study aimed to develop an artificial intelligence model using CNN-based transfer learning to identify AE in lateral neck radiographs. All cases in this study are from two hospitals, a medical center, and a local teaching hospital in Taiwan. In this retrospective study, we collected 251 lateral neck radiographs of patients with AE and 936 individuals without AE. Neck radiographs obtained from patients without and with AE were used as the input for model transfer learning in a pre-trained CNN including Inception V3, Densenet201, Resnet101, VGG19, and Inception V2 to select the optimal model. We used five-fold cross-validation to estimate the performance of the selected model. The confusion matrix of the final model was analyzed. We found that Inception V3 yielded the best results as the optimal model among all pre-train models. Based on the average value of the fivefold cross-validation, the confusion metrics were obtained: accuracy = 0.92, precision = 0.94, recall = 0.90, and area under the curve (AUC) = 0.96. Using the Inception V3-based model can provide an excellent performance to identify AE based on radiographic images. We suggest using the CNN-based model which can offer a non-invasive, accurate, and fast diagnostic method for AE in the future.

Effectiveness of electrical stimulation therapy in improving arm function after stroke: a systematic review and a meta-analysis of randomised controlled trials.

OBJECTIVE: The aim of this study is to investigate the effectiveness of electrical stimulation in arm function recovery after stroke. METHODS: Data were obtained from the PubMed, Cochrane Library, Embase, and Scopus databases from their inception until 12 January 2019. Only randomized controlled trials (RCTs) reporting the effects of electrical stimulation on the recovery of arm function after stroke were selected. RESULTS: Forty-eight RCTs with a total of 1712 patients were included in the analysis. The body function assessment, Upper-Extremity Fugl-Meyer Assessment, indicated more favorable outcomes in the electrical stimulation group than in the placebo group immediately after treatment (23 RCTs (n = 794): standard mean difference (SMD) = 0.67, 95% confidence interval (CI) = 0.51-0.84) and at follow-up (12 RCTs (n = 391): SMD = 0.66, 95% CI = 0.35-0.97). The activity assessment, Action Research Arm Test, revealed superior outcomes in the electrical stimulation group than those in the placebo group immediately after treatment (10 RCTs (n = 411): SMD = 0.70, 95% CI = 0.39-1.02) and at follow-up (8 RCTs (n = 289): SMD = 0.93, 95% CI = 0.34-1.52). Other activity assessments, including Wolf Motor Function Test, Box and Block Test, and Motor Activity Log, also revealed superior outcomes in the electrical stimulation group than those in the placebo group. Comparisons between three types of electrical stimulation (sensory, cyclic, and electromyography-triggered electrical stimulation) groups revealed no significant differences in the body function and activity. CONCLUSION: Electrical stimulation therapy can effectively improve the arm function in stroke patients.

Intermittent Cervical Traction for Treating Neck Pain: A Meta-analysis of Randomized Controlled Trials.

STUDY DESIGN: A meta-analysis. OBJECTIVE: The aim of this study was to perform a comprehensive search of current literature and conduct a meta-analysis of randomized controlled trials (RCTs) to assess the neck pain relieving effect of intermittent cervical traction (ICT). SUMMARY OF BACKGROUND DATA: Neck pain is a common and disabling problem with a high prevalence in general population. It causes a considerable burden on the health care system with a substantial expenditure. ICT is a common component of physical therapy for neck pain in the outpatient clinic. However, the evidence regarding the effectiveness of ICT for neck pain is insufficient. METHODS: Data were obtained from the PubMed, Cochrane Library, Embase, and Scopus databases from the database inception date to July 02, 2016. RCTs reporting the effects of ICT on neck pain, including those comparing the effects of ICT with those of a placebo treatment, were included. Two reviewers independently reviewed the studies, conducted a risk of bias assessment, and extracted data. The data were pooled in a meta-analysis by using a random-effects model. RESULTS: The meta-analysis included seven RCTs. The results indicated that patients who received ICT for neck pain had significantly lower pain scores than those receiving placebos did immediately after treatment (standardized mean difference = -0.26, 95% confidence interval = -0.46 to -0.07). The pain scores during the follow-up period and the neck disability index scores immediately after treatment and during the follow-up period did not differ significantly. CONCLUSION: ICT may have a short-term neck pain-relieving effect. Some risks of bias were noted in the included studies, reducing the evidence level of this meta-analysis. Additional high-quality RCTs are required to clarify the long-term effects of ICT on neck pain. LEVEL OF EVIDENCE: 1.