Objective: We developed a few-shot learning (FSL) framework for the diagnosis of osteopenia and osteoporosis in knee X-ray images.
Methods: Computer vision models containing deep convolutional neural networks were fine-tuned to enable generalization from natural images (ImageNet) to chest X-ray images (normal vs. pneumonia, base images). Then, a series of automated machine learning classifiers based on the Euclidean distances of base images were developed to make predictions for novel images (normal vs. osteopenia vs. osteoporosis). The performance of the FSL framework was compared with that of junior and senior radiologists. In addition, the gradient-weighted class activation mapping algorithm was used for visual interpretation.
Results: In Cohort #1, the mean accuracy (0.728) and sensitivity (0.774) of the FSL models were higher than those of the radiologists (0.512 and 0.448). A diagnostic pipeline of FSL model (first)-radiologists (second) achieved better performance (0.653 accuracy, 0.582 sensitivity, and 0.816 specificity) than radiologists alone. In Cohort #2, the diagnostic pipeline also showed improved performance.
Conclusions: The FSL framework yielded practical performance with respect to the diagnosis of osteopenia and osteoporosis in comparison with radiologists. This retrospective study supports the use of promising FSL methods in computer-aided diagnosis tasks involving limited samples.
Wernicke encephalopathy (WE) is an acute life-threatening neurological condition caused by thiamine (vitamin B1) deficiency. Patients with WE often present with a triad of symptoms consisting of ophthalmoplegia, gait ataxia, and mental confusion. If WE is not treated in a timely manner, it can lead to serious complications such as confusion, coma, or death. Although alcohol abuse is the most commonly reported cause of WE, nonalcoholic causes-although rare-do exist. Herein, we present the case of a nonalcoholic woman with medullary infarctions who presented with intractable vomiting. Her clinical state subsequently progressed to include ophthalmoplegia and gait ataxia. A diagnosis of WE was suspected based on her clinical presentation; this was confirmed by brain magnetic resonance imaging (MRI) and the finding of decreased serum thiamine levels. Brain magnetic resonance imaging demonstrated the complete resolution of abnormal hyperintensities during a follow-up visit, 6 months after treatment.
Objectives: We aimed to evaluate the association of interleukin-6 (IL-6) expression levels with stroke.
Methods: According to the set search strategy, we systematically screened relevant studies using PubMed and extracted study results regarding IL-6 from the literature for comprehensive quantitative analysis to explore the relationship between IL-6 level and stroke risk.
Results: This study included 15 publications with a total of 1696 participants, with 975 cases in the case group and 721 cases in the control group. Meta-analysis showed that IL-6 levels were significantly higher in the stroke population than those in the control group (standardized mean difference = 1.22, 95% confidence interval = 0.79-1.64). Subgroup analysis showed that there was no significant difference in heterogeneity for IL-6 detection methods between the two groups (I2 = 0, P = 0.47). The difference in heterogeneity test results regarding geographic region was statistically significant (I2 = 89.7%, P < 0.01). The results of heterogeneity testing for mean participant age were also statistically significant (I2 = 84.3%, P = 0.01).
Conclusion: The present study results showed that IL-6 may be significantly associated with stroke development.