Precision Errors and Monitoring Time Interval in Pediatric Muscle Imaging and Neuromuscular Performance Assessment.

OBJECTIVES: To determine precision errors and monitoring time intervals in imaged muscle properties and neuromuscular performance, and to explore growth-related factors associated with precision errors in children. METHODS: We included 35 children (mean age 10.5yrs) in the precision study cohort and 40 children (10.7yrs) in the follow-up study cohort. We assessed forearm and lower leg muscle properties (area, density) with peripheral quantitative computed tomography. We measured neuromuscular performance via maximal pushup, grip force, countermovement and standing long jump force, power, and impulse along with long jump length. We calculated precision errors (root-mean-squared coefficient of variation) from the precision cohort and monitoring time intervals using annual changes from the follow-up cohort. We explored associations between precision errors (coefficient of variation) and maturity, time interval (between repeated measures), and anthropometric changes using Spearman's rank correlation (p<0.05). RESULTS: Muscle measures exhibited precision errors of 1.3-14%. Monitoring time intervals were 1-2.6yrs, except muscle density (>43yrs). We identified only one association between precision errors and maturity (maximal pushup force: rho=-0.349; p=0.046). CONCLUSIONS: Imaging muscle properties and neuromuscular performance measures had precision errors of 1-14% and appeared suitable for follow-up on ~2yr scales (except muscle density). Maximal pushup force appeared more repeatable in mature children.

Performance evaluation of a deep learning-based cascaded HRNet model for automatic measurement of X-ray imaging parameters of lumbar sagittal curvature.

PURPOSE: To develop a deep learning-based cascaded HRNet model, in order to automatically measure X-ray imaging parameters of lumbar sagittal curvature and to evaluate its prediction performance. METHODS: A total of 3730 lumbar lateral digital radiography (DR) images were collected from picture archiving and communication system (PACS). Among them, 3150 images were randomly selected as the training dataset and validation dataset, and 580 images as the test dataset. The landmarks of the lumbar curve index (LCI), lumbar lordosis angle (LLA), sacral slope (SS), lumbar lordosis index (LLI), and the posterior edge tangent angle of the vertebral body (PTA) were identified and marked. The measured results of landmarks on the test dataset were compared with the mean values of manual measurement as the reference standard. Percentage of correct key-points (PCK), intra-class correlation coefficient (ICC), Pearson correlation coefficient (r), mean absolute error (MAE), mean square error (MSE), root-mean-square error (RMSE), and Bland-Altman plot were used to evaluate the performance of the cascade HRNet model. RESULTS: The PCK of the cascaded HRNet model was 97.9-100% in the 3 mm distance threshold. The mean differences between the reference standard and the predicted values for LCI, LLA, SS, LLI, and PTA were 0.43 mm, 0.99°, 1.11°, 0.01 mm, and 0.23°, respectively. There were strong correlation and consistency of the five parameters between the cascaded HRNet model and manual measurements (ICC = 0.989-0.999, R = 0.991-0.999, MAE = 0.63-1.65, MSE = 0.61-4.06, RMSE = 0.78-2.01). CONCLUSION: The cascaded HRNet model based on deep learning algorithm could accurately identify the sagittal curvature-related landmarks on lateral lumbar DR images and automatically measure the relevant parameters, which is of great significance in clinical application.

Noise-robustness test for ultrasound breast nodule neural network models as medical devices.

Background: Deep learning technology has been widely applied to medical image analysis. But due to the limitations of its own imaging principle, ultrasound image has the disadvantages of low resolution and high Speckle Noise density, which not only hinder the diagnosis of patients' conditions but also affect the extraction of ultrasound image features by computer technology. Objective: In this study, we investigate the robustness of deep convolutional neural network (CNN) for classification, segmentation, and target detection of breast ultrasound image through random Salt & Pepper Noise and Gaussian Noise. Methods: We trained and validated 9 CNN architectures in 8617 breast ultrasound images, but tested the models with noisy test set. Then, we trained and validated 9 CNN architectures with different levels of noise in these breast ultrasound images, and tested the models with noisy test set. Diseases of each breast ultrasound image in our dataset were annotated and voted by three sonographers based on their malignancy suspiciousness. we use evaluation indexes to evaluate the robustness of the neural network algorithm respectively. Results: There is a moderate to high impact (The accuracy of the model decreased by about 5%-40%) on model accuracy when Salt and Pepper Noise, Speckle Noise, or Gaussian Noise is introduced to the images respectively. Consequently, DenseNet, UNet++ and Yolov5 were selected as the most robust model based on the selected index. When any two of these three kinds of noise are introduced into the image at the same time, the accuracy of the model will be greatly affected. Conclusions: Our experimental results reveal new insights: The variation trend of accuracy with the noise level in Each network used for classification tasks and object detection tasks has some unique characteristics. This finding provides us with a method to reveal the black-box architecture of computer-aided diagnosis (CAD) systems. On the other hand, the purpose of this study is to explore the impact of adding noise directly to the image on the performance of neural networks, which is different from the existing articles on robustness in the field of medical image processing. Consequently, it provides a new way to evaluate the robustness of CAD systems in the future.

Bone deficits in children and youth with type 1 diabetes: A systematic review and meta-analysis.

Deficits in bone mineral and weaker bone structure in children with type 1 diabetes (T1D) may contribute to a lifelong risk of fracture. However, there is no meta-analysis comparing bone properties beyond density between children with T1D and typically developing children (TDC). This meta-analysis aimed to assess differences and related factors in bone mineral content (BMC), density, area, micro-architecture and estimated strength between children with T1D and TDC. We systematically searched MEDLINE, Embase, CINAHL, Web of Science, Scopus, Cochrane Library databases, and included 36 in the meta-analysis (2222 children and youth with T1D, 2316 TDC; mean age ≤18 yrs., range 1-24). We estimated standardized mean differences (SMD) using random-effects models and explored the role of age, body size, sex ratio, disease duration, hemoglobin A1c in relation to BMC and areal density (aBMD) SMD using meta-regressions. Children and youth with T1D had lower total body BMC (SMD: -0.21, 95% CI: -0.37 to -0.05), aBMD (-0.30, -0.50 to -0.11); lumbar spine BMC (-0.17, -0.28 to -0.06), aBMD (-0.20, -0.32 to -0.08), bone mineral apparent density (-0.30, -0.48 to -0.13); femoral neck aBMD (-0.21, -0.33 to -0.09); distal radius and tibia trabecular density (-0.38, -0.64 to -0.12 and -0.35, -0.51 to -0.18, respectively) and bone volume fraction (-0.33, -0.56 to -0.09 and -0.37, -0.60 to -0.14, respectively); distal tibia trabecular thickness (-0.41, -0.67 to -0.16); and tibia shaft cortical content (-0.33, -0.56 to -0.10). Advanced age was associated with larger SMD in total body BMC (-0.13, -0.21 to -0.04) and aBMD (-0.09; -0.17 to -0.01) and longer disease duration with larger SMD in total body aBMD (-0.14; -0.24 to -0.04). Children and youth with T1D have lower BMC, aBMD and deficits in trabecular density and micro-architecture. Deficits in BMC and aBMD appeared to increase with age and disease duration. Bone deficits may contribute to fracture risk and require attention in diabetes research and care. STUDY REGISTRATION: PROSPERO (CRD42020200819).

Higher Body Fat in Children and Adolescents With Type 1 Diabetes-A Systematic Review and Meta-Analysis.

Aims: Higher prevalence of overweight and obesity in children and adolescents with type 1 diabetes (T1D) suggests alterations are required in body composition. However, differences in body composition between children with T1D and typically developing children (TDC) have not been synthesized using meta-analysis. Therefore, we conducted a systematic review and meta-analysis to compare body composition between children with T1D and TDC, and to explore the role of disease and non-disease related factors in potential body composition differences. Methods: Studies were performed comparing dual-energy x-ray absorptiometry-acquired total body fat and lean mass, absolute (kg) and relative (%) values, between children with T1D and TDC. We reported mean differences with 95% confidence intervals (CI) from meta-analysis and relative between-group %-differences. We used meta-regression to explore the role of sex, age, height, body mass, body mass index, Hemoglobin A1c, age of onset, disease duration, and insulin dosage in the potential body composition differences between children with T1D and TDC, and subgroup analysis to explore the role of geographic regions (p < 0.05). Results: We included 24 studies (1,017 children with T1D, 1,045 TDC) in the meta-analysis. Children with T1D had 1.2 kg more fat mass (kg) (95%CI 0.3 to 2.1; %-difference = 9.3%), 2.3% higher body fat % (0.3-4.4; 9.0%), but not in lean mass outcomes. Age of onset (ß = -2.3, -3.5 to -1.0) and insulin dosage (18.0, 3.5-32.6) were negatively and positively associated with body fat % mean difference, respectively. Subgroup analysis suggested differences among geographic regions in body fat % (p < 0.05), with greater differences in body fat % from Europe and the Middle East. Conclusion: This meta-analysis indicated 9% higher body fat in children with T1D. Earlier diabetes onset and higher daily insulin dosage were associated with body fat % difference between children with T1D and TDC. Children with T1D from Europe and the Middle East may be more likely to have higher body fat %. More attention in diabetes research and care toward body composition in children with T1D is needed to prevent the early development of higher body fat, and to minimize the cardiovascular disease risk and skeletal deficits associated with higher body fat.

Acupuncture for Cancer-Induced Bone Pain in Animal Models: A Systemic Review and Meta-Analysis.

BACKGROUND: Cancer-induced bone pain (CIBP) is a highly prevalent symptom, which afflicts vast majority of patients who suffer from cancer. The current treatment options failed to achieve satisfactory effect and the side effects were prominent. Recent randomized controlled trials (RCTs) of animal demonstrate the benefit of acupuncture for CIBP. We sought to determine if the pooled data from available RCTs supports the use of acupuncture for CIBP. METHODS: A literature search for randomized controlled trials was conducted in six electronic databases from inception to May 31, 2019. Meta-analysis was performed with Review Manager 5.3 software; the publication bias was assessed by Stata 12.0 software. We used random effects model for pooling data because heterogeneity is absolute among studies to some extent. RESULTS: Twenty-four trials were included in the review, of which 12 trials provided detailed data for meta-analyses. Preliminary evidence indicates that compared to wait list/sham group, acupuncture was effective on increasing paw withdrawal threshold (PWT) and paw withdrawal latency (PWL). Compared to medicine, acupuncture was less effective on PWT, but as effective as medicine on PWL. Acupuncture can reinforce medicine's effect on PWT and PWL. Compared to the control group, acupuncture was superior to increase body weight (BW), decrease spinal cord glial fibrillary acidic protein (GFAP), and interleukin-1ß (IL-1ß). Furthermore, some studies showed acupuncture delay or partially reverse morphine tolerance. Three studies found acupuncture has no effect on PWT, but 2 of them found acupuncture could enhance small dose of Celebrex's effect on CIBP. CONCLUSIONS: Acupuncture was superior to wait list/sham acupuncture on increasing PWT and has no less effect on increasing PWL compared to medicine; acupuncture improved the efficacy of drugs, increased the CIBP animals' body weight, and decreased their spinal cord GFAP and IL-1ß. High-quality studies are necessary to confirm the results.