Information extraction from German radiological reports for general clinical text and language understanding.

Recent advances in deep learning and natural language processing (NLP) have opened many new opportunities for automatic text understanding and text processing in the medical field. This is of great benefit as many clinical downstream tasks rely on information from unstructured clinical documents. However, for low-resource languages like German, the use of modern text processing applications that require a large amount of training data proves to be difficult, as only few data sets are available mainly due to legal restrictions. In this study, we present an information extraction framework that was initially pre-trained on real-world computed tomographic (CT) reports of head examinations, followed by domain adaptive fine-tuning on reports from different imaging examinations. We show that in the pre-training phase, the semantic and contextual meaning of one clinical reporting domain can be captured and effectively transferred to foreign clinical imaging examinations. Moreover, we introduce an active learning approach with an intrinsic strategic sampling method to generate highly informative training data with low human annotation cost. We see that the model performance can be significantly improved by an appropriate selection of the data to be annotated, without the need to train the model on a specific downstream task. With a general annotation scheme that can be used not only in the radiology field but also in a broader clinical setting, we contribute to a more consistent labeling and annotation process that also facilitates the verification and evaluation of language models in the German clinical setting.

Reproducibility of artificial intelligence models in computed tomography of the head: a quantitative analysis.

When developing artificial intelligence (AI) software for applications in radiology, the underlying research must be transferable to other real-world problems. To verify to what degree this is true, we reviewed research on AI algorithms for computed tomography of the head. A systematic review was conducted according to the preferred reporting items for systematic reviews and meta-analyses. We identified 83 articles and analyzed them in terms of transparency of data and code, pre-processing, type of algorithm, architecture, hyperparameter, performance measure, and balancing of dataset in relation to epidemiology. We also classified all articles by their main functionality (classification, detection, segmentation, prediction, triage, image reconstruction, image registration, fusion of imaging modalities). We found that only a minority of authors provided open source code (10.15%, n 0 7), making the replication of results difficult. Convolutional neural networks were predominantly used (32.61%, n = 15), whereas hyperparameters were less frequently reported (32.61%, n = 15). Data sets were mostly from single center sources (84.05%, n = 58), increasing the susceptibility of the models to bias, which increases the error rate of the models. The prevalence of brain lesions in the training (0.49 ± 0.30) and testing (0.45 ± 0.29) datasets differed from real-world epidemiology (0.21 ± 0.28), which may overestimate performances. This review highlights the need for open source code, external validation, and consideration of disease prevalence.

Short-Term Treatment with Alirocumab, Flow-Dependent Dilatation of the Brachial Artery and Use of Magnetic Resonance Diffusion Tensor Imaging to Evaluate Vascular Structure: An Exploratory Pilot Study.

BACKGROUND: Short-term effects of alirocumab on vascular function have hardly been investigated. Moreover, there is a scarce of reliable non-invasive methods to evaluate atherosclerotic changes of the vasculature. The ALIROCKS trial was performed to address these issues using standard ultrasound-based procedures and a completely novel magnetic resonance-based imaging technique. METHODS: A total of 24 patients with an indication for treatment with PCSK9 antibodies were recruited. There were 2 visits to the study site, the first before initiation of treatment with alirocumab and the second after 10 weeks of treatment. The key outcome measures included the change of carotid vessel wall fractional anisotropy, a novel magnetic resonance-based measure of vascular integrity, and the changes of carotid intima-media thickness and flow-dependent dilatation of the brachial artery measured with ultrasound. RESULTS: A total of 19 patients completed the trial, 2 patients stopped treatment, 3 patients did not undergo the second visit due to the COVID pandemic. All of them had atherosclerotic vascular disease. Their mean (standard deviation) LDL-cholesterol concentration was 154 (85) mg/dL at baseline and was reduced by 76 (44) mg/dL in response to alirocumab treatment (p < 0.001, n = 19). P-selectin and vascular endothelial growth factors remained unchanged. Flow-dependent dilatation of the brachial artery (+41%, p = 0.241, n = 18), carotid intima-media thickness (p = 0.914, n = 18), and fractional anisotropy of the carotid artery (p = 0.358, n = 13) also did not significantly change. CONCLUSION: Despite a nominal amelioration for flow-dependent dilatation, significant effects of short-term treatment with alirocumab on vascular function were not detectable. More work would be needed to evaluate, whether fractional anisotropy may be useful in clinical atherosclerosis research.

Sex-Specific Association of Serum Anti-Oxidative Capacity and Leukocyte Telomere Length.

Telomeres are a crucial factor in the preservation of genomic integrity, and an elevated risk for diseases such as cancer and cardiovascular events is related to shortened telomeres. However, telomere deterioration could be caused by factors such as chronic oxidative stress and inflammation, which are promoted by an imbalance among reactive oxygen species (ROS) and antioxidants. In this cross-sectional study, we investigated the relationship between telomeres and oxidative stress. The serum leucocyte telomer length (LTL), serum total antioxidant capacity (TAC) and the total serum lipid panel of 180 healthy athletic volunteers (90 males, 90 females) were measured Additionally, a questionnaire about sports behaviour and the type of training was completed. We observed a positive significant relation between serum LTL and TAC in the male group (cc = 3.4/p = 0.001) but not in females. There was no statistically significant correlation between age and physical activity and LTL in both groups. This is the first cross sectional study demonstrating an association between total serum TAC and LTL in healthy males, but interestingly, not in the females. Nevertheless, these results should be interpreted as preliminary, and further studies in independent cohorts are needed to investigate the sex-specific effects of oxidative stress on telomere length and telomerase activity.

Distribution of subcutaneous and intermuscular fatty tissue of the mid-thigh measured by MRI-A putative indicator of serum adiponectin level and individual factors of cardio-metabolic risk.

Obesity and metabolic syndrome (MetS) are associated with hypoadiponectinemia. On the contrary, studies revealed correlations between the amount of subcutaneous adipose tissue (SAT) and higher serum adiponectin levels. Furthermore, independent association of intermuscular adipose tissue (IMAT) deposit in the thigh with cardiometabolic risk factors (including total blood cholesterol, low-density lipoprotein (LDL), and triglycerides), and decreased insulin sensitivity, as MetS components, are sufficiently described. The combined relationship of thigh IMAT and SAT with serum adiponectin, leptin levels, and cardiometabolic risk factors have not been investigated till date. Since both SAT and IMAT play a role in fat metabolism, we hypothesized that the distribution pattern of SAT and IMAT in the mid-thigh might be related to adiponectin, leptin levels, and serum lipid parameters. We performed adipose tissue quantification using magnetic resonance imaging (MRI) of the mid-thigh in 156 healthy volunteers (78 male/78 female). Laboratory measurements of lipid panel, serum adiponectin, and leptin levels were conducted. Total serum adiponectin level showed a significant correlation with the percentage of SAT of the total thigh adipose tissue (SAT/ (IMAT+SAT)) for the whole study population and in sex-specific analysis. Additionally, SAT/(IMAT+SAT) was negatively correlated with known cardiometabolic risk factors such as elevated total blood cholesterol, LDL, and triglycerides; but positively correlated with serum high-density lipoprotein. In multiple linear regression analysis, (SAT/(IMAT+SAT)) was the most strongly associated variable with adiponectin. Interestingly, leptin levels did not show a significant correlation with this ratio. Adipose tissue distribution in the mid-thigh is not only associated to serum adiponectin levels, independent of sex. This proposed quantitative parameter for adipose tissue distribution could be an indicator for individual factors of a person`s cardiometabolic risk and serve as additional non-invasive imaging marker to ensure the success of lifestyle interventions.

Investigation of the Relationship between the Mid_Thigh Adipose Tissue Distribution Measured by MRI and Serum Osteocalcin-A Sex-Based Approach.

Osteocalcin, in its non-carboxylated form, has a positive effect on glucose metabolism. Additionally, osteocalcin levels are related to body composition, especially muscle mass. The relation to the distribution of different adipose tissue types, such as subcutaneous, intermuscular, and visceral adipose tissue, is unclear. This study aimed to investigate associations between serum osteocalcin and the distribution of subcutaneous and intermuscular adipose tissue of the mid-thigh. Furthermore, the influence of different training methods on osteocalcin levels was investigated. We performed adipose tissue quantification of subcutaneous adipose tissue (SAT) and intramuscular adipose tissue (IMAT) using MRI measurements of the mid-thigh in 128 volunteers (63 male/65 female). Laboratory analysis included blood lipid panel, serum insulin, adiponectin, and osteocalcin measurements. The main observation was a significant correlation of total serum osteocalcin (TOC) and the distribution of adipose tissue of the mid-thigh (SAT/(SAT + IMAT)) (cc = -0.29/p-value = 0.002), as well as the cross-sectional muscle area (MA), increasing with the weekly resistance training duration in males. Additionally, TOC (p-value = 0.01) and MA (p-value = 0.03) were negatively related to serum insulin. The significant relationship between TOC and SAT/(SAT + IMAT) is a new finding and confirms the negative influence of IMAT on glucose metabolism in a sex-specific approach. We could substantiate this by the negative relation of TOC with serum insulin.