This study demonstrates the utilization of deep learning techniques for binary semantic segmentation of pores in carbon fiber reinforced polymers (CFRP) using X-ray computed tomography (XCT) ...datasets. The proposed workflow is designed to generate efficient segmentation models with reasonable execution time, applicable even for users using consumer-grade GPU systems. First, U-Net, a convolutional neural network, is modified to handle the segmentation of XCT datasets. In the second step, suitable hyperparameters are determined through a parameter analysis (hyperparameter tuning), and the parameter set with the best result was used for the final training. In the final step, we report on our efforts of implementing the testing stage in open_iA, which allows users to segment datasets with the fully trained model within reasonable time. The model performs well on datasets with both high and low resolution, and even works reasonably for barely visible pores with different shapes and size. In our experiments, we could show that U-Net is suitable for pore segmentation. Despite being trained on a limited number of datasets, it exhibits a satisfactory level of prediction accuracy.
Alpha-1 antitrypsin (AAT) deficiency causes/predisposes individuals to advanced chronic liver disease (ACLD). However, the role of the SERPINA1 Pi∗Z allele in patients who have already progressed to ...ACLD is unclear. Thus, we aimed to evaluate the impact of the Pi∗Z allele on the risk of liver transplantation/liver-related death in patients with ACLD, while adjusting for the severity of liver disease at inclusion.
A total of 1,118 patients with ACLD who underwent hepatic venous pressure gradient (HVPG) measurement and genotyping for the Pi∗Z/Pi∗S allele at the Vienna Hepatic Hemodynamic Lab were included in this retrospective analysis. The outcome of interest was liver transplantation/liver-related death, while non-liver-related death and removal/suppression of the primary etiological factor were considered as competing risks.
Viral hepatitis was the most common etiology (44%), followed by alcohol-related (31%) and non-alcoholic fatty liver disease (11%). Forty-two (4%) and forty-six (4%) patients harboured the Pi∗Z and Pi∗S variants, respectively. Pi∗Z carriers had more severe portal hypertension (HVPG: 19±6 vs.15±7 mmHg; p <0.001) and hepatic dysfunction (Child-Turcotte-Pugh: 7.1±1.9 vs. 6.5±1.9 points; p = 0.050) at inclusion, compared to non-carriers. Contrarily, the Pi∗S allele was unrelated to liver disease severity. In competing risk regression analysis, harbouring the Pi∗Z allele was significantly associated with an increased probability of liver transplantation/liver-related death, even after adjusting for liver disease severity at inclusion. The detrimental impact of the common Pi∗MZ genotype (adjusted subdistribution hazard ratio: ≈1.56 vs. Pi∗MM) was confirmed in a fully adjusted subgroup analysis. In contrast, Pi∗S carriers had no increased risk of events.
Genotyping for the Pi∗Z allele identifies patients with ACLD at increased risk of adverse liver-related outcomes, thereby improving prognostication. Therapies targeting the accumulation of abnormal AAT should be evaluated as disease-modifying treatments in Pi∗Z allele carriers with ACLD.
Alpha-1 antitrypsin deficiency is a genetic disease that affects the lung and the liver. Carrying two dysfunctional copies of the gene causes advanced liver disease. Harbouring one dysfunctional copy increases disease severity in patients with other liver illness. However, the significance of this genetic defect in patients who already suffer from advanced liver disease is unclear. Our study found that harbouring at least one dysfunctional copy of the alpha-1 antitrypsin gene increases the risk of requiring a liver transplantation or dying from a liver disease. This indicates the need for medical therapies aimed at treating the hepatic consequences of this genetic defect.
Display omitted
•Pi∗Z allele is significantly associated with liver-related events in patients with ACLD.•This finding extends to patients harbouring the Pi∗MZ genotype.•Genotyping for the Pi∗Z allele might improve prognostication in patients with ACLD.•Therapies targeting accumulation of abnormal AAT should be assessed in Pi∗Z carriers with ACLD.
Genetic factors such as the transmembrane 6 superfamily 2 (TM6SF2) rs58542926 single nucleotide variant(SNV) modulate the susceptibility for (advanced) chronic liver disease (ACLD). However, the ...impact of this variant in patients who have already progressed to ACLD is unknown.
The association between TM6SF2-rs58542926 genotype and liver-related events was evaluated in 938 ACLD patients undergoing hepatic venous pressure gradient (HVPG) measurement.
Mean HVPG was 15±7 mmHg and mean UNOS MELD (2016) 11±5 points. Viral hepatitis (n = 495, 53%) was the most common cause of ACLD, followed by alcohol-related (ARLD; n = 342, 37%) and non-alcoholic fatty liver disease (NAFLD; n = 101, 11%).
While 754 (80%) patients harboured the TM6SF2 wild-type (C/C), 174 (19%) and 10 (1%) patients had one or two T-alleles. At baseline, patients with at least one TM6SF2 T-allele had more pronounced portal hypertension (HVPG: 16±7 vs. 15±7 mmHg; p = 0.031), higher gamma-glutamyl transferase levels (123 63-229 vs. 97 55-174 UxL−1; p = 0.002), and more commonly hepatocellular carcinoma (17% vs. 12%; p = 0.049).
Harbouring the TM6SF2 T-allele was associated with the composite endpoint hepatic decompensation/liver transplantation/liver-related death (SHR: 1.44 95%CI: 1.14-1.83; p = 0.003). This was confirmed in multivariable competing risk regression analyses that were adjusted for severity of portal hypertension and hepatic dysfunction at baseline.
The TM6SF2 variant modulates liver disease progression beyond the development of ACLD, as it modifies the risks of hepatic decompensation and liver-related death, independently of baseline liver disease severity.
Display omitted
This work introduces a novel Augmented Reality (AR) approach to visualize material data alongside real objects in order to facilitate detailed material analyses based on spatial non-destructive ...testing (NDT) data as generated in X-ray computed tomography (XCT) imaging. For this purpose, we introduce a framework that leverages the potential of AR devices, visualization and interaction techniques to seamlessly explore complex primary and secondary XCT data matched with real-world objects. The overall goal of the proposed analysis scheme is to enable researchers and analysts to inspect material properties and structures onsite and in-place. Coupling immersive visualization techniques with real physical objects allows for highly intuitive workflows in material analysis and inspection, which enables the identification of anomalies and accelerates informed decision making. As a result, this framework generates an immersive experience, which provides a more engaging and more natural analysis of material data. A case study on fiber-reinforced polymer datasets was used to validate the AR framework and its new workflow. Initial results revealed positive feedback from experts, in particular regarding improved understanding of spatial data and a more natural interaction with material samples, which may have significant potential when combined with conventional analysis systems.
