Search

In patients with interstitial lung disease (ILD), accurate pattern assessment from their computed tomography (CT) images could help track lung abnormalities and evaluate treatment efficacy. Based on excellent image classification performance, convolutional neural networks (CNNs) have been massively investigated for classifying and labeling pathological patterns in the CT images of ILD patients. However, previous studies rarely considered the three-dimensional (3D) structure of the pathological patterns of ILD and used two-dimensional network input. In addition, ResNet-based networks such as SE-ResNet and ResNeXt with high classification performance have not been used for pattern classification of ILD. This study proposed a SE-ResNeXt-SA-18 for classifying pathological patterns of ILD. The SE-ResNeXt-SA-18 integrated the multipath design of the ResNeXt and the feature weighting of the squeeze-and-excitation network with split attention. The classification performance of the SE-ResNeXt-SA-18 was compared with the ResNet-18 and SE-ResNeXt-18. The influence of the input patch size on classification performance was also evaluated. Results show that the classification accuracy was increased with the increase of the patch size. With a 32 × 32 × 16 input, the SE-ResNeXt-SA-18 presented the highest performance with average accuracy, sensitivity, and specificity of 0.991, 0.979, and 0.994. High-weight regions in the class activation maps of the SE-ResNeXt-SA-18 also matched the specific pattern features. In comparison, the performance of the SE-ResNeXt-SA-18 is superior to the previously reported CNNs in classifying the ILD patterns. We concluded that the SE-ResNeXt-SA-18 could help track or monitor the progress of ILD through accuracy pattern classification. [ABSTRACT FROM AUTHOR]

Objective: Diffusion tensor imaging (DTI) is a useful approach for studying neuronal integrity in animals. However, the test–retest reproducibility of DTI techniques in animals has not been discussed. Therefore, the first part of this work was to systematically elucidate the reliability of DTI-derived parameters in an animal study. Subsequently, we applied the DTI approach to an animal model of diabetes in a longitudinal manner. Materials and methods: In Study 1, nine rats underwent two DTI sessions using the same scanner and protocols, with a gap of 4 weeks. The reliability of the DTI-derived parameters was evaluated in terms of sessions and raters. In Study 2, nine rats received a single intraperitoneal injection of 70 mg/kg streptozotocin (STZ) to develop diabetes. Longitudinal DTI scans were used to assess brain alterations before and 4 weeks after STZ administration. Results: In the test–retest evaluation, the inter-scan coefficient of variation (CoV) ranged from 3.04 to 3.73% and 2.12–2.59% for fractional anisotropy (FA) and mean diffusivity (MD), respectively, in different brain regions, suggesting excellent reproducibility. Moreover, rater-dependence had minimal effects on FA and MD quantification, with all inter-rater CoV values less than 4%. Following the onset of diabetes, FA in striatum and cortex were noted to be significantly lower relative to the period where they had not developed diabetes (both P < 0.05). However, when compared to the control group, a significant change in FA caused by diabetes was detected only in the striatum (P < 0.05), but not in the cortex. Conclusion: These results demonstrate good inter-rater and inter-scan reliability of DTI in animal studies, and the longitudinal setting has a beneficial effect on detecting small changes in the brain due to diseases. [ABSTRACT FROM AUTHOR]