Zhihui Fu

Objective: To evaluate the robustness of Diffusion Tensor Image Analysis Along the Perivascular Space (DTI-ALPS) index, and to investigate its correlation with clinical characteristics in patients following basal ganglia hemorrhage. Methods: (1) Twenty-three age- and sex-matched healthy controls were enrolled and underwent DTI scans using both 15-direction and 32-direction sequences. The differences and consistency of ALPS index between the two protocols were compared. (2) A mixed cross-sectional and longitudinal study was conducted on 56 patients. Patients were divided into three groups by time since hemorrhage: ≤ 14 days, 3 months and 1 year. The ALPS index was calculated and compared among patient groups and against healthy controls. Correlations between the ALPS index with hematoma volume and the Brunnstrom recovery stage for the hand (BRS-H) were analyzed. Results: (1) Comparison of ALPS index calculated from 15-direction and 32-direction DTI data revealed significant differences in the left hemisphere (ALPS-L) between the two protocols, while no statistically significant difference was observed in the right hemisphere (ALPS-R) and the bilateral mean (ALPS-mean). The consistency between the two protocols was excellent across all groups (ICC = 0.959–0.981, p < 0.05). (2) ① ALPS index differed significantly among the three patient groups ( p < 0.05). ALPS index for each group were as follows: Group A ( n = 34): 1.24 ± 0.03; Group B ( n = 12): 1.14 ± 0.04; Group C ( n = 10): 1.18 ± 0.04. The ALPS index in all patient groups were significantly lower than in controls ( p < 0.001). ② No significant correlation was found between the ALPS index and the BRS-H ( r = 0.219, p > 0.05). ③ Overall, no significant correlation was observed between ALPS index and hematoma volume ( r = −0.252, p > 0.05). However, stratified analysis showed a significant negative correlation in the ≤ 10 mL subgroup ( r = −0.529, p < 0.05), but not in the > 10 mL subgroup ( r = −0.091, p > 0.05). Conclusion: Basal ganglia hemorrhage is associated with impaired glymphatic function, with partial recovery after 3 months. The ALPS index showed a significant negative correlation with hematoma volume, supporting its sensitivity to hematoma-induced brain injury.

Rib fractures are common injuries caused by chest trauma, which may cause serious consequences. It is essential to diagnose rib fractures accurately. Low-dose thoracic computed tomography (CT) is commonly used for rib fracture diagnosis, and convolutional neural network- (CNN-) based methods have assisted doctors in rib fracture diagnosis in recent years. However, due to the lack of rib fracture data and the irregular, various shape of rib fractures, it is difficult for CNN-based methods to extract rib fracture features. As a result, they cannot achieve satisfying results in terms of accuracy and sensitivity in detecting rib fractures. Inspired by the attention mechanism, we proposed the CFSG U-Net for rib fracture detection. The CSFG U-Net uses the U-Net architecture and is enhanced by a dual-attention module, including a channel-wise fusion attention module (CFAM) and a spatial-wise group attention module (SGAM). CFAM uses the channel attention mechanism to reweight the feature map along the channel dimension and refine the U-Net’s skip connections. SGAM uses the group technique to generate spatial attention to adjust feature maps in the spatial dimension, which allows the spatial attention module to capture more fine-grained semantic information. To evaluate the effectiveness of our proposed methods, we established a rib fracture dataset in our research. The experimental results on our dataset show that the maximum sensitivity of our proposed method is 89.58%, and the average FROC score is 81.28%, which outperforms the existing rib fracture detection methods and attention modules.

Burn is a common traumatic disease with high morbidity and mortality. The treatment of burns requires accurate and reliable diagnosis of burn wounds and burn depth, which can save lives in some cases. However, due to the complexity of burn wounds, the early diagnosis of burns lacks accuracy and difference. Therefore, we use deep learning technology to automate and standardize burn diagnosis to reduce human errors and improve burn diagnosis. First, the burn dataset with detailed burn area segmentation and burn depth labelling is created. Then, an end-to-end framework based on deep learning method for advanced burn area segmentation and burn depth diagnosis is proposed. The framework is firstly used to segment the burn area in the burn images. On this basis, the calculation of the percentage of the burn area in the total body surface area (TBSA) can be realized by extending the network output structure and the labels of the burn dataset. Then, the framework is used to segment multiple burn depth areas. Finally, the network achieves the best result with IOU of 0.8467 for the segmentation of burn and no burn area. And for multiple burn depth areas segmentation, the best average IOU is 0.5144.