Shengyu Fan

To automatically and quantitatively evaluate the venous oxygen saturation (SvO2) in cerebral ischemic tissues and explore its value in predicting prognosis. A retrospective study was conducted on 48 AIS patients hospitalized in our hospital from 2015–2018. Based on quantitative susceptibility mapping and perfusion-weighted imaging, this paper measured the cerebral SvO2 in hypoperfusion tissues and its change after intraarterial rt-PA treatment. The cerebral SvO2 in different hypoperfusion regions between the favorable and unfavorable clinical outcome groups was analyzed using an independent t-test. Relationships between cerebral SvO2 and clinical scores were determined using the Pearson correlation coefficient. The receiver operating characteristic process was conducted to evaluate the accuracy of cerebral SvO2 in predicting unfavorable clinical outcomes. Cerebral SvO2 in hypoperfusion (Tmax > 4 and 6 s) was significantly different between the two groups at follow-up (p < 0.05). Cerebral SvO2 and its changes before and after treatment were negatively correlated with clinical scores. The positive predictive value, negative predictive value, accuracy, and area under the curve of the cerebral SvO2 were higher than those predicted by the ischemic core. Therefore, the cerebral SvO2 of hypoperfusion regions was a stronger imaging predictor of unfavorable clinical outcomes after stroke.

This study investigated the quantitative distribution of cerebral venous oxygen saturation (SvO2) based on quantitative sensitivity mapping (QSM) and determined its prognostic value in patients with acute ischemic stroke (AIS). A retrospective study was conducted on 39 hospitalized patients. Reconstructed QSM was used to calculate the cerebral SvO2 of each region of interest (ROI) in the ischemic hemisphere. The intraclass correlation coefficient (ICC) and Bland–Altman analysis were conducted to define the best resolution of the distribution map. The correlation between the cerebral SvO2 in hypoxic regions (SvO2_ROI < 0.7) and clinical scores was obtained by Spearman and power analysis. The associations between cerebral SvO2 and unfavorable prognosis were analyzed using multivariate logistic regression. Excellent agreement was found between the cerebral SvO2 in hypoxic regions with a resolution of 7.18 × 7.18 × 1.6 mm³ and asymmetrically prominent cortical veins regions (ICC: 0.879 (admission), ICC: 0.906 (discharge)). The cerebral SvO2 was significantly negative with clinical scores (all |r| > 0.3). The cerebral SvO2 and its changes at discharge were significantly associated with an unfavorable prognosis (OR: 0.812 and 0.866). Therefore, the cerebral SvO2 in hypoxic regions measured by the quantitative distribution map can be used as an indicator for evaluating the early prognosis of AIS.

Automated cerebrovascular segmentation of time-of-flight magnetic resonance angiography (TOF-MRA) images is an important technique, which can be used to diagnose abnormalities in the cerebrovascular system, such as vascular stenosis and malformation. Automated cerebrovascular segmentation can direct show the shape, direction and distribution of blood vessels. Although deep neural network (DNN)-based cerebrovascular segmentation methods have shown to yield outstanding performance, they are limited by their dependence on huge training dataset. In this paper, we propose an unsupervised cerebrovascular segmentation method of TOF-MRA images based on DNN and hidden Markov random field (HMRF) model. Our DNN-based cerebrovascular segmentation model is trained by the labeling of HMRF rather than manual annotations. The proposed method was trained and tested using 100 TOF-MRA images. The results were evaluated using the dice similarity coefficient (DSC), which reached a value of 0.79. The trained model achieved better performance than that of the traditional HMRF-based cerebrovascular segmentation method in binary pixel-classification. This paper combines the advantages of both DNN and HMRF to train the model with a not so large amount of the annotations in deep learning, which leads to a more effective cerebrovascular segmentation method.