Aim: The atherogenic index of plasma (AIP) was significantly related to adverse outcomes in patients with cardiovascular disease. Our aim was to investigate the association between AIP and adverse outcomes in acute ischemic stroke.Methods: Patients with acute ischemic stroke (AIS) admitted between 2015 and 2018 were prospectively enrolled in this study. Functional outcomes were evaluated by the modified Rankin Scale (mRS). Poor outcomes were defined as mRS 3–6. The relationship of AIP with the risk of outcomes was analyzed by multivariate logistic regression models.Results: A total of 1,463 patients with AIS within 24 h of symptom onset were enrolled. The poor outcome group had a significantly higher level of AIP [0.09 (−0.10 to 0.27) vs. 0.04 (−0.09 to 0.18), p < 0.001] compared with the good outcome group. Multivariable logistic regression analysis showed that higher AIP was associated with poor outcomes in all the stroke patients (OR 1.84, 95% CI, 1.23–2.53, p = 0.007), which was more evident in patients with large-artery atherosclerosis subtype (OR 1.90, 95% CI, 1.53–2.62, p = 0.002), but not in the other subtypes. Receiver operating curve (ROC) analysis revealed that the best predictive cutoff value of AIP was 0.112, with a sensitivity of 70.8% and a specificity of 59.2%, and the area under the ROC curves for AIP was 0.685.Conclusion: AIP may be an important and independent predictor of the outcome of dysfunction in patients with AIS, especially the stroke subtype of large-artery atherosclerosis.

Background Lymphocyte to monocyte ratio (LMR) is associated with functional outcome in patients with stroke. But the relationship between the LMR value and the prognosis of cerebral venous sinus thrombosis (CVST) has not been investigated. Methods CVST patients, admitted to the First Affiliated Hospital of Zhengzhou University, were retrospectively identified from November 2010 to January 2017. Functional outcomes of patients were evaluated with the modified Rankin Scale (mRS). Patients were divided into good (mRS 0–2) and poor (mRS 3–6) outcomes groups. Univariate and multivariate Cox regression analyses were used to assess the relationship between LMR and the poor survival outcome. Results A total of 228 patients were included of which 41 had poor outcomes (18.0%). The duration of follow-up was 22 months (6–66 months). LMR (2.3±1.2 vs 3.2±1.8, p<0.01) was significantly lower in the poor outcome group. Multivariate Cox regression analysis showed that LMR (HR 0.726, 95% CI 0.546 to 0.964, p=0.027) was a independent predictor of poor prognosis. Conclusions LMR may be a predictor of poor prognosis in CVST patients.

The relationship between plasma level of total homocysteine (tHcy) and white matter hyperintensities (WMHs), especially in patients with acute ischemic stroke (AIS), is controversial. The present study investigated the association between these two as well as WMH locations in a large cohort of patients with AIS. Consecutive patients were reviewed from a prospective ischemic stroke database. Clinical data, including tHcy level and WMHs, were assessed. WMHs were assessed using the Fazekas scale and Age-Related White Matter Changes (ARWMC) visual grading scale. The association between tHcy and WMH locations was investigated by using multivariate logistic regression analyses. A total of 923 out of 1,205 patients were examined. The average age was 58.9 ± 11.9 years; 31.6% were female. Elevated tHcy level was significantly associated with WMHs. For the highest tHcy quartile, the odds ratio (OR) (95% confidence interval; CI) was 1.891 (1.257; 2.843) according to the Fazekas scale and 1.781 (1.185; 2.767) according to the ARWMC scale when compared to the lowest quartile. However, in a subgroup analysis, only WMHs in the periventricular area and left or right frontal areas were found to be independently associated with tHcy level. For the highest tHcy quartile, the OR (95% CI) was 1.761 (1.172; 2.648) for the periventricular WMHs, 1.768 (1.134; 2.756)for the left frontal WMHs, and 1.890 (1.206; 2.960)for the right frontal WMHs. In patients with AIS, plasma tHcy level is related to WMHs, especially WMHs distributed within the periventricular and frontal areas.

Background: TOAST subtype classification is important for diagnosis and research of ischemic stroke. Limited by experience of neurologist and time-consuming manual adjudication, it is a big challenge to finish TOAST classification effectively. We propose a novel active deep learning architecture to classify TOAST. Methods: To simulate the diagnosis process of neurologists, we drop the valueless features by XGB algorithm and rank the remaining ones. Utilizing active learning framework, we propose a novel causal CNN, in which it combines with a mixed active selection criterion to optimize the uncertainty of samples adaptively. Meanwhile, KL-focal loss derived from the enhancement of Focal loss by KL regularization is introduced to accelerate the iterative fine-tuning of the model. Results: To evaluate the proposed method, we construct a dataset which consists of totally 2310 patients. In a series of sequential experiments, we verify the effectiveness of each contribution by different evaluation metrics. Experimental results show that the proposed method achieves competitive results on each evaluation metric. In this task, the improvement of AUC is the most obvious, reaching 77.4. Conclusions: We construct a backbone causal CNN to simulate the neurologist process of that could enhance the internal interpretability. The research on clinical data also indicates the potential application value of this model in stroke medicine. Future work we would consider various data types and more comprehensive patient types to achieve fully automated subtype classification.

The goal of this study was to determine whether the stress hyperglycemia ratio (SHR) is associated with early neurological deterioration (END) and poor outcomes in patients with single subcortical infarct (SSI). For this study, we prospectively enrolled patients with SSI admitted between 2015 and 2021. SHR was distributed in quartiles according to the size of each subgroup. END was defined as an increase of ≥ 2 total points in the National Institutes of Health Stroke Scale (NIHSS) or ≥ 1 point in the motor items of the NIHSS within 7 days of hospital admission. The modified Rankin Scale (mRS) was used to evaluate patient prognosis. Good and poor outcomes were defined as mRS scores ≤2 and >2, respectively. The relationships between SHR and risk of END as well as outcomes were analyzed using multivariate logistic regression models. A total of 1049 patients with SSI with an average age of 59.49 years met the inclusion criteria for the analysis. The incidence of END markedly increased with increasing SHR. Multivariate logistic regression analysis revealed that the highest SHR quartile was independently associated with an increased risk of both END (OR 4.04, 95% CI, 2.43-6.69, P < 0.001) and 3-month poor outcomes (OR 2.34, 95% CI, 1.44-3.82, P = 0.001), compared to the lowest quartile. A receiver operating characteristic curve analysis of the SHR based on the area under the curve showed a diagnostic accuracy equal or greater than fasting plasma glucose. SHR is a reliable predictor of END and poor outcomes in patients with SSI.

Clinically, physicians collect the benchmark medical data to establish archives for a stroke patient and then add the follow up data regularly. It has great significance on prognosis prediction for stroke patients. In this paper, we present an interpretable deep learning model to predict the one-year mortality risk on stroke. We design sub-modules to reconstruct features from original clinical data that highlight the dissimilarity and temporality of different variables. The model consists of Bidirectional Long Short-Term Memory (Bi-LSTM), in which a novel correlation attention module is proposed that takes the correlation of variables into consideration. In experiments, datasets are collected clinically from the department of neurology in a local AAA hospital. It consists of 2,275 stroke patients hospitalized in the department of neurology from 2014 to 2016. Our model achieves a precision of 0.9414, a recall of 0.9502 and an F1-score of 0.9415. In addition, we provide the analysis of the interpretability by visualizations with reference to clinical professional guidelines.