A machine learning algorithm detected infarction in patients with acute stroke on baseline nonenhanced CT images with precision similar to that of diffusion-weighted MRI. Background Identifying the presence and extent of infarcted brain tissue at baseline plays a crucial role in the treatment of patients with acute ischemic stroke (AIS). Patients with extensive infa...

BACKGROUND AND PURPOSE: Alberta Stroke Program Early CT Score (ASPECTS) was devised as a systematic method to assess the extent of early ischemic change on noncontrast CT (NCCT) in patients with acute ischemic stroke (AIS). Our aim was to automate ASPECTS to objectively score NCCT of AIS patients. MATERIALS AND METHODS: We collected NCCT images with a 5-mm thickness of 257 patients with acute ischemic stroke (<8 hours from onset to scans) followed by a diffusion-weighted imaging acquisition within 1 hour. Expert ASPECTS readings on DWI were used as ground truth. Texture features were extracted from each ASPECTS region of the 157 training patient images to train a random forest classifier. The unseen 100 testing patient images were used to evaluate the performance of the trained classifier. Statistical analyses on the total ASPECTS and region-level ASPECTS were conducted. RESULTS: For the total ASPECTS of the unseen 100 patients, the intraclass correlation coefficient between the automated ASPECTS method and DWI ASPECTS scores of expert readings was 0.76 (95% confidence interval, 0.67–0.83) and the mean ASPECTS difference in the Bland-Altman plot was 0.3 (limits of agreement, −3.3, 2.6). Individual ASPECTS region-level analysis showed that our method yielded κ = 0.60, sensitivity of 66.2%, specificity of 91.8%, and area under curve of 0.79 for 100 × 10 ASPECTS regions. Additionally, when ASPECTS was dichotomized (>4 and ≤4), κ = 0.78, sensitivity of 97.8%, specificity of 80%, and area under the curve of 0.89 were generated between the proposed method and expert readings on DWI. CONCLUSIONS: The proposed automated ASPECTS scoring approach shows reasonable ability to determine ASPECTS on NCCT images in patients presenting with acute ischemic stroke.

BACKGROUND AND PURPOSE: Thrombus characteristics identified on non-contrast CT (NCCT) are potentially associated with recanalization with intravenous (IV) alteplase in patients with acute ischemic stroke (AIS). Our aim was to determine the best radiomics-based features of thrombus on NCCT and CT angiography associated with recanalization with IV alteplase in AIS patients and proximal intracranial thrombi. MATERIALS AND METHODS: With a nested case-control design, 67 patients with ICA/M1 MCA segment thrombus treated with IV alteplase were included in this analysis. Three hundred twenty-six radiomics features were extracted from each thrombus on both NCCT and CTA images. Linear discriminative analysis was applied to select features most strongly associated with early recanalization with IV alteplase. These features were then used to train a linear support vector machine classifier. Ten times 5-fold cross-validation was used to evaluate the accuracy of the trained classifier and the stability of the selected features. RESULTS: Receiver operating characteristic curves showed that thrombus radiomics features are predictive of early recanalization with IV alteplase. The combination of radiomics features from NCCT, CTA, and radiomics changes is best associated with early recanalization with IV alteplase (area under the curve = 0.85) and was significantly better than any single feature such as thrombus length (P < .001), volume (P < .001), and permeability as measured by mean attenuation increase (P < .001), maximum attenuation in CTA (P < .001), maximum attenuation increase (P < .001), and assessment of residual flow grade (P < .001). CONCLUSIONS: Thrombus radiomics features derived from NCCT and CTA are more predictive of recanalization with IV alteplase in patients with acute ischemic stroke with proximal occlusion than previously known thrombus imaging features such as length, volume, and permeability.

Lesion segmentation is a fundamental step for the diagnosis of acute ischemic stroke (AIS). Non-contrast CT (NCCT) is still a mainstream imaging modality for AIS lesion measurement. However, AIS lesion segmentation on NCCT is challenging due to low contrast, noise and artifacts. To achieve accurate AIS lesion segmentation on NCCT, this study proposes a hybrid convolutional neural network (CNN) and Transformer network with circular feature interaction and bilateral difference learning. It consists of parallel CNN and Transformer encoders, a circular feature interaction module, and a shared CNN decoder with a bilateral difference learning module. A new Transformer block is particularly designed to solve the weak inductive bias problem of the traditional Transformer. To effectively combine features from CNN and Transformer encoders, we first design a multi-level feature aggregation module to combine multi-scale features in each encoder and then propose a novel feature interaction module containing circular CNN-to-Transformer and Transformer-to-CNN interaction blocks. Besides, a bilateral difference learning module is proposed at the bottom level of the decoder to learn the different information between the ischemic and contralateral sides of the brain. The proposed method is evaluated on three AIS datasets: the public AISD, a private dataset and an external dataset. Experimental results show that the proposed method achieves Dices of 61.39% and 46.74% on the AISD and the private dataset, respectively, outperforming 17 state-of-the-art segmentation methods. Besides, volumetric analysis on segmented lesions and external validation results imply that the proposed method is potential to provide support information for AIS diagnosis.

BACKGROUND AND PURPOSE: Infarct volume is an important predictor of clinical outcome in acute stroke. We hypothesized that the association of infarct volume and clinical outcome changes with the magnitude of infarct size. MATERIALS AND METHODS: Data were derived from the Safety and Efficacy of Nerinetide in Subjects Undergoing Endovascular Thrombectomy for Stroke (ESCAPE-NA1) trial, in which patients with acute stroke with large-vessel occlusion were randomized to endovascular treatment plus either nerinetide or a placebo. Infarct volume was manually segmented on 24-hour noncontrast CT or DWI. The relationship between infarct volume and good outcome, defined as mRS 0–2 at 90 days, was plotted. Patients were categorized on the basis of visual grouping at the curve shoulders of the infarct volume/outcome plot. The relationship between infarct volume and adjusted probability of good outcome was fitted with linear or polynomial functions as appropriate in each group. RESULTS: We included 1099 individuals in the study. Median infarct volume at 24 hours was 24.9 mL (interquartile range [IQR] = 6.6–92.2 mL). On the basis of the infarct volume/outcome plot, 4 infarct volume groups were defined (IQR = 0–15 mL, 15.1–70 mL, 70.1–200 mL, >200 mL). Proportions of good outcome in the 4 groups were 359/431 (83.3%), 219/337 (65.0%), 71/201 (35.3%), and 16/130 (12.3%), respectively. In small infarcts (IQR = 0–15 mL), no relationship with outcome was appreciated. In patients with intermediate infarct volume (IQR = 15–200 mL), there was progressive importance of volume as an outcome predictor. In infarcts of > 200 mL, outcomes were overall poor. CONCLUSIONS: The relationship between infarct volume and clinical outcome varies nonlinearly with the magnitude of infarct size. Infarct volume was linearly associated with decreased chances of achieving good outcome in patients with moderate-to-large infarcts, but not in those with small infarcts. In very large infarcts, a near-deterministic association with poor outcome was seen.