Introduction: Despite improvements in device technology, only one-third of stroke patients undergoing endovascular thrombectomy (EVT) achieve first-pass effect (FPE). We investigated the effect of arterial tortuosity and thrombus characteristics on the relationship between first-line EVT strategy and angiographic outcomes. Patients and methods: Patients with thin-slice baseline CT-angiography from the ESCAPE-NA1 trial (Efficacy and safety of nerinetide for the treatment of acute ischemic stroke) were included. Tortuosity was estimated using the tortuosity index extracted from catheter pathway, and radiological thrombus characteristics were length, non-contrast density, perviousness and hyperdense artery sign. We assessed the association of first-line EVT strategy (stent-retriever [SR] versus contact aspiration [CA] versus combined SR+CA) with FPE (eTICI score 2c/3 after one pass), final eTICI 2b/3, number of passes and procedure duration using multivariable regression. Interaction of tortuosity and thrombus characteristics with first-line technique were assessed using interaction terms. Results: Among 520 included patients, SR as a first-line modality was used in 165 (31.7%) patients, CA in 132 (25.4%), and combined SR+CA in 223 (42.9%). FPE was observed in 166 patients (31.9%). First-line strategy was not associated with FPE. Tortuosity had a significant effect on FPE only in the CA group (aOR = 0.90 [95% CI 0.83–0.98]) compared with stent-retrievers and combined first-line approach (p interaction = 0.03). There was an interaction between thrombus length and first-line strategy for number of passes (p interaction = 0.04). Longer thrombi were associated with higher number of passes only in the CA group (acOR 1.03 [95% CI 1.00–1.06]). Conclusion: Our study suggests that vessel tortuosity and longer thrombi may negatively affect the performance of first-line contact aspiration catheters in acute stroke patients undergoing EVT.

BACKGROUND AND PURPOSE: Identifying the presence and extent of intracranial thrombi is crucial in selecting patients with acute ischemic stroke for treatment. This article aims to develop an automated approach to quantify thrombus on NCCT and CTA in patients with stroke. MATERIALS AND METHODS: A total of 499 patients with large-vessel occlusion from the Safety and Efficacy of Nerinetide in Subjects Undergoing Endovascular Thrombectomy for Stroke (ESCAPE-NA1) trial were included. All patients had thin-section NCCT and CTA images. Thrombi contoured manually were used as reference standard. A deep learning approach was developed to segment thrombi automatically. Of 499 patients, 263 and 66 patients were randomly selected to train and validate the deep learning model, respectively; the remaining 170 patients were independently used for testing. The deep learning model was quantitatively compared with the reference standard using the Dice coefficient and volumetric error. The proposed deep learning model was externally tested on 83 patients with and without large-vessel occlusion from another independent trial. RESULTS: The developed deep learning approach obtained a Dice coefficient of 70.7% (interquartile range, 58.0%–77.8%) in the internal cohort. The predicted thrombi length and volume were correlated with those of expert-contoured thrombi (r = 0.88 and 0.87, respectively; P < .001). When the derived deep learning model was applied to the external data set, the model obtained similar results in patients with large-vessel occlusion regarding the Dice coefficient (66.8%; interquartile range, 58.5%–74.6%), thrombus length (r = 0.73), and volume (r = 0.80). The model also obtained a sensitivity of 94.12% (32/34) and a specificity of 97.96% (48/49) in classifying large-vessel occlusion versus non-large-vessel occlusion. CONCLUSIONS: The proposed deep learning method can reliably detect and measure thrombi on NCCT and CTA in patients with acute ischemic stroke.

BACKGROUND: Early ischemic change and collateral extent are colinear with ischemic core volume (ICV). We investigated the relationship between a combined score using the Alberta Stroke Program Early Computed Tomography Score and multiphase computed tomography angiography (mCTA) collateral extent, named mCTA-ACE score, on functional outcomes in endovascular therapy–treated patients. METHODS: We performed a post hoc analysis of a subset of endovascular therapy–treated patients from the Alteplase Compared to Tenecteplase trial which was conducted between December 2019 and January 2022 at 22 centers across Canada. Ten-point mCTA collateral corresponding to M2 to M6 regions of the Alberta Stroke Program Early Computed Tomography Score grid was evaluated as 0 (poor), 1 (moderate), or 2 (normal) and additively combined with the 10-point Alberta Stroke Program Early Computed Tomography Score to produce a 20-point mCTA-ACE score. We investigated the association of mCTA-ACE score with modified Rankin Scale score ≤2 and return to prestroke level of function at 90 to 120 days using mixed-effects logistic regression. In the subset of patients who underwent baseline computed tomography perfusion imaging, we compared the mCTA-ACE score and ICV for outcome prediction. RESULTS: Among 1577 intention-to-treat population in the trial, 368 (23%; 179 men; median age, 73 years) were included, with Alberta Stroke Program Early Computed Tomography Score, mCTA collateral, and combination of both (mCTA-ACE score: median [interquartile range], 8 [7–10], 9 [8–10], and 17 [16–19], respectively). The probability of modified Rankin Scale score ≤2 and return to prestroke level of function increased for each 1-point increase in mCTA-ACE score (odds ratio, 1.16 [95% CI, 1.06–1.28] and 1.22 [95% CI, 1.06–1.40], respectively). Among 173 patients in whom computed tomography perfusion data was assessable, the mCTA-ACE score was inversely correlated with ICV (ρ=−0.46; P <0.01). The mCTA-ACE score was comparable to ICV to predict a modified Rankin Scale score ≤2 and return to prestroke level of function (C statistics 0.71 versus 0.69 and 0.68 versus 0.64, respectively). CONCLUSIONS: The mCTA-ACE score had a significant positive association with functional outcomes after endovascular therapy and had a similar predictive performance as ICV.

Although Deep Convolutional Neural Networks (DCNNs) facilitate the evolution of 3D human pose estimation, ambiguity remains the most challenging problem in such tasks. Inspired by the Human Perception Mechanism (HPM), we propose an image-to-pose coding method to fill the gap between image cues and 3D poses, thereby alleviating the ambiguity of 3D human pose estimation. First, in 3D pose space, we divide the whole 3D pose space into multiple subregions named pose codes , turning a disambiguation problem into a classification problem. The proposed coding mechanism covers multiple camera views and provides a complete description for 3D pose space. Second, it is noteworthy that the articulated structure of the human body lies on a sophisticated product manifold and the error accumulation in the chain structure will undoubtedly affect the coding performance. Therefore, in image space, we extract the image cues from independent local image patches rather than the whole image. The mapping relationship between image cues and 3D pose codes is established by a set of DCNNs. The image-to-pose coding method transforms the implicit image cues into explicit constraints. Finally, the image-to-pose coding method is integrated into a linear matching mechanism to construct a 3D pose estimation method that effectively alleviates the ambiguity. We conduct extensive experiments on widely used public benchmarks. The experimental results show that our method effectively alleviates the ambiguity in 3D pose recovery and is robust to the variations of view.

The stroke lesion volume is a key radiologic measurement for assessing the prognosis of acute ischemic stroke (AIS) patients, which is challenging to be automatically measured on noncontrast CT (NCCT) scans. Recent diffusion probabilistic models (DPMs) in the domain of image generation have shown potentials of being used for lesion volume segmentation on medical images. In this article, a novel synchronous image-label diffusion probability model (SDPM) is proposed for stroke lesion segmentation on NCCT using a dual-Markov diffusion process with shared noise. The proposed SDPM is fully based on a generative latent variable model (LVM), offering a probabilistic elaboration from stem to stem. To fit into our segmentation tasks using the strength from generation models, we develop the architecture of the network where an additional net-stream, parallel with a noise prediction stream, is introduced to obtain the initial label estimates with noise for efficiently inferring the final labels. By optimizing the specified variational boundaries, the trained model can infer the final label estimates given the input images at any scale of time in four different label-inference methods, which gives more flexibility to the proposed SDPM. The proposed model was assessed on three stroke lesion datasets including one public and two private datasets. Compared with several U-Net, transformer, and DPM-based segmentation methods, our proposed SDPM model is able to achieve the state-of-the-art accuracy.