Dataset Information

Posterior circulation stroke: machine learning-based detection of early ischemic changes in acute non-contrast CT scans.

ABSTRACT:

Objectives

Triage of patients with basilar artery occlusion for additional imaging diagnostics, therapy planning, and initial outcome prediction requires assessment of early ischemic changes in early hyperacute non-contrast computed tomography (NCCT) scans. However, accuracy of visual evaluation is impaired by inter- and intra-reader variability, artifacts in the posterior fossa and limited sensitivity for subtle density shifts. We propose a machine learning approach for detecting early ischemic changes in pc-ASPECTS regions (Posterior circulation Alberta Stroke Program Early CT Score) based on admission NCCTs.

Methods

The retrospective study includes 552 pc-ASPECTS regions (144 with infarctions in follow-up NCCTs) extracted from pre-therapeutic early hyperacute scans of 69 patients with basilar artery occlusion that later underwent successful recanalization. We evaluated 1218 quantitative image features utilizing random forest algorithms with fivefold cross-validation for the ability to detect early ischemic changes in hyperacute images that lead to definitive infarctions in follow-up imaging. Classifier performance was compared to conventional readings of two neuroradiologists.

Results

Receiver operating characteristic area under the curves for detection of early ischemic changes were 0.70 (95% CI [0.64; 0.75]) for cerebellum to 0.82 (95% CI [0.77; 0.86]) for thalamus. Predictive performance of the classifier was significantly higher compared to visual reading for thalamus, midbrain, and pons (P value < 0.05).

Conclusions

Quantitative features of early hyperacute NCCTs can be used to detect early ischemic changes in pc-ASPECTS regions. The classifier performance was higher or equal to results of human raters. The proposed approach could facilitate reproducible analysis in research and may allow standardized assessments for outcome prediction and therapy planning in clinical routine.

SUBMITTER: Kniep HC

PROVIDER: S-EPMC7419359 | biostudies-literature | 2020 Sep

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Posterior circulation stroke: machine learning-based detection of early ischemic changes in acute non-contrast CT scans.

Kniep Helge C HC Sporns Peter B PB Broocks Gabriel G Kemmling André A Nawabi Jawed J Rusche Thilo T Fiehler Jens J Hanning Uta U

Journal of neurology 20200511 9

<h4>Objectives</h4>Triage of patients with basilar artery occlusion for additional imaging diagnostics, therapy planning, and initial outcome prediction requires assessment of early ischemic changes in early hyperacute non-contrast computed tomography (NCCT) scans. However, accuracy of visual evaluation is impaired by inter- and intra-reader variability, artifacts in the posterior fossa and limited sensitivity for subtle density shifts. We propose a machine learning approach for detecting early ...[more]

PMID: 32394015

Similar Datasets

Project description:ImportanceChildhood arterial ischemic stroke (CAIS) affects approximately 1.6 per 100 000 children per year, while stroke recurs in up to 20% of patients at 5 years. Factors determining the risk of recurrence are incompletely understood.ObjectiveTo investigate the incidence of the recurrence of CAIS in the posterior and anterior circulations to determine if the risk differs between the 2 locations.Design, setting, and participantsA retrospective analysis of CAIS was conducted among children enrolled in a single-center prospective consecutive cohort at The Children's Hospital of Philadelphia between January 1, 2006, and January 1, 2015. Children with confirmed CAIS occurring between 29 days and 17.99 years were evaluated for inclusion. Patients were excluded if infarcts were located in both the anterior and posterior distributions or if CAIS occurred as a complication of intracranial surgery or brain tumor.Main outcomes and measuresStroke recurrence.ResultsThe study population included 107 patients (75 boys [70.1%] and 32 girls [29.9%]; median age at AIS, 7.7 years [interquartile range, 3.1-13.6 years]). Sixty-one children had anterior circulation CAIS (ACAIS) and 46 had posterior circulation CAIS (PCAIS). Median follow-up was 20.9 months (interquartile range, 8.7-40.4 months). For ACAIS, recurrence-free survival was 100% at 1 month and 96% (95% CI, 85%-99%) at 1 and 3 years. For PCAIS, recurrence-free survival was 88% (95% CI, 75%-95%) at 1 month and 81% (95% CI, 66%-90%) at 1 and 3 years. The hazard ratio for recurrence after PCAIS compared with ACAIS was 6.4 (95% CI, 1.4-29.8; P = .02) in univariable analysis and 5.3 (95% CI, 1.1-26.4; P = .04) after adjusting for sex and cervical dissection.Conclusions and relevanceWe identified a subgroup of patients that comprise more than 80% of recurrences of CAIS. Three years after incident stroke, 19% of children with PCAIS had a recurrence compared with 4% of patients with ACAIS. Different mechanisms of stroke may account for this difference. Children with PCAIS may warrant increased monitoring. This study highlights the necessity for further research focused on recurrence prevention.

Project description:(1) Background: Posterior circulation ischemic stroke has high mortality and disability rates and requires an early prediction prognosis to provide the basis for an interventional approach. Current quantitative measures are only able to accurately assess the prognosis of patients using magnetic resonance imaging (MRI). However, it is difficult to obtain MRI images in critically urgent cases. Therefore, the development of a noncontrast CT-based rapid-assist tool is needed to enhance the value of the clinical application. (2) Objective: This study aimed to develop an auxiliary-annotating noncontrast CT-efficient tool, which is based on a deep learning model, to provide a quantitative scale and the prognosis of posterior circulation ischemic stroke patients. (3) Methods: A total of 31 patients with posterior circulation ischemic stroke, diagnosed in the stroke registry at the Tri-Service General Hospital from November 2019 to July 2020, were included in the study, with a total of 578 CT images collected from noncontrast CT and MRI that were ≤ 3 days apart. A 5-fold cross validation was used to develop an image segmentation model to identify nine posterior circulation structures, and intersection over union (IoU) was used to assess the ability of the model to identify each structure. A quantitative score was integrated to assess the importance of the proportion of ischemic lesions in each posterior circulation structure, and the ROC curve was compared with the semiquantitative score for prognostic power. The prognoses of the patients were defined into two groups of 18 patients. An mRS score of 0-2 at discharge was defined as a good prognosis, while an mRS score of 3-6 was deemed to be a poor prognosis. (4) Results: The performance of the image segmentation model for identifying the nine posterior circulation structures in noncontrast CT images was evaluated. The IoU of the left cerebellum was 0.78, the IoU of the right cerebellum was 0.79, the IoU of the left occipital lobe was 0.74, the IoU of the right occipital lobe was 0.68, the IoU of the left thalamus was 0.73, the IoU of the right thalamus was 0.75, the IoU of the medulla oblongata was 0.82, and the IoU of the midbrain was 0.83. The prognostic AUC of posterior circulation patients predicted using a quantitative integrated score was 0.74, which was significantly higher than that of the pc-ASPECTS (AUC = 0.63, p = 0.035), with a sensitivity of 0.67 and a specificity of 0.72. (5) Conclusions: In this study, a deep learning model was used to develop a noncontrast CT-based quantitative integrated score tool, which is an effective tool for clinicians to assess the prognosis of posterior circulation ischemic stroke.

Project description:PurposeReperfusion therapies for acute ischemic stroke due to large-vessel occlusion (AIS-LVO) are highly time-dependent, and large infarction is related to poor outcomes and risk of symptomatic hemorrhage. It is of significance to investigate and optimize the screening means and selection criteria for reperfusion therapies to identify more appropriate patients with better outcomes. This study aimed to compare the performance of attenuation changes vs. automated Alberta Stroke Program Early CT Score (ASPECTS) and using CT angiography (CTA) source images vs. non-contrast CT (NCCT) in distinguishing the infarction extent of ischemic core volumes ≥ 70 ml within different time windows.MethodsA total of 73 patients with AIS-LVO who received multimodal CT were analyzed. The automated software was used to calculate ASPECTS. Attenuation change was defined as the sum of products of relative Hounsfield unit (rHU) values times weighting factors of all 10 ASPECTS regions. rHU value of each region was the HU of the ischemic side over that of the contralateral. The corresponding weighting factors were the regression coefficients derived from a multivariable linear regression model which was used to correlate regional rHU with ischemic core volumes, because each region in the ASPECTS template is weighted disproportionally in the ASPECTS system. Automated ASPECTS and attenuation changes were both calculated using CTA and NCCT, respectively.ResultsAttenuation changes were correlated with ischemic core volumes within different time windows (Rho ranging from 0.439 to 0.637). In classification of the ischemic core ≥ 70 ml, the performances of attenuation changes were comparable with ASPECTS (area under the curve [AUC] ranging from 0.799 to 0.891), with DeLong's test (P = 0.079, P = 0.373); using CTA (AUC = 0.842) was not different from NCCT (AUC = 0.838).ConclusionAttenuation changes in ASPECTS regions were correlated with ischemic core volumes. In the classification of infarction volumes, attenuation changes had a high diagnostic ability comparable with automated ASPECTS. Measurement of attenuation changes is not involved in complicated scoring algorithms. This measurement can be used as an available, rapid, reliable, and accurate means to evaluate infarction extent within different time windows. The usefulness of infarction volumes measured by attenuation changes to identify more appropriate patients for reperfusion therapies can be validated in future clinical trials.

Dataset Information

Posterior circulation stroke: machine learning-based detection of early ischemic changes in acute non-contrast CT scans.

Objectives

Methods

Results

Conclusions

Publications

Posterior circulation stroke: machine learning-based detection of early ischemic changes in acute non-contrast CT scans.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets