In recent years, numerous methods have been proposed to predict tissue outcome in acute stroke patients using machine learning methods incorporating multiparametric imaging data. Most methods include ...diffusion and perfusion parameters as image-based parameters but do not include any spatial information although these parameters are spatially dependent, e.g. different perfusion properties in white and gray brain matter. This study aims to investigate if including spatial features improves the accuracy of multi-parametric tissue outcome prediction.
Acute and follow-up multi-center MRI datasets of 99 patients were available for this study. Logistic regression, random forest, and XGBoost machine learning models were trained and tested using acute MR diffusion and perfusion features and known follow-up lesions. Different combinations of atlas coordinates and lesion probability maps were included as spatial information. The stroke lesion predictions were compared to the true tissue outcomes using the area under the receiver operating characteristic curve (ROC AUC) and the Dice metric.
The statistical analysis revealed that including spatial features significantly improves the tissue outcome prediction. Overall, the XGBoost and random forest models performed best in every setting and achieved state-of-the-art results regarding both metrics with similar improvements achieved including Montreal Neurological Institute (MNI) reference space coordinates or voxel-wise lesion probabilities.
Spatial features should be integrated to improve lesion outcome prediction using machine learning models.
Automated brain volumetric analysis based on high-resolution T1-weighted MRI datasets is a frequently used tool in neuroimaging for early detection, diagnosis, and monitoring of various neurological ...diseases. However, image distortions can corrupt and bias the analysis. The aim of this study was to explore the variability of brain volumetric analysis due to gradient distortions and to investigate the effect of distortion correction methods implemented on commercial scanners.
36 healthy volunteers underwent brain imaging using a 3T magnetic resonance imaging (MRI) scanner, including a high-resolution 3D T1-weighted sequence. For all participants, each T1-weighted image was reconstructed directly on the vendor workstation with (DC) and without (nDC) distortion correction. For each participant's set of DC and nDC images, FreeSurfer was used for the determination of regional cortical thickness and volume.
Overall, significant differences were found in 12 cortical ROIs comparing the volumes of the DC and nDC data and in 19 cortical ROIs comparing the thickness of the DC and nDC data. The most pronounced differences for cortical thickness were found in the precentral gyrus, the lateral occipital and postcentral ROI (2.69, -2.91% and -2.79%, respectively) while cortical volumes differed most prominently in the paracentral, the pericalcarine and lateral occipital ROI (5.52%, -5.40% and -5.11%, respectively).
Correcting for gradient non-linearities can have significant influence on volumetric analysis of cortical thickness and volume. Since the distortion correction is an automatic feature of the MR scanner, it should be stated by each study that applies volumetric analysis which images were used.
Mean kurtosis (MK), one of the parameters derived from diffusion kurtosis imaging (DKI), has shown increased sensitivity to tissue microstructure damage in several neurological disorders.
...Thirty-seven patients with relapsing-remitting MS and eleven healthy controls (HC) received brain imaging on a 3T MR scanner, including a fast DKI sequence. MK and mean diffusivity (MD) were measured in the white matter of HC, normal-appearing white matter (NAWM) of MS patients, contrast-enhancing lesions (CE-L), FLAIR lesions (FLAIR-L) and black holes (BH).
Overall 1529 lesions were analyzed, including 30 CE-L, 832 FLAIR-L and 667 BH. Highest MK values were obtained in the white matter of HC (0.814 ± 0.129), followed by NAWM (0.724 ± 0.137), CE-L (0.619 ± 0.096), FLAIR-L (0.565 ± 0.123) and BH (0.549 ± 0.12). Lowest MD values were obtained in the white matter of HC (0.747 ± 0.068 10-3mm2/sec), followed by NAWM (0.808 ± 0.163 10-3mm2/sec), CE-L (0.853 ± 0.211 10-3mm2/sec), BH (0.957 ± 0.304 10-3mm2/sec) and FLAIR-L (0.976 ± 0.35 10-3mm2/sec). While MK differed significantly between CE-L and non-enhancing lesions, MD did not.
MK adds predictive value to differentiate between MS lesions and might provide further information about diffuse white matter injury and lesion microstructure.
An accurate prediction of tissue outcome in acute ischemic stroke patients is of high interest for treatment decision making. To date, various machine learning models have been proposed that combine ...multi-parametric imaging data for this purpose. However, most of these machine learning models were trained using voxel information extracted from the whole brain, without taking differences in susceptibility to ischemia into account that exist between brain regions. The aim of this study was to develop and evaluate a local tissue outcome prediction approach, which makes predictions using locally trained machine learning models and thus accounts for regional differences.
Multi-parametric MRI data from 99 acute ischemic stroke patients were used for the development and evaluation of the local tissue outcome prediction approach. Diffusion (ADC) and perfusion parameter maps (CBF, CBV, MTT, Tmax) and corresponding follow-up lesion masks for each patient were registered to the MNI brain atlas. Logistic regression (LR) and random forest (RF) models were trained employing a local approach, which makes predictions using models individually trained for each specific voxel position using the corresponding local data. A global approach, which uses a single model trained using all voxels of the brain, was used for comparison. Tissue outcome predictions resulting from the global and local RF and LR models, as well as a combined (hybrid) approach were quantitatively evaluated and compared using the area under the receiver operating characteristic curve (ROC AUC), the Dice coefficient, and the sensitivity and specificity metrics.
Statistical analysis revealed the highest ROC AUC and Dice values for the hybrid approach. With 0.872 (ROC AUC; LR) and 0.353 (Dice; RF), these values were significantly higher (p < 0.01) than the values of the two other approaches. In addition, the local approach achieved the highest sensitivity of 0.448 (LR). Overall, the hybrid approach was only outperformed in sensitivity (LR) by the local approach and in specificity by both other approaches. However, in these cases the effect sizes were comparatively small.
The results of this study suggest that using locally trained machine learning models can lead to better lesion outcome prediction results compared to a single global machine learning model trained using all voxel information independent of the location in the brain.
Myelin Water Fraction (MWF) measurements derived from quantitative Myelin Water Imaging (MWI) may detect demyelinating changes of the cerebral white matter (WM) microstructure. Here, we investigated ...age-related alterations of the MWF in normal aging brains of healthy volunteers utilizing two fast and clinically feasible 3D gradient and spin echo (GRASE) MWI sequences with 3 mm and 5 mm isotropic voxel size. In 45 healthy subjects (age range: 18-79 years), distinct regions of interest (ROI) were defined in the cerebral WM including corticospinal tracts. For the 3 mm sequence, significant correlations of the mean MWF with age were found for most ROIs (r < -0.8 for WM ROIs; r = -0.55 for splenium of corpus callosum; r = -0.75 for genu of corpus callosum; p < 0.001 for all ROIs). Similar correlations with age were found for the ROIs of the 5 mm sequence. No significant correlations were found for the corticospinal tract and the occipital WM (p > 0.05). Mean MWF values obtained from the 3 mm and 5 mm sequences were strongly comparable. The applied 3D GRASE MWI sequences were found to be sensitive for age-dependent myelin changes of the cerebral WM microstructure. The reported MWF values might be of substantial use as reference for further investigations in patient studies.
Infarct growth from the early ischemic core to the total infarct lesion volume (LV) is often used as an outcome variable of treatment effects, but can be overestimated due to vasogenic edema. The ...purpose of this study was (1) to assess two components of early lesion growth by distinguishing between water uptake and true net infarct growth and (2) to investigate potential treatment effects on edema-corrected net lesion growth. Sixty-two M1-MCA-stroke patients with acute multimodal and follow-up CT (FCT) were included. Ischemic lesion growth was calculated by subtracting the initial CTP-derived ischemic core volume from the LV in the FCT. To determine edema-corrected net lesion growth, net water uptake of the ischemic lesion on FCT was quantified and subtracted from the volume of uncorrected lesion growth. The mean lesion growth without edema correction was 20.4 mL (95% CI: 8.2–32.5 mL). The mean net lesion growth after edema correction was 7.3 mL (95% CI: −2.1–16.7 mL; p < 0.0001). Lesion growth was significantly overestimated due to ischemic edema when determined in early-FCT imaging. In 18 patients, LV was lower than the initial ischemic core volume by CTP. These apparently “reversible” core lesions were more likely in patients with shorter times from symptom onset to imaging and higher recanalization rates.
Most false negative findings in DWI of ischemic stroke are in patients with minor deficits clinically localized to the brainstem. Our goal was to evaluate the benefit of a thin-sliced sagittal DWI in ...addition to conventional axial DWI at 1.5T for the detection of brainstem infarctions.
Data of patients with symptoms consistent with acute and subacute brainstem infarction and an MRI examination including standard axial DWI and thin-sliced sagittal DWI were retrospectively analyzed. Patients with the later diagnosis of a TIA, an inflammation or a tumor of the brainstem were excluded from analysis. Diffusion restrictions were identified by two independent raters blinded for the final clinical diagnosis in three separate reading steps: First, only axial DWI, secondly only sagittal DWI, and lastly both DWIs together. Presence and size of DWI-lesions were documented for each plane. Differences between the observers were settled in consensus in a separate joint reading.
Of 73 included patients, 46 patients were clinically diagnosed with brainstem infarction. Inter-observer agreement was excellent for the detection of brainstem lesions in axial and sagittal DWI (kappa = 0.94 and 0.97). In 28/46 patients (60.9%) lesions were detected in the axial plane alone, whereas in 6 more patients (73.9%) lesions were detected in the review of both sequences together. All lesions undetectable in the axial plane were smaller than 5 mm in cranio-caudal direction.
Thin-sliced sagittal DWI in addition to axial DWI improves the detection rate of brainstem infarction with little additional expenditure of time.
Cortical and thalamic pathologies have been associated with cognitive impairment in patients with multiple sclerosis (MS).
We aimed to quantify cortical and thalamic damage in patients with MS using ...a high-resolution T1 mapping technique and to evaluate the association of these changes with clinical and cognitive impairment.
The study group consisted of 49 patients with mainly relapsing-remitting MS and 17 age-matched healthy controls who received 3T MRIs including a T1 mapping sequence (MP2RAGE). Mean T1 relaxation times (T1-RT) in the cortex and thalami were compared between patients with MS and healthy controls. Additionally, correlation analysis was performed to assess the relationship between MRI parameters and clinical and cognitive disability.
Patients with MS had significantly decreased normalized brain, gray matter, and white matter volumes, as well as increased T1-RT in the normal-appearing white matter, compared to healthy controls (
< 0.001). Partial correlation analysis with age, sex, and disease duration as covariates revealed correlations for T1-RT in the cortex (
= -0.33,
< 0.05), and thalami (right thalamus:
= -0.37, left thalamus:
= -0.50, both
< 0.05) with working memory and information processing speed, as measured by the Symbol-Digit Modalities Test.
T1-RT in the cortex and thalamus correlate with information processing speed in patients with MS.
Abstract
Mechanical thrombectomy (MT) for acute ischemic stroke with medium vessel occlusions is still a matter of debate. We sought to identify factors associated with clinical outcome after MT for ...M2-occlusions based on data from the German Stroke Registry-Endovascular Treatment (GSR-ET). All patients prospectively enrolled in the GSR-ET from 05/2015 to 12/2021 were analyzed (NCT03356392). Inclusion criteria were primary M2-occlusions and availability of relevant clinical data. Factors associated with excellent/good outcome (modified Rankin scale mRS 0–1/0–2), poor outcome/death (mRS 5–6) and mRS-increase pre-stroke to day 90 were determined in multivariable logistic regression. 1348 patients were included. 1128(84%) had successful recanalization, 595(44%) achieved good outcome, 402 (30%) had poor outcome. Successful recanalization (odds ratio OR 4.27 95% confidence interval 3.12–5.91,
p
< 0.001), higher Alberta stroke program early CT score (OR 1.25 1.18–1.32,
p
< 0.001) and i.v. thrombolysis (OR 1.28 1.07–1.54,
p
< 0.01) increased probability of good outcome, while age (OR 0.95 0.94–0.95,
p
< 0.001), higher pre-stroke-mRS (OR 0.36 0.31–0.40,
p
< 0.001), higher baseline NIHSS (OR 0.89 0.88–0.91,
p
< 0.001), diabetes (OR 0.52 0.42–0.64,
p
< 0.001), higher number of passes (OR 0.75 0.70–0.80,
p
< 0.001) and intracranial hemorrhage (OR 0.26 0.14–0.46,
p
< 0.001) decreased the probability of good outcome. Additional predictors of mRS-increase pre-stroke to 90d were dissections, perforations (OR 1.59 1.11–2.29,
p
< 0.05) and clot migration, embolization (OR 1.67 1.21–2.30,
p
< 0.01). Corresponding to large-vessel-occlusions, younger age, low pre-stroke-mRS, low severity of acute clinical disability, i.v. thrombolysis and successful recanalization were associated with good outcome while diabetes and higher number of passes decreased probability of good outcome after MT in M2 occlusions. Treatment related complications increased probability of mRS increase pre-stroke to 90d.
BACKGROUND AND PURPOSE—The early growth of ischemic lesions has been described as being nonlinear, with lesion growth rates at their highest during the earliest period after stroke onset. We ...hypothesized that the time gap from imaging to revascularization results in higher lesion growth in patients with hyperacute presentation.
METHODS—Fifty-one patients with ischemic stroke with initial multimodal computed tomography (CT), follow-up CT after 24 hours, and successful endovascular recanalization were included and separated into 2 groups according to their median time from symptom onset to imaging (eg, hyperacute versus acute). The difference in Alberta Stroke Program Early CT Score (ASPECTS) between initial CT and follow-up CT was assessed, as well as volumetric lesion growth from early ischemic core in admission perfusion CT and total lesion volume in follow-up CT.
RESULTS—The median time from onset to imaging was 1.85 hours. There was no significant difference in admission ASPECTS (mean, 8.5 versus 8.2) or time from imaging to recanalization in both groups (median, 2.7 versus 2.4 hours; P=0.4). The mean (SD) lesion growth assessed by ASPECTS difference was 2.7 (2.3) in the hyperacute group and 1.6 (1.3) in the acute group (P=0.03). The mean (SD) volumetric difference in the hyperacute group was 26.6 mL (43.2 mL) and 17.2 mL (26.3 mL; P=0.36) in the acute group, respectively. For every passing hour after onset, ASPECTS lesion growth was reduced by 0.4.
CONCLUSIONS—Patients in the hyperacute phase showed increased ASPECTS lesion growth from imaging to recanalization suggesting a particular benefit of faster recanalization times in this group of patients with stroke.