Pr Omer EKER

learn more

Aims To evaluate the performance of three MR perfusion software packages (A: RAPID; B: OleaSphere; and C: Philips) in predicting final infarct volume (FIV). Methods This cohort study included patients treated with mechanical thrombectomy following an admission MRI and undergoing a follow-up MRI. Admission MRIs were post-processed by three packages to quantify ischemic core and perfusion deficit volume (PDV). Automatic package outputs (uncorrected volumes) were collected and corrected by an expert. Successful revascularization was defined as a modified Thrombolysis in Cerebral Infarction (mTICI) score ≥2B. Uncorrected and corrected volumes were compared between each package and with FIV according to mTICI score. Results Ninety-four patients were included, of whom 67 (71.28%) had a mTICI score ≥2B. In patients with successful revascularization, ischemic core volumes did not differ significantly from FIV regardless of the package used for uncorrected and corrected volumes (p>0.15). Conversely, assessment of PDV showed significant differences for uncorrected volumes. In patients with unsuccessful revascularization, the uncorrected PDV of packages A (median absolute difference −40.9 mL) and B (median absolute difference −67.0 mL) overestimated FIV to a lesser degree than package C (median absolute difference −118.7 mL; p=0.03 and p=0.12, respectively). After correction, PDV did not differ significantly from FIV for all three packages (p≥0.99). Conclusions Automated MRI perfusion software packages estimate FIV with high variability in measurement despite using the same dataset. This highlights the need for routine expert evaluation and correction of automated package output data for appropriate patient management.

learn more

Introduction: We performed a non-inferiority study comparing magnetic resonance angiography (MRA) techniques including contrast-enhanced (CE) and time-of-flight (TOF) with brain digital subtraction arteriography (DSA) in localizing occlusion sites in acute ischemic stroke (AIS) with a prespecified inferiority margin taking into account thrombus migration. Materials and methods: HIBISCUS-STROKE (CoHort of Patients to Identify Biological and Imaging markerS of CardiovascUlar Outcomes in Stroke) includes large-vessel-occlusion (LVO) AIS treated with mechanical thrombectomy (MT) following brain magnetic resonance imaging (MRI) including both CE-MRA and TOF-MRA. Locations of arterial occlusions were assessed independently for both MRA techniques and compared to brain DSA findings. Number of patients needed was 48 patients to exclude a difference of more than 20%. Discrepancy factors were assessed using univariate general linear models analysis. Results: The study included 151 patients with a mean age of 67.6±15.9years. In all included patients, TOF-MRA and CE-MRA detected arterial occlusions, which were confirmed by brain DSA. For CE-MRA, 38 (25.17%) patients had discordant findings compared with brain DSA and 50 patients (33.11%) with TOF-MRA. The discordance factors were identical for both MRA techniques namely, tandem occlusions (OR=1.29, P=0.004 for CE-MRA and OR=1.61, P<0.001 for TOF-MRA), proximal internal carotid artery occlusions (OR=1.30, P=0.002 for CE-MRA and OR=1.47, P<0.001 for TOF-MRA) and time from MRI to MT (OR=1.01, P=0.01 for CE-MRA and OR=1.01, P=0.02 for TOF-MRA). Conclusion: Both MRA techniques are inferior to brain DSA in localizing arterial occlusions in LVO-AIS patients despite addressing the migratory nature of the thrombus.

learn more

Background: Predictive maps of the final infarct may help therapeutic decisions in acute ischemic stroke patients. Our objectives were to assess whether integrating the reperfusion status into deep learning models would improve their performance, and to compare them to current clinical prediction methods. Methods: We trained and tested convolutional neural networks (CNNs) to predict the final infarct in acute ischemic stroke patients treated by thrombectomy in our center. When training the CNNs, non-reperfused patients from a non-thrombectomized cohort were added to the training set to increase the size of this group. Baseline diffusion and perfusion-weighted magnetic resonance imaging (MRI) were used as inputs, and the lesion segmented on day-6 MRI served as the ground truth for the final infarct. The cohort was dichotomized into two subsets, reperfused and non-reperfused patients, from which reperfusion status specific CNNs were developed and compared to one another, and to the clinically-used perfusion-diffusion mismatch model. Evaluation metrics included the Dice similarity coefficient (DSC), precision, recall, volumetric similarity, Hausdorff distance and area-under-the-curve (AUC). Results: We analyzed 109 patients, including 35 without reperfusion. The highest DSC were achieved in both reperfused and non-reperfused patients (DSC = 0.44 ± 0.25 and 0.47 ± 0.17, respectively) when using the corresponding reperfusion status-specific CNN. CNN-based models achieved higher DSC and AUC values compared to those of perfusion-diffusion mismatch models (reperfused patients: AUC = 0.87 ± 0.13 vs 0.79 ± 0.17, P < 0.001; non-reperfused patients: AUC = 0.81 ± 0.13 vs 0.73 ± 0.14, P < 0.01, in CNN vs perfusion-diffusion mismatch models, respectively). Conclusion: The performance of deep learning models improved when the reperfusion status was incorporated in their training. CNN-based models outperformed the clinically-used perfusion-diffusion mismatch model. Comparing the predicted infarct in case of successful vs failed reperfusion may help in estimating the treatment effect and guiding therapeutic decisions in selected patients.

learn more

Background: After radiochemotherapy, 30% of patients with early worsening MRI experience pseudoprogression (Psp) which is not distinguishable from early progression (EP). We aimed to assess the diagnostic value of radiomics in patients with suspected EP or Psp. Methods: Radiomics features (RF) of 76 patients (53 EP and 23 Psp) retrospectively identified were extracted from conventional MRI based on four volumes-of-interest. Subjects were randomly assigned into training and validation groups. Classification model (EP versus Psp) consisted of a random forest algorithm after univariate filtering. Overall (OS) and progression-free survivals (PFS) were predicted using a semi-supervised principal component analysis, and forecasts were evaluated using C-index and integrated Brier scores (IBS). Results: Using 11 RFs, radiomics classified patients with 75.0% and 76.0% accuracy, 81.6% and 94.1% sensitivity, 50.0% and 37.5% specificity, respectively, in training and validation phases. Addition of MGMT promoter status improved accuracy to 83% and 79.2%, and specificity to 63.6% and 75%. OS model included 14 RFs and stratified low- and high-risk patients both in the training (hazard ratio [HR], 3.63; P = .002) and the validation (HR, 3.76; P = .001) phases. Similarly, PFS model stratified patients during training (HR, 2.58; P = .005) and validation (HR, 3.58; P = .004) phases using 5 RF. OS and PFS forecasts had C-index of 0.65 and 0.69, and IBS of 0.122 and 0.147, respectively. Conclusions: Conventional MRI radiomics has promising diagnostic value, especially when combined with MGMT promoter status, but with moderate specificity. In addition, our results suggest a potential for predicting OS and PFS.

learn more

Background: Neuroprotection for acute ischemic stroke remains an elusive goal. Intracranial collaterals may favor neuroprotective drugs delivery at the acute stage of ischemic stroke. A recent phase 2 study showed that cyclosporine A (CsA) reduced ischemic damage in patients with a proximal occlusion who experienced effective recanalization. Collateral flow may improve this benefit. Materials & methods: Collateral supply was assessed using dynamic susceptibility contrast MRI in 47 patients among the 110 patients from the original study and were graded in two groups: good collaterals and poor collaterals. Patients with good collaterals had significantly smaller initial infarct in both CsA group (p = 0.003) and controls (p = 0.016). Similarly, the final lesion volume was significantly lower in patients with good collaterals in both groups. Results: In patients with either good or poor collaterals CsA showed no additional benefit on ischemic lesion progression and final infarct size at day 30. Conclusion: We failed to demonstrate any significant additional benefit of CsA in patients with good collateral circulation.

learn more

Background and Purpose- The development of leptomeningeal collateral artery network might be adversely affected by small vessel wall alteration. We sought to determine whether small vessel disease (SVD) burden may impact collateral development in patients treated by mechanical thrombectomy for anterior circulation acute ischemic stroke. Methods- The patients admitted in our center for anterior circulation acute ischemic stroke and (1) treated by mechanical thrombectomy with or without thrombolysis and (2) who underwent a baseline magnetic resonance imaging were included in the study. The SVD burden and the pial collaterality were assessed through the cerebral SVD score (severe when ≥1) and the Higashida score (favorable when ≥ 3) on magnetic resonance imaging and digital subtraction angiography, respectively. Any association between the cerebral SVD score and the collaterality were assessed through comparative and regression analyses. Results- Between January 2013 and March 2018, 240 patients met the inclusion criteria (68.7±16.1 years old; 49.2 % female). The cerebral SVD scores were of 0 in 125 (52.1%), 1 in 74 (30.8%), 2 in 30 (12.5%), and 3 in 11 (4.6%) patients. Hundred and thirty-six patients (58.1%) presented a favorable collaterality score. The favorable collaterality subgroup presented a significantly higher proportion of female (79%), lower baseline National Institutes of Health Stroke Scale ( P<0.001), and higher Diffusion-Weighted Imaging-Alberta Stroke Program Early CT Scores ( P<0.001). The regression analyses showed no impact of the cerebral SVD score on the collaterality pattern (odds ratio, 1.11, 95% CI, 0.82-1.50; P=0.51). Conclusions- In patients with anterior circulation acute ischemic stroke, collateral flow status does not seem to be influenced by SVD burden.

learn more

Background and purpose: Hemodynamic and metabolic impairment in intracranial atherosclerotic stenosis (ICAS) may promote stroke vulnerability particularly in borderzone areas. Perfusion and oxygen mapping magnetic resonance imaging (MRI) may provide useful information in this setting. Methods: In this pilot study, patients with symptomatic atherosclerotic anterior circulation stenosis ≥60%, without other sources of ischemic stroke, were included. High-resolution vessel wall MRI quantified the stenosis degree, and hemodynamic and metabolic impairment was assessed at baseline using dynamic susceptibility contrast perfusion and multiparametric quantitative blood-oxygen-level-dependent (mqBOLD) oxygenation MRI. All parameters were assessed within both hemispheres and in borderzone areas. Results: Forty-three subjects with intracranial artery narrowing were screened from November 2014 to January 2016. Eleven patients met the study criteria (mean ± standard deviation age = 64.4 ± 10.6 years, the mean degree of stenosis was 76.9 ± 23.4%). No interhemispheric differences were observed across oxygen (cerebral metabolic rate of oxygen and tissular saturation of oxygen) or perfusion (mean transit time, time to maximum, Tmax , normalized cerebral blood volume [nCBV], and normalized cerebral blood flow) parameters. A positive correlation was observed between the stenosis degree and ipsilateral nCBV (R = .77, P = .008). In addition, a significant increase in CBV was observed in anterior cortical borderzones ipsilateral to stenosis (nCBV = 7.20 ± 1.81 vs. 5.45 ± 1.40 mL/100 g, P = .02). Conclusion: Symptomatic ICAS had no global impact on perfusion and oxygen mapping MRI at resting state. A significant increase in nCBV was found within anterior borderzone areas.

learn more

Objectives The aim of the present study was to assess the association between collateral status and DWI-FLAIR mismatch in patients with acute ischemic stroke within the 4.5 h time-window. Methods We analysed DWI, FLAIR, and PWI data in patients within 4.5 h after symptom onset from the I-KNOW European database. Collateral flow maps were graded by analyzing contrast ‘staining’ extent over the early, mid and late perfusion phases. ADC values, DWI lesion volume, and normalised perfusion parameters (CBV,Tmax) within DWI lesions were determined. Visibility of parenchymal hyperintensivty on FLAIR was evaluated (“FLAIR positive”), and DWI-FLAIR mismatch was assessed. Spontaneously reperfused regions were defined as voxels with Tmax <6 s within the DWI lesion. Final infarct size was assessed on day-30 FLAIR images. Results Of the 168 patients included in I-KNOW database, 87 were eligible for this study. DWI-FLAIR mismatch was present in 69 patients. There was no difference between poor and good collaterals status according to age, sex, baseline NIHSS score, time to MRI and DWI lesion volume. Collateral status was significantly better in the FLAIR positive group (p = .001). Patients with poor collaterals had significantly increased Tmax (p = .005). Baseline DWI lesion volume and final lesion volume were significantly smaller in patients with good collateral status (p < .001 and 0.01, respectively). Conclusions We found that patients with early FLAIR lesion visibility have a better collateral status. This finding has implications for the management of stroke patients with unknown time-of-onset, and more widely should be considered in the current context of extending the therapeutic window.